← Course overviewFree lesson 1 of 3 · lesson 1 of 10Next lesson →

Unlock all 10 lessons of Crashworthiness Analysis using HyperMesh and RadiossFull course ₹40,000 · paid

Price & syllabus
Free lesson · 296 min · no payment

FEA Basics and HyperMesh Geometry, 2D and 3D Meshing

A free lesson from Skill-Lync's Crashworthiness Analysis using HyperMesh and Radioss course, on Grow by Skill-Lync · Last updated 16 Nov 2026 · Session: “1.0 - Basics of FEA”

What will you learn in this free Crashworthiness Analysis using HyperMesh and Radioss lesson?

The instructor introduces himself, lists the engineering prerequisites, and explains finite element analysis, its linear, non-linear, static and dynamic types and the idea of time integration. He then moves into HyperMesh geometry cleanup, mid-surface creation, 2D shell meshing and a 3D solid meshing example.

  • Linear versus non-linear and static versus dynamic
  • Mid-surface extraction and geometry cleanup
  • Quad and tria shell elements, 6 mm thickness rule
  • Building a 3D solid mesh with solid map

What comes after lesson 1?

The next free lesson is lesson 2, “1D Elements, Radioss Input Files and Critical Time Step” (66 min). Later in Skill-Lync's Crashworthiness Analysis using HyperMesh and Radioss course (paid): lesson 4, Materials Part 1; lesson 5, Hypercrash Part 1; lesson 6, Kinematic Conditions Part 1. The full course has 10 lessons (25 h 22 min of video); see the course overview or its price and syllabus.

Lesson transcript

Machine transcript of the lesson audio (lightly cleaned). Technical terms may contain errors; the video is the reference.

Part 1

Hello everyone, welcome to the crashworthiness analysis class using HyperMesh and radios. So in the first part of this class, we'll be learning HyperMesh, the basics of HyperMesh and how to create geometries, how to preprocess some of the load cases and submit them for initial analysis. So while we're doing that, first of all, let me introduce to you. My name is Sandesh and I have been working in the automotive crash and safety industry for three years now. Currently, I'm associated with Ford Motor Company and I work with mainly the passenger car crashworthiness and safety.

So this Ford opportunity, just before that, I used to work with Altair. That is where I learned a great deal about radios as well as HyperMesh as well or hyperworks as well. And my main FEA knowledge, it all started from Tata Technologies back in Pune who gave me an opportunity on Jaguar and Rad Rovers. And that was really the biggest step up for me as a FEA engineer. So from there, the projects until now I have completed are consisting of airbags and injury analysis, off-road impacts of the cars and one of my master thesis project, bullet impact on a glass cylinder and then modeling its crack propagation.

So with that, recently I have completed my master's in mechanical engineering from Michigan Technological University. And now today, I'm sharing my knowledge in HyperMesh and crashworthiness in radios along with you. So welcome to this class. While starting this class, there are a few prerequisites that you should definitely go through while attending this class or before attending this class. First of all, the most important are engineering basics, which includes applied mechanics or engineering mechanics. So degrees of freedom, rigid body motion or friction, what exactly the friction means or if one rigid body is impacting another rigid body, what would be the momentum of first rigid body after impact and before impact?

Such basic calculations and such basic concepts, you should definitely know about those. What is kinetic energy and internal energy or let's say strain energy? Those are few of the important concepts that you should definitely learn, even in strength of material aspects. So apart from the strain energies, the types of stresses and strengths such as shear, tension, compression or any other modes of failure or material properties. So what is engineering stress? What is true stress and strength? Those are few of the important concepts that you should definitely know.

Then shear force diagram, bending moment diagrams. Now, I don't expect you guys to exactly learn all the formulas by heart, but you should definitely know what goes on behind the scenes when we are looking at any particular deformation. So for that purpose, it is good to have that understanding right from the basics. Then apart from that, the design principles for designing any particular component, the theories of failure that go behind that, what is factor of safety or how do you decide that the factor of safety should be 2 while your yield stress or yield strength is going to be at one particular value.

So what should be your design stress? All those calculations, that basics in mind would be a handy information or knowledge for you guys. Apart from that, matrices or matrix operations. So what is a matrix multiplication? How complicated it can be? What is matrix inversion? What could be the methods for that? I don't expect to know the methods in detail, but you should know definitely the challenges that are associated with that. And apart from that, FEA, just the introductory part or theoretical knowledge is sufficient because right in the next few slides, we'll be learning more about FEA.

So in FEA, if you know already what exactly FEA means and what are the element types, stiffness and damping and mass matrix, that will be really handy thing to do. All right. With this, let's start with the basics of FEA. So, FEA by itself, what it means is finite element analysis. So, essentially, finite element means that we are going to separate or split a problem into smaller parts and we are going to solve each and every small part to combine into a solution. Now, while doing this, for the analysis part, there are multiple ways a problem can be analyzed.

And depending on the type of problem, there are different various types of analysis as well. So, the types of analysis, roughly, they can be described as either linear or non-linear or on the other side, they can be static or dynamic. Then, the FEA method can also vary depending on what time integration you want. Now, what is time integration? That is something that we'll be studying very soon. Implicit method versus explicit time integration method. What exactly is that? We'll be looking at it. Then, the shape functions and Gaussian integration.

This is one of the important parts that you should understand right in the beginning that will help you get a good understanding of the engineering or analysis result that a solver is going to throw at you. And looking at those results, you should understand what is exactly going on on one particular element. That could give you an idea while debugging any particular problem in your solution. And then, after completion of this, we'll be starting our HyperMesh sessions right away. So, starting with the types of analysis, let's look at the linear analysis first of all.

So, as the word suggests, linear analysis, it is going to follow the exact word. The software is going to follow the linear path for static calculations. So, as we know, popular stress and strain curve, the stress versus strain curve is very much linear until a yield point. And after that, we see a plasticity zone until the material breaks at the breaking point. Now, in the linear analysis, what software assumes is that the stress and strain curve is going to be linear even after the yield point.

So, as we see, this whole particular region is completely assumed because this being the linear analysis. Now, obviously, this is not always practical. Now, in cases where you have a static load standing on a beam or a bar or a component where you don't expect too much of deformation, you only want to see what kind of stresses are getting developed. In such cases, it could be a good idea to use linear analysis. But in real life, if you have a problem which crosses yield point for sure, then it is a very bad idea to look at linear analysis because once you cross this particular point, whatever value you get is actually a garbage value because the value you should be getting for stresses should be in this particular region and not in this particular region at all.

The linear analysis after yield point becomes pretty much invalid. So, for that, what we had to develop or FEA world had to develop is nonlinear analysis. Now, when we look at nonlinear analysis, there are multiple sources which lead to nonlinearity in a problem. So, just now we saw that material could be nonlinear. That is exactly giving one of the nonlinearities. But apart from that, there are geometric and contact nonlinearities as well. So, the geometric nonlinearities arise from the large deformation problems. Now, in a geometry, as complicated, let's say, as an automotive body, the geometries, there are so many components coming together or assembling together to undergo deformation.

Now, if it was a small deformation, the problem could have been much simpler. But the cases or the analysis which an engineer typically does while designing any particular automotive component is always about some fatigue analysis or durability analysis just to check how durable that component is going to be in under usual load cases such as if you are slamming your car door multiple times, your door still has to sustain those kind of impacts. So, the durability or fatigue analysis becomes very important in such cases. Now, when you are in your car and you hit another obstacle, your car is undergoing a crash.

Now, during crash, your structure also has one particular role to follow. And while doing that or while the structure is getting deformed, such are the cases of large deformation. And during such large deformation cases, the geometric nonlinearities behave or lead a tremendous role in contributing to the nonlinearity of the problem. So, once the structure start getting deformed, the stiffness of the structure goes on changing because the shape of the structure or the shape of the section of one particular component goes on changing and that simply changes the whole stiffness of the structure.

Apart from that, material nonlinearity we just saw for a metallic stress-strand curve because after yield point, the stresses and strain never follow a linear path but the slope of the line is never going to be constant after that. On that front, the material nonlinearity could be from a metal material going beyond elastic limit. So, it is a very important thing for the structure and the emphasis on the structure if it is not Then, the surface yn yw i'w i'w i'w i'w i'w i'w i'w i'w i'w i'w i'w i'w i'w i'w i'w i'w i'w i'w i'w iw i'w i'w iw i'w i'w i'w i'w iw i'w i'w iw Or for non-metals, even within elastic limit, the E-value may not be constant, such as some of the plastics, you see that there are few hysteresis cases where while loading you follow one particular curve and while unloading you follow a different curve.

Now that is one way of expressing non-linearity and that can be a challenge. If you follow a linear analysis for that, you are not going to get the accurate results. And then the creep behavior, that is deformation of material under constant strays. So this is one case of what can be looked at during a fatigue or durability analysis, that if you are constantly loading one particular component in one particular manner, then over the time, you see a permanent deformation in the component. Though your stress and strain are still predictable, they are still within your reach, you can still predict the stress at one particular strain.

But if you keep doing that loading cycle again and again, over the time, you will see that the stiffness of the component has changed in that particular region. And that kind of non-linearity comes from the creep non-linearity, which again goes through material non-linearities. Apart from that, the contact non-linearities. So, in dynamic problems such as a crash analysis, automotive body is made of thousands of components, individual components. Now, they are bound together using welds or bolts. But while binding together, if there is a loading case such as a crash, when your front bumper is getting in contact with the obstacle, once the front bumper gets deformed, it is going to deform the internal components as well, such as hood of the car, then your fenders, your wheels, or even your air conditioning system.

So, all these components have to come in contact with the deforming component or the barrier component. And to define that, a contact is very crucial. So, during, while defining this contact, stiffness of this contact, the way this contact is going to happen or the way the components are going to be pushed against the rigid wall or the barrier, that stiffness of this contact is going to differ depending on how the components are meeting each other. Now, that particular stiffness, we are going to study when we study the crash analysis in more details.

But this is one type of non-linearity that should be definitely looked at. So, after looking at non-linear analysis, we'll look at what static analysis means. So, static, as the word suggests, there is something stationary in this particular problem definition. Now, that stationary thing will obviously will be the loading. So, in static analysis, the loading is typically pretty much stationary. So, over the time, you see that the forces in the system are almost constant. So, one case of static analysis could be you have a cantilever bar and on that bar, you are hanging a hook and on that hook, you have a clock or hanging to it.

So, this clock is never going to be touched. The clock is always going to stay hanging to that particular hook which hangs on the bar. And if you are analyzing that particular bar, the load on that particular bar is always static. That is a very, very small example of static analysis. But in real life, there are many other static analysis cases. So, if you have designed a bracket to mount your fuel tank for a car, that bracket, when the car is not moving, is under a static load of the fuel tank.

When the car is moving, the tank will continuously undergo vibration. It will continuously keep moving. And that time, the case could be different. But until then, when the car is stationary, it is a pure case of static analysis. Now, when the car is moving, obviously, the problem becomes dynamic. And the loads keep varying over the time. And that is where the dynamic analysis comes into picture. So, linear dynamic exactly means the same thing. The forces increase linearly with parameters such as position and velocity. So, not just increase but decrease as well.

So, the force keeps increasing or decreasing depending on the road contour, depending on the velocity of the car. If the car breaks, the fuel tank will try to jump ahead of the bracket on which it is mounted. But the bracket has a task to hold it together, hold it at the same location. And that leads to another type of loading. So, linear dynamic is one such analysis case. Non-linear dynamic is a case where the force, well, they may not vary linearly with the space parameter. So, such as the case of now coming back to a crash analysis.

There is no linearity between the space parameter and the forces that you will experience if you are an occupant. So, when the crash happens, when the bumper is hitting the barrier, you get an immediate shock inside the car. But after that shock, while your front structure is getting deformed, once it starts deforming, the shock starts getting reduced. And over the time, whatever energy you receive inside your passenger compartment, that is the energy that will cause you any injuries, if at all. But the way this whole dynamic is going to work is that as the structure gets deforming, goes on deforming, the opposing force or to the external forces, those are also changing.

The frictional values of many contacts are going to change. So, that is why the stiffness of your car is going to change. That will impact or that will really affect the way you experience things inside your car. So, that is one case of non-linear dynamic analysis. After that, we will quickly move to time discretization. So, if you realize until now, we have looked at linear analysis or non-linear analysis. Then we have discussed static and dynamic analysis. Now, we are looking at the next chapter that is methods of time integration.

So, under that, we have two main points to discuss. Implicit method or explicit method. So, implicit method, again, as the name suggests, so many terms in FEA are going to be self-explanatory. Once you understand this particular topic, just by looking at the word, you will have a clear meaning of it in the back of your mind. And that will really help you over the time when you are trying to decide what kind of analysis you should be doing or what method you should be applying to solve one particular challenge in FEA.

So, implicit time integration. When we are looking at a dynamic analysis problem, very typically, the acceleration or the loading function will be varying with respect to time. Now, because the loading is varying or the load is varying with respect to time, there are obviously some displacement, non-linear displacement or some non-linear velocities that are going to get involved. Now, when the velocity is changing rapidly with respect to time, that means there is some acceleration involved in the problem as well. Now, acceleration could be in the form of actual acceleration, positive acceleration, or it could be in the form of deceleration.

Now, when I drag a cart rapidly towards myself, whatever forces are going to be experienced on the cart, so those forces are due to some positive acceleration that is towards me. Now, gravity is another example of continuous positive acceleration, though we call it negative acceleration in z direction, but those are completely different technical aspects. But the way gravity works, it is always acting positively on every object. Apart from that, deceleration, when you are breaking your car or when you are hitting into some object or cases like that, during that deceleration phase, you experience challenging in stopping or in such cases, that can also be a dynamic analysis.

And again, that leads to many challenges. Now, to solve those challenges, there are two different methods. First of all, implicit method or explicit method. When we look at implicit method, the problems in such cases are typically small deformation problems. So obviously, if you are looking at a crash analysis cases, the deformation are at such a huge quantity that it cannot be termed as a small deformation problem. But small deformation problem could be more like if I am slamming the door, what deformation I am getting on the door rim.

Or on the external surface of the door, so those are very minor deformations and over the time they accumulate together to give you a permanent failure of the door. So such are the small deformation problems. So for small deformation problems, the acceleration in such cases is assumed to vary linearly over the time step. What that means is that just by the words of if I have an acceleration acting on one particular body for 10 seconds, and if the acceleration is increasing from 0 to 10 over the 10 seconds, for every 1 second, if I am assuming that my time step is 1 second each, so every 1 second, the acceleration, what I am going to assume is that acceleration will vary from 0 to 1 during the first second linearly.

Now other cases if you but that may not be the actual case all the time the acceleration could be varying non-linearly or the acceleration could be increasing or decreasing depending on how the structure is behaving so there are many factors the way acceleration could be changing over the time but then the major assumption in this case is that we are assuming that the acceleration is going to vary linearly over that particular time step now when I say that when I say that it is going to vary many calculations that could get involved. Now considering a 1D system a single degree of freedom system with a spring mass with a force applied in one particular direction I can calculate the velocity at one particular time dt with respect to or starting from a time t and that can be calculated using very simple kinematic equations v is equal to u plus a t.

Now in this case a has to be calculated and it has to be assumed that because it is varying linearly or particular that dt time I have to say that the initial acceleration and the final acceleration at and at plus dt over the time dt because it is linearly varying I can simply average it and I can get v is equal to u plus a t. Now to get this a I had to average the first and the last term and depending on this velocity equation now I can derive lot many quantities such as displacement then I can also go for the forces now if I say what could be the force or final force acting on this particular body what I have to calculate is first of all the F0 F0 is the external force force and I have to subtract the force and I have to subtract the spring resistance from that now this resistance of inside the spring can be calculated using the simplest equation stiffness of the spring into displacement if I subtract this from the external force and if I subtract this from the external force quantity that is F0 I'll get the force which is acting on mass m in the end this force is actually it can be expressed in terms of mass and acceleration and developing that equation further I end up I end up with this particular equation over here F0 is equal to m at plus k into x t plus dt and plus k into x t plus dt now if I develop this F0 further I end up with this final equation where I have only acceleration on one side and rest of the terms on the other side now to solve this problem obviously it becomes a matrix problem when when I have multiple degrees of freedom of system so make it a matrices problem now I see that though acceleration is on one side I have to first of all inverse the K matrix and I have to multiply it with rest of the components in the system to get that answer for my acceleration now this acceleration is going to help me calculate the final displacement as well or this acceleration is going to help me calculate the forces acting on this particular mass now this force or this force Ft will tell me what kind of strain is going to come or what kind of stresses that are going to form on that particular component and that is why at plus dt needs to be calculated now in this process k needs to be inverted and to invert the k now you can imagine the complexity of this matrix inversion if you have one degree of freedom system if you have one degree of freedom system your matrix is actually 1d that is why you can simply carry out these operations by hand but when you have multiple degree of freedom system let's say if you had seven different masses acting in in a row and you had only one force acting at the end of it in such cases because you have seven degree of freedom system this k matrix will be a seven by seven matrix and not just a diagonal matrix but it will have values everywhere in the matrix now inverting that matrix now inverting that matrix becomes very tricky now just for a seven degree of freedom issue or seven degree of freedom problem we are experiencing that kind of a challenge then you can imagine when you are looking at a problem with number of nodes like of the order of millions you can say that the stiffness matrix in that particular case will be a million rows and million columns now to invert that matrix you can imagine what kind of computational resources it could take and that exact is the reason the convergence or getting a solution in the case in the end becomes very difficult in implicit time integration now let's compare that with explicit time integration now in implicit time integration if you remember in the beginning i said that we are going to accelerate a body from 0 to 10 seconds with one particular acceleration that also starts from 0 and reaches some value so until now we have been assuming that it varies linearly over the time over the time step and time step we have been considering was only one second each every time now many of the complexities in this particular equation arise because we have two different terms for simply calculating the acceleration now what if we had this acceleration assumed to be linear over this particular time step for that obviously i have to reduce this time step to such a small uh small level that the acceleration will be almost linear acceleration will be varying from 0 to 0.00001 quantity or unit in such cases i can really assume that over this this particular time step the acceleration is almost linear and then i can see that during the next time step i am getting the acceleration directly at the next level or the next value and not at the previous value and that is perhaps perfectly acceptable because the difference or in the jump in the acceleration is not too high anyway or the from one time step to another time step so with that particular assumption now my velocity calculation becomes very easy that leads to even easier calculation for displacement and then the final force acting on that particular mass now if i develop that equation further for the same acceleration at at at the end of the time step i have to calculate the initial force and then the k matrix now this time the k matrix is in the denominator or rather numerator and i don't have to i don't see any inversion sign on that so i don't have to invert it anymore whereas i see the m in this case is uh in the denominator that means in the multiple degree of freedom system the k matrix being in the numerator doesn't need to be inverted but the m matrix need will uh will need to be inverted but the the good thing about this m matrix is that this is always just a diagonal matrix so if i have three degree of freedom system i i will have in the m matrix or the mass matrix i have only three different values uh existing at only at the diagonal elements and that makes it very easy for me to invert that matrix so if i look at the final equation for multiple degree of freedom system for at plus dt i see that only mass matrix needs to be inverted rest of the things can stay as they are that makes this explicit time integration a very simpler process uh to solve for this particular equation but obviously this comes at a cost now because you are assuming the acceleration is constant over the time step you have to uh you have to take those time step very very small and that small time step will lead you to a consistent solution if you jump this time step to a higher values then you will see that you are getting error in your calculation because since you are assuming the acceleration to be constant from first time time step to the next time step i see that there is a huge difference between the times uh accelerations and that is not really matching the loading condition and that will simply create a instability in your model so just to summarize the differences between implicit and explicit which could be confusing concepts many times the deformation essentially uh for explicit analysis there are large deformations so typical example could be a crash analysis case then what causes this large deformation the way the problem has to be defined or the way the loading condition are that can also help you decide whether to follow implicit or explicit method so if i have implicit uh velocities or accelerations acting on the mass which are varying over the time and there are there is inertia of the body involved then i i can say that uh it is definitely an explicit problem than being an implicit problem the k matrix inversion that is that is about the mathematical process that goes behind this behind the scenes the k matrix inversion is needed in implicit not needed in explicit that makes it easier to solve the time step size whereas large time step size in implicit analysis makes it perhaps faster to solve in simpler analysis cases whereas in explicit because you have to assume that acceleration to be constant your time step will be very small now what limits the time step as i mentioned earlier uh accuracy of the problem will be defining your uh implicit time step whereas the stability of the problem will be defining your uh time step in in the explicit analysis the cost per time step because now the time step are larger in implicit even if it exists it We'll be right back.

The cost will be large, whereas in explicit, the cost per time step is small, but now we have a huge or tremendous number of time steps. So in the end, the cost could still be an issue. What domain is this cost? Matrix inversion in implicit? Yes, element processing because now in this case, you have huge number of time steps and for every time step, you have to calculate the results on each and every node and element. Now that creates a lot of ruckus and that really creates a lot of challenges for the computation or for the computers.

So, just for an example, in an FEA model, if you have a mesh and if you double this density, that means if you reduce the mesh size by half, the cost wise or the time that could, that the implicit analysis would take now could be of the order of 128 times. Whereas for explicit analysis, it could be only as 16 times because in the end, your end calculation or the way it is trying to calculate your acceleration is simpler equation and that makes it easier. Whereas if you increase the mesh density in implicit analysis, you can imagine the stiffness matrix reaching another level, another complexity level and that will lead to more time.

Parallel processing, well, it is possible in implicit, not possible in explicit. Parallel processing only means that in one particular problem, you can split the problems into many, many parts. And that we will try to solve problems or that can be tricky at times, but for now, we can, we can skip that topic. And when the time comes, we'll discuss this parallel processing in more details. So just to summarize implicit and explicit with few more examples, there are, there is a good overlap between implicit method application and explicit method application.

So if there is any structural problem where the signal, where the force is stationary and you just want to see if the structure is going on, undergoing any plastic deformation, then it's a good, good enough problem for a static analysis case. Quasi and quasi static analysis is something where the force is almost constant over the time. So if you see the symbol over here, it is approximation symbol and the force is almost zero or the change in the force is almost zero over the time. So metal forming process, if it is, the process is having very small metal of a deformation can be used, can be solved using implicit method.

Whereas having large deformation, it can go to explicit method analysis as well. So quasi static problems, that, that is the exact term in such cases where quasi means almost and your, your force is linear, varying over the time. But because it's the, the rate of variation is very low, it is the change in the force is almost called constant. Now in explicit method, impact problem becomes very typical example. Your force is actually, um, your external force, that is external mass, uh, mass impacting onto your body, uh, with an acceleration M and, uh, with an acceleration A and that gives you the forces acting on the, on the system.

So in another terminology, if you have elastic or plastic or buckling, uh, problems, um, or to, up to certain extent with static loading, you can solve it using implicit analysis. Whereas for, uh, if you have loading conditions dynamic or if your velocity is dynamic and the non-linearity in the problems can lead to rupture or damage. Then that becomes a definite explicit analysis problem and cost wise, uh, the complexity of, of the problem, if it is small or if it is a less complicated problem where the deformations are small and the loading condition are not too stringent.

In such cases, you will see that the implicit method is actually very economical compared to explicit, whereas explicit method, if you increase the complexity, uh, over the time, the, the, the cost per effort becomes smaller. And that makes it economical with, uh, as compared to implicit method, as I mentioned just now, the quasi static simulation quasi means almost. And the force or the time is actually varying, uh, in a very slow manner. And that, that makes it a quasi static, uh, analysis simulation. Now, after that, we'll be looking at the discretization systems.

So if you, if you look at a continuous body, uh, as you know, a finite element analysis, it's all about finite elements. So that means that you are dividing or splitting a problem into a definite number of elements. Now this discretization, that means, uh, splitting a problem into smaller, smaller parts. This discretization, uh, there are many types of this discretization. First of all, uh, or the most important, the one which we see by, with our eyes is the space discretization. Where you are, when, where you are splitting your objects into small elements and you are trying to solve a problem.

With this space discretization, uh, the continuous system gets converted into a discrete, uh, discrete system. And during this, uh, the materials of this particular, uh, continuous system model need to be modeled in a way that they can be applied on an FEA model as well. So for a FEA model, there are essential properties that are needed to be defined. And that becomes a challenge many times to get those properties. Now for, uh, for elements which are connecting the nodes and the nodes are actually splitting your problem into smaller parts.

So the calculations are actually made on the nodes. So this discretization process, it is actually separating the problem on the nodes. And all the calculations of the forces on the acceleration or the stress and the strain will happen at the nodes. And the results are actually interpolated and they are reflected on an element. So whenever you see the contours of an FEA result, you actually see the element, the results on the element. But what goes behind that is actually the result has been calculated on the nodes.

And then it is, it is getting interpolated on the element. Now for this particular fin, uh, the temperature is varying over the length of the fin and the way temperature is, uh, varying is not really linear. So to capture that, uh, you see that this fin is getting split into smaller components. And those components are now, uh, are bound by nodes and two adjacent nodes are, are constituting to an element over here. So now we see that this bar is, uh, getting, getting split into five nodes and those five nodes form four elements.

And now those four elements, uh, if we follow a linear approach with them, then we see that those four elements, uh, will, uh, will show you that the temperature is varying linearly over them. And, uh, while we move from one element to another, you see that the profile of the temperatures try, uh, uh, is getting followed very roughly, but somewhat in the same direction. Whereas if you make these elements quadratic, that means they can also capture the curve. Uh, for the temperature. So with a quadratic element, you get little more details on the temperature profile now, but ultimately, uh, what drives the cost is the actual solving time and the accuracy as well.

What has been found over time is that while using four, uh, instead of using two elements over here, two quadratic elements, if you use four different elements over the bar, you get more predictive and more accurate results. Uh, as compared to this, because when you involve two quadratic approximation, then it triggers a lot, many other mathematical calculation that leads to more complexity. Now the shape function. Now, when we see, uh, that there are elements and over the element, we'll be looking at few results in between these two nodes.

Now, what, what gives the results between these two nodes, as I mentioned earlier, the results will be calculated only at nodes and not on the element. So what gives, uh, this interpolation over the element. So this, the, the function of this, this interpolation over the element is actually leading us to the next topic. That is shape function. And this interpolation method over this element length, because this is being 1d element being linear. We see that the function over the, uh, the temperature function or the length is linear.

So the shape function for this is linear. Now, in cases for a 2d, uh, or quadratic element, the shape function will be more like a quadratic curve. And so if you have a, uh, result at node, uh, node one and node two, or T, uh, then, uh, using that particular quadratic equation, you can easily jump to some point in between them. And that gives you, that can give you more accurate results, perhaps. So the shape functions or the information on this shape functions as to, as to what equation to follow to calculate the results in between the end nodes of the element.

And, uh, so those expressions or the equations are carried, uh, with the element itself. So. In popular FEA course, all the FEA course in fact, if you select an element, you have to select the property of that particular element and that property or the option where you select, you want to make a 2D element or a 3D element or 1D element, when you call for that particular type, it will trigger the shape function associated with that element type and that way it gets assigned automatically. So, not that we have to worry about shape function assignment, but we should definitely know what exactly the shape function is doing when you change the element types for some particular problem.

So, the shape function is giving us the interpolated results. Now, when we look at the nodal results though, if you look at the explicit time integration function, Now, just for calculating acceleration, we have to undergo such huge matrix calculations. Now, to simplify those, many times, it is possible to express that same problem in terms of a series problem. So, very typically, if you are trying to solve an integration problem, you can actually use the Taylor series or some other series and you can get a series of a number of elements, a number of terms and if you add those as per what the series rule says, you get the final result which is much closer to your actual result.

In case of implicit analysis, you have inversion of K matrix. In case of explicit analysis, you have inversion of M matrix. So, this inversion can demand a significant computational time. In terms of stiffness, that computational time is too high, whereas for mass, it is much easier, much lower. Instead of solving this problem analytically by inverting the whole matrix, what the solver can do, FES solver can do, is follow the series approach to calculate the results or approximate results. And this is exactly where all the approximation in FEA comes into picture.

So, what it does, instead of calculating the results at the nodes, it will try to put some integration points over the element space and using the shape functions, using loading conditions, the loads at these particular integration points will be calculated first. The stresses and strain will be calculated in integration points first. And then over the time, now you can imagine that this element is going to be connected to some other element. And on this side also, it will be connected to some other element. Now, in such cases, these integration points around this particular node, there could be four or five different integration points around this particular node.

Those integration points will combine together and that will give us the nodal result over this particular node. Now, if the distance, so if I have a different type of element just besides this particular node, which is sharing this node as well. So, while deciding these integration points or quadrature points, there is a weighing function assigned to it. And this weighing function will come into picture when we are trying to calculate the nodal result. So, if you have this particular integration point associated very close to this particular node, and then you also have this particular integration point, which is quite far from this particular node compared to the first integration point, then you see that the weighing function for them is also different.

And that will, so all these components will be added in the series, and in the end, you get the final accurate results or almost accurate result. So, that way, the Gaussian quadrature is going to help you with nodal results, whereas the shape function is going to help you with the results that come up over the whole element space. So, that way, as of now, what we know from today's session is that the types of analysis, we have linear, non-linear, static, and dynamic. Then we have two different integration methods, implicit method and explicit method.

The crash analysis cases, all of them fall under explicit method. The shape function for an element is responsible for interpolation of nodal results. And this interpolation is going to help us increase the accuracy of the solution. So, better the shape function, more the accuracy, but more the cost as well. As for Gaussian integration, it helps calculate the nodal results more accurately and more easily. And that is why Gaussian integration, again, becomes very important parameter. So, in some cases, now we saw for Gaussian integration for the shell element, we had, or the rectangular element, we had four different Gaussian integration points.

In some cases, such as triads, they could be less. And in some other cases, there could be only one integration point. In such cases, the nodal results could be more difficult to calculate. And that leads to another level of challenges and that we are going to study in the near future. So, after completion of the basics of APA, we'll be moving to HyperMesh in the very next session. Stay tuned and see you again. Thank you for joining me today.

Part 2

Hello everyone, welcome back to our HyperMesh class. So today we'll be starting with HyperMesh software with its GUI and where it is used in FEA process. So continuing from the last lecture where we learned about the basics of FEA, what exactly are the types of FEA analysis that are needed to be done. After that, for every analysis, while submitting any particular problem to a solver, we have to first set up the problem in FEA software. Now this setting up of problem is called pre-processing. So HyperMesh is actually a pre-processor.

FEA process, it runs from a pre-processor to a solver to post-processor. So after HyperMesh, you submit the whole file to a solver and the solver can be many options depending on the type of analysis that you are doing. And then the result file can be read in a post-processor to read the animation and the plots or the graphs from that. So in this particular session, we'll be looking at the basic of FEA process quickly. And after that, we'll be moving to HyperMesh Graphical User Interface. We'll be trying to understand toolbars, panels, and the basic functionality of them.

And after that, we are planning to move to geometry pre-processing. After that, we'll try creating the mesh. After that, we'll check the mesh for quality criteria. And after that, we'll be looking at the types of elements over the time as well while creating the mesh itself. So a typical FEA process, well, it goes from importing a CAD or CAE data all the time. So in your project, once you create a CAD model in some CAD software such as, there are many examples in the market. So once you create that, you want to convert that into a FEA model.

And for that, you need to import that data to an FEA software. So you get that CAD and many times what happens is, a user already has a CAE model. He wants to modify that and run another analysis on that. So in that case, even CAE data can be a good data for an FEA analysis. So both types of data, they can be imported. And after that, you try modifying the mesh or you try creating the mesh using the CAD data or CAE data. And after that, you apply boundary condition, loads, and any other conditions that need to be applied.

And then the model becomes ready for solving. So by solving, what I mean is, once you create all those loads, the solver gets to understand what load is to be applied, what particular node, and after that, what forces or what kind of stresses that are needed to be calculated. So to understand that, every solver needs the data to be parsed in a particular manner or in a particular fashion. And that is why preprocessor play a very important role in that particular case. So preprocessor, which is HyperMesh in this case, it exactly does this preprocessing for many industrial solvers such as Radios or Optistruct or LS Dyna or Abacus or Ansys and many others.

So if you look at these names, they are known or they are specialized for different types of analysis, whether it is explicit analysis or linear on non-linear implicit analysis. LS Dyna is known for explicit or implicit analysis as well. Abacus for the same case. And Ansys also for the similar cases. So for all those, each of these solvers have their own particular format in which the data needs to be parsed. And that data writing task is a pretty tedious task and that can be carried out using HyperMesh.

So what HyperMesh does is it will give you a graphical user interface. You do the operations the way you like, the way you can visualize. And then HyperMesh will create or convert that information into a text-based file as per the format specified by the solver. And then we can submit the job to the solver using that particular text file. So after you get the results from solver, it can be post-processed. And for post-processing, again, in the market, there are many softwares available such as HyperView, Hypergraph, then MetaPost or even D3Plot and THIS and many others.

So all these post-processing softwares also come with many pre-processors together. So many times once you operate HyperMesh, it's easier for you to switch to hypergraph just to quickly check the post-processing results. So to start with HyperMesh, the basic step to follow to start HyperMesh session is first of all you go to start and then you simply type hyper and you'll simply get the shortcut to hyperworks or HyperMesh desktop. But before that, I hope that you must have already installed HyperMesh on your machine. If you have student edition of HyperMesh, then you will see this kind of name, hyperworks 2017 student edition, depending on what version you have.

And after that, it shows the kind of solver that I have selected in the beginning. Now we'll see step by step how to select such things. So HyperMesh, as I mentioned earlier, it is a pre-processor in FEA and it can be used to convert a CAD geometry or FEA model, another FEA model into a useful load case, which we want to solve in some other solver or some particular solver. So during that process, you can set a boundary conditions as well as create a solver input file.

So good thing about HyperMesh. Yes, it supports hyperworks solvers such as AccuSolve or Radius or OptiStruct or even MotionSolve. Apart from that, it also supports the external solvers such as LS Dyna or Abacus or Ansys as well and many others. While starting HyperMesh, let's follow the step quickly. We will go to start and I'll type hyper and I get the student edition shortcut on my computer. So usually the loading of this file can take some time depending on whether you are opening your first session after starting the computer or whether you are opening a new session, but there was one session already run on your computer.

Now in my case, because one session was already run on this computer, it takes a little less time to load. This is like your first ever session, then it could take some more time to load. So please be patient while it loads. Now while loading HyperMesh, there are three things that we should understand. When I'm trying to search the shortcut for HyperMesh or hyperworks, if I right click on this and go to the properties of this, then I'll see three different files. So if I take the shortcut of hyperworks to my desktop right here, if I go to the properties of this, I see that there is a target file which is linking to the application for hyperworks, but after that, it has a start in directory.

This is a very important location of each and every hyperworks session will be opened in your document folder. So you can always change this folder by using some path on your own as long as it is valid, it will accept that path and each and every session will start in that folder. So by starting a session in that folder, what it exactly means is that now I'll try to go to my documents and I'll show you what exactly is happening over there. So I'll go to my documents, which I have already opened for demonstration.

So in my documents, I see there are command.tcl, HM menu, HW settings and some backup file for HW settings. So HW means hyperworks, HM means HyperMesh and command.tcl is a general purpose backend file for all the hyperworks applications or many hyperworks applications, not all. A good explanation for those files is as per what the help documentation always suggests. So HM menu.set, the first one of those files, what it does is while you are using hyperworks session or HyperMesh session on your computer, it will constantly store the information about panel options or panel settings.

So all the hyperworks internal settings that you do as per your own preferences or whether you select some solver for your load case setup or if you are doing any graphic settings such as you are changing the background color or whether you are setting some quality criteria for elements. So all such settings, they will be retained for each and every session once you set them. And it is all possible because this HM menu.set file is getting created right after you start your first session. So when you complete your session, when you close hyperworks, it saves all the settings in this particular file.

And when you open a new session, so it will first try to refer to this particular file to get hold of all the settings. And that way you, you save a lot of time doing those settings again. So very handy file. Many times if you run into some issue with your hyperworks, let's say you are not able to set something or you are getting some abnormality, some error is happening on hyperworks itself. It is usually good idea to delete this file and then restart the session.

So in that case, all these settings get reset and that way it is much more like a soft reset in computer language. HM settings, it does a similar job, but it stores information about the browsers or user interface layout. So this is more about the default look of HyperMesh. If you are changing that, then then that will be stored in HM settings dot TCL. Now TCL is itself a program. Language. So whatever you store over here or whatever gets stored over here, it runs like a script in the background when you open a new session of Hyperworks. And that way Hyperworks just tries to keep that continuity all the time in picture. So another handy file for everyone to use. And after hwsettings.tcl, so the third file in the list comes as command.tcl. And command.tcl is a very interesting file. So it stores information.

information about the live session and open session in your HyperMesh window. So I have a HyperMesh session open over here. If I hit new model, there is some command that goes into the software which runs in the backend and it does some operation on the screen. When I say change something to, let's say if I hit open button and if I open some file, then there is some command that is going in the background to open this file. Now this command.tcl is exactly taking care of such things. So if I open this file using a Notepad++ or any kind of text editor software and if I look at all the history, it has stored the history of all my sessions until now. So the operation which I just did, first of all, I opened a new model. I tried to open a new model. So it gave the delete model signal to the software which cleared any previous data that was there on the screen.

After that I tried to open a new file. So the read file operation went on and then on this particular path, I tried to open this particular file. And that way it has stored all the backend information that is needed. So if I literally take this line, copy this line and run this in the command window of Hyperworks, then it will do the same operation all over again. So this is much more like, again, a TCL language, TCL programming language syntax. So if I copy these and if I can literally make a script out of this to do a repetitive operation. So another useful file is command.tcl. And as I mentioned earlier, if we ever feel that there is a there is some issue with my settings in Hyperworks, you can simply delete these three files. It doesn't hurt you much, but it just that some of the settings will be missing next time. So you'll have to set few things again. Shouldn't be a big deal at such because Hyperworks itself by default comes with many shortcuts. So there is less work to do for all of us. So just now in front of you, I opened a file. But before that, let's get back to what exactly are the things inside the graphical user interface. So when I open the window for Hyperworks, there is a lot of information that I see on the screen. First of all, there is title bar. And right after that, just like any usual software on Windows platform, after title bar, we have a menu bar where I have many options such as file operation menu or edit menu or to view something on the screen such as browsers or toolbars.

So view button, then collectors, geometry and all such things that are required all over the operation during pre-processing. So such options are available in the menu bar. So essentially all the operations in Hyperworks that are possible, their shortcut is always available through the menu bar. If you select any of these options and go to that, you can always navigate to that particular option. Apart from menu bar, we also have a standard toolbar where again, some operations such as file opening or new file, creating a new file altogether by deleting previous, opening a new file, saving a file or importing or exporting data, then setting up user and then views or handling of the model.

So there are many options available on the standard toolbar as well. And then we'll be looking at them one by one very soon. After that, there is browser area. So model browser, solver browser, connector browser, entity state browser, there are many browsers available. And depending on what is the kind of operation that we are doing, many of these browsers come very useful. Apart from that, then we come to the main area, the main graphics area. This is where you'll always see your model being opened. And or if you do the fit skin operation, it will try to fit your model within this window, particular window itself. Apart from that, we have panel area and sub panels. So there are many pages of menu pages in the panel.

So such as if you try and match the menu bar options with this, there is geometry panel or geometry menu, then under mesh, we will have 1D, 2D, 3D over here, such is the case with this. And after that, there is connectors, material properties and such things, which are again available through these options. So when I go to new page, I get to see a new panel. And inside that, I have so many features available right at one click. And then I can access the features and do some operation. After or below the below the panel area, there is command window. So commands from command.tcl or such things.

If I want to run them in HyperWorks, just to repeat some operation, I can paste those commands over here and hit enter. And that command will run in HyperWorks as easy as that. So if any of you have handled any other CAD software before, such as AutoCAD, it also has a command prompt itself and you can do some programming on that front. Similarly, this command window can be used for programming in HyperWorks or HyperMesh. So another very useful feature is just that it may not be useful for our sessions right now.

Apart from that, right below the model, all the browsers, there is entity editor. So this entity editor comes very handy while creating properties or assigning properties to your components or any kind of materials to your components or creation of many entities, preprocessing entities. So this entity editor always comes very handy. So this is all about the graphical user interface of HyperMesh. So in this graphical user interface, there are many toolbars available. We'll be looking at them one by one now. So first of all, we have a title bar. It is telling us what product is active. So title bar, it is first of all telling me if there is any particular file open in my current session.

So untitled means there is no file currently open. Even if you import some CAD or CA data, preexisting data, it will show untitled until you save a particular file in HM format that is HyperMesh format. It won't show any name over here. After that, I see the application that is open. So in my case, because it is student edition, I see that HyperWorks 2017 student edition name is there. And right after that, I see the solver name. Now this Opti struct is the solver currently set in this particular window.

But if I go to my own window, which I opened right in front of you, I see a radius as the solver currently set. And after that, the version of radius, if there is any such information to be applied in those cases, so that will be available over here. After that, we have menu bar. And menu bar obviously will have a lot of options for a user to explore and play with. So typically, now since I have already opened a model over here using open button from standard bar, using menu bar, I'll try to show you a few more things.

So I can create a new session, new model and simply go to new session. So this delete model and all those operation must have happened in the TCL file again. And if I open the same file. So apart from that, it has many import or export operations available on loading few user profiles, user profile corresponds to the solver in the student edition. Only radio solvers can be or only alter solvers can be accessed or their profiles can be accessed. But apart from that, there are other industrial solvers that can be pre processed in HyperMesh.

So out of these four options, well, I have selected radius for now. I'll hit OK. And then I can always check or uncheck this option. If I check this, every time I start a new session of HyperMesh, I'll get to see this window right in front. First of all, and once I select some profile then and hit OK, then it will go and next time it will open again. So unless I uncheck this, it will keep happening. So for now, I'll keep this as always show up at startup.

It is a better idea to do. When I select any profile, it tries to reapply that particular profile in the screen. So because it was already radius and I selected radius again, it won't make any change over here. So after file, there is edit where we can cut, copy and paste some stuff just like any common window application. In view, I have many browsers options available in HyperMesh. As I mentioned just now, there are many browsers available. The mask model, there are browsers which are already switched on and then there are browsers which are hidden which can be switched on once I want them.

Until then, I'll just keep them switched off because they are not required right now. After that, we have collectors menu where we can organize our model. We can rename, renumber few things, reorder few things. I can change the color from here and then organization. So if I want to move few elements or few surfaces from one component to another just for the sake of understanding, I can do that from here. I can split apart into two different parts or two different components. After that, all the geometry, mesh, connector operations available from here.

Creation of material, assigning material to some particular components and then renumbering them is also available. As I mentioned earlier, each and every operation that is possible in HyperMesh is available from here. So while doing this, very important option from the menu bar is the last two operations. that if it were in a second, a normal absolutamente bad, like a lot ofnaturalica rygots. We use it right now. Just go using to the wings for that essentially bias, like best you can do. From application, you can switch to different application inside HyperWorks Suite.

So I see in Student Edition, there is Hyper Crash available, Hyper Study available as a separate application. But from that, HyperMath can be switched on too. And inside Help, this HyperWorks Help Home will take me to the HyperWorks Help Home page. So once I click on that, in the default browser, it will open a window. So my Firefox default browser will be showing me the HyperWorks Help Home page. Apart from that, this particular option will tell me about what version I have right now on my machine.

So if you have an industrial license or any other type of license in such cases, you get to see more details here. So HyperMesh is of this version, HyperView is of this particular version, all such cases, such as this. So right now, I see that the version is pretty consistent and all those applications. So right now, their names are as per the application file name or executed file name. But from this, I can tell you that this HMOpngl corresponds to HyperMesh, HyperStudy, then batch versions of them.

Apart from that, HyperWorksession or HyperView Processor Session. So all those, they have this particular version right now. So the page which we opened just now, HyperWorks Help Home, I can always go to HyperMesh button. And now this will give me access to all the features in HyperMesh. So I can literally go to index and find any feature that I want by just by typing. And I can search for it using a search algorithm. And then while searching it, I need to check HyperMesh for a better search.

It focuses its search on that particular application. All right. So right after that, we have standard toolbar or standard bars. In that, we have file operation bars just like file menu. All those operations are available at one single click. I can export whatever is on my screen to a presentation or some other operations that is available. And then I can export data to the solver version. So right now I have selected Radius. But apart from that, I can export the same data to OptiStock or CFD solver input files as well.

So because Radius has the solver input file as .rad, 000.rad, that is why any name that I type here, while exporting, it will store that with that particular version. Apart from that, I have various versions that can be written out. So I can access those versions while exporting. And then there are other options available. So in export, then I can have different options available, exporting everything or exporting any particular or selected entities or exporting only what is displayed on the screen. So just like how the name suggests on the screen.

Apart from that, the HyperMesh commands, many other things. So, well, apart from that, again, I have the user profile button over here at single click. Loading results, something that is not directly useful right now in our sessions, but something that a CAE analyst or CA engineer will definitely use in their day-to-day life. So another important feature, if you want to import a deformed result in your FEA model and then work on that, then this load result comes very handy. Now, this button in Student Edition or HyperWorks session gives me an option to split the window into two and I can model different models over here.

I can open a different model over here and work on it while this model is still open. So usually a good case for comparing two models at a time. Right now I'll stick to one. But these operations are available only based on once I spread the page. Image capture. So there is no ambiguity. The name itself suggests what exactly it is going to do. Save image. It will save your screenshot images to wherever you want. And this saving can happen by window or by some selected area selected by you.

And then you can also record videos of whatever you are doing on your screen. And that can be shared with your colleagues to follow the similar steps. So it has many applications. Now, before we go to these options over here. So, first of all, let me also show you the mouse settings for HyperMesh. So while we were discussing this and we just now selected, checked how mesh or all the menu bars are coming into picture and how they are giving us access to many features in HyperMesh.

The mouse controls are pretty easy once we get used to it. Until then, it can be an exercise for everyone. So left mouse button is always used for some selection and the right click button or right mouse button will always be used for some deselection. But in general, while you are not deselecting or selecting anything, if you want to rotate the model or drag the model somewhere or you want to pan the model, then in such cases, these mouse settings come into picture. Middle click or middle roller button in many mouse cases now.

That is mainly used for zooming options, fitting everything on the screen. So middle button is always associated with zoom or fitting everything on one particular on screen. Left button. So using left button, I can rotate the model the way I want. It's just that the control button needs to be pressed all the time. Right click, whereas it will help me pan the model all over the place. Just to orient the way I want. So this was about the mouse settings. And now we'll go back to the toolbar.

So we have already seen what standard toolbar does. It helps us with file management. And now we'll look at the standard views toolbar. So the standard views toolbar, it starts with fit and previous view button. So right now I'm at this particular view. If I hit previous view, it will take me to one step before. So right now it is toggling between only two steps because that is what its capacity is for saving any particular view. Now I'm at this. If I hit back, you see that every moment on screen is being captured.

So this comes from the command.tcl and when I press it here, it looks for what was the last moment that was done by user it can get back to. So that way it tries to get back to our previous view. Accidentally, we may change the view sometimes and it can be very tricky to get back to the view by proper zooming in and zooming out all the time. After that, we have views. So in this particular corner in the graphics area, we have the global coordinate system, a global system.

So all these buttons are in reference to the global system. When I hit Y X button, I'll get the X Y top plane view. If I click on that again, nothing will happen. It will just keep on Y X or X Y view. Apart from that, there are other views available. And then there is isometric view in the end. Zooming in buttons. And then there are many options available to zoom in or zoom out. Whether you click them or not. That is up to every user.

There is different level of comfort zone for every user. Some users like to access these buttons because they are one click single short buttons. Some users like using the keyboard shortcuts for all of these. So keyboard shortcuts is something that I'll be telling you very soon in today's session. Apart from that, I have rotation, the same thing that I do with my mouse, using my mouse, right click and left click. That can be done in HyperMesh with single click. The pan operation or rotation operation. And then clockwise or anti-clockwise rotations.

Whether I click right or left. That way it will change the orientation in that particular direction. So with this, we have pretty much a good knowledge about standard toolbars now. So if you go back to the presentation, after standard toolbars, we have visualization toolbars, which are right below our graphics area. We have display controls right beside visualization toolbar. So we'll try to look at it right now. The display toolbar though, located in a vertical fashion, just besides the graphics area. So let's get back to HyperMesh.

So this guy over here, this is the display toolbar that we have been talking about now. Well, so from the browser area, somewhat some part of this component is pre-meshed. So this display toolbar will help me. It will help me mask few elements. And once I hit mask, it will hide them from this model right now until I switch off the whole component and switch on again. Until then, this will stay masked off. Then there are options available. Once you mask few elements, if you hit reverse, then only those elements will stay and rest of the elements will be hidden.

Then there is adjust and or such option available in here. So, well, if I hit mask and if I say that unmask adjacent, it tries to find the location where there is some element created to that particular, on that particular component. So there are different. Types of clipping methods. Right now I'm using rectangular, but you also have an option to go circular. Depends on the user preference completely. The find button though, it will help me find the entities based on their IDs. Then there is display numbers option. If I select a few nodes and I'll hit display on, then it shows me the node number for that particular thing. So here instead of nodes, if I had the selection as elements or any particular component or any particular entity that can be selected from here and I can display the number for it. Okay. So after displaying the numbers, I can turn off those numbers just by using this button over here. All of is also option. So I'll go and drag this entity back to its original location. So drag and drop option is all possible because of the left mouse click and that will help us move the stuff around.

Apart from this, after search and numbering, I have scale button. On the graphics area, I always see a scale. I can either switch it or switch on. Then there are other options like element handles displaying them or not displaying them. The load handles display and then loading vectors to display. So there are many other entities that can be displayed separately than showing it with the mesh. Because then it creates unnecessary confusion. So until then it is a good idea to hide those entities. And the display fixed points will help me determine what points are necessary to cover during FE analysis. So these points can be switched off here or better graphics performance. But in the long run, when you are trying to create any mesh, it is always a good idea to consider those points on the geometry.

Just that they will understand quicker. All right. So let's get back to the presentation. Apart from display toolbar. Now let's take a look at visualization toolbar. So in hyperworks, the toolbars right over here are the visualization toolbar. So I'll talk only mainly about the first row over here. So the first button comes from what application is on right now in hyperworks student edition. So I have hypermatch session currently going on. That is why the hypermatch symbol over here, but I can switch to hyperview or any other applications available in the software package.

And then these buttons will help me create new component material or material properties or parts. Such things can be created from this particular button. The load cases can be created from here. And then everyone tries to keep many things in the local coordinate system. So system collector will help us create a local coordinate system time and again and it will help us manage our pre-processing analysis or pre-processing model pretty much in a better manner. And then there comes delete options and layer options and many other, but most important here, which is always handy for every student is checking the geometry color mode and mesh color mode.

So geometry mode or the auto will show you whatever the color of that particular component is with this switch to mixed. Now I see that there are green lines, red lines, which have an implication on that particular bracket. All right. So after turning this to mixed where I see green and red lines, they always have a signal influence on the geometry or the way we see geometry. So while doing the geometry cleanup, we'll be looking at that explanation apart from this, well, these buttons over here, they will help me whether I want to see my geometry as a solid or whether I want to see it as a wireframe and inside wireframe. If I have, if I have these lines turned on, then there is green lines coming on, on that particular component as well.

And that creates a lot of ruckus for, for a general user to look at things. So always be careful while selecting these options. So solid geometry is good enough for me right now. Well, there is no other need right now. This button right here, it will help me set the transparency. So this component, the yellow component, if you want to make it more transparent and components, I select that particular component. I'll hit yes, select, and then I can set the transparency level. So this almost goes to other extreme. It is good idea to keep it in between and make use of it nicely.

After that, I can change the part, the coloring of the, all the objects that are on the screen by, by different criteria. So by component is the one which is switched on right now. But I can not do the coloring by, based on materials or element properties. So those can be handy when, when you are suspecting some issue with one of those quantities in a huge assembly. Well, you can put that property or color with, or component with one particular color. And then when you go by property, you will get that team in that particular color itself.

Now this was for the geometry, the CAD geometry. Only CAD geometry was getting affected. Now when I go here, this is about the mesh geometry. So if I click this until I hit return, there is nothing else that I can turn or switch on. So this particular button over here, it will convert the mesh into a wireframe element as a skin. So when you check here, this option will tell you if there is anything fishy happening behind, behind the layer of this. So that is why wireframe can be easy or easy to maintain.

Apart from that, we have another option to switch the shaded geometry as, as the way to go. Shaded geometry is also very useful feature. And then the element representation for 1D. So this will help me or allow me to change my element formulation and many other things. So this option can come useful in the, in such cases. So apart from this collector toolbar, as we just saw, it will help me create new components, materials, properties or any other load cases and image capture. I have already explained them. So alright with this, now let's take a quick look at shortcuts.

And after that, we'll be looking at browsers and entity editor as well. So commonly used shortcuts or readily available shortcuts are always listed on alter help page and as well as in the keyboard settings. So I can always go to preferences and keyboard settings. And that way I can click on some button and I can understand if there is any particular command that is already assigned to that particular button. So when I press that button myself, I should see that particular profit coming into, into that.

Because people come by and throw it away. As a, as I mentioned earlier, we have different browsers. We have mask browser, which will help me mask few things on the screen. So under geometry, the geometry, which I see over here, if I want to mask that and click this and every geometry or all the surface have been masked. Now, if I hit plus, they will come back all the geometries and on the mesh side, I can always say that hide the mesh, then hide the geometry.

And then I see only three nodes, which are free. So I can clear them all the time. So this mask button or mask option will help me with that. Inside 2Ds or 1Ds, we can have different material or element types based on the solver type. And we can individually hide them or isolate them or switch them on. So that way there are different options available here. Plus obviously means switch on the display of something that is selected over there. Minus would mean hiding the same thing.

And one will mean isolate option. That means all the other entities will be hidden except for the entity that we selected. The model browser, whereas it will list down everything that is there in this particular model. So there is original geometry. And then I have done some task on that. That will learn during the geometry cleanup, which will be right in the next session, if not now. And then the property that is applied. And apart from that, the material. So material is missing. So I'll probably go and create a new material card.

And then try assigning it to that particular component. So even that material will show up here. And based on the available entities, they can be applied on the components. Or a new entity can be created so that it can be applied on the component. So the good thing about model browser, if I try creating some material, I can simply right-click. I have another test for application, which is San Francisco, January, and A根 if it stops, And I can point two, I have one way to improve the area.

Now I can select a particular window and I can get the create option right away. So now I want to create a material. I can hit material, give some name. After model browser, once I close all the panels over here, in the model browser I can see the entity browser. So on the steel, I can select different types and then I can probably select all the density values and elasticity values and the new values from here. For that, there should be some unit system and there are many other aspects that need to be checked before applying any material properties. So always be careful while creating this and with the unit system of this property. Whereas, I see what property and material has been applied. So I can click, I can hit create to create a new material or I can click on this to select some material that is already defined in the model. So that way I can apply that particular material.

So after this, there is browsers available depending on what is the requirement. So in our case, mask and model browser happen to be the most useful ones. And when we go to model browser, the entity browser which pops up for every selected entity, that will help us a lot in deciding what properties or what numbers should go in here. So for steel, so for steel, whatever values I want to apply, if I want to change the material card for that steel. So that can be done from here.

Apart from that, we have different panels available and the pages. So the geometry editing is something that we'll be learning in the next class. During that class, we'll be using many of these options readily available. And that way we can access all those features with one click away or with single click. Apart from that, the 1D panel will suggest everything about 1D connection, the springs or the welds, whatever is modeled as 1D element that can be created from here from this particular panel. Same thing for 2D mesh, anything right from scratch for 2D mesh that can be created over here.

And different operations for creating 2D mesh are also available over here and that can be used effectively. So that way 3D mesh analysis, tools, those are the general options available. Well, tools will have many different specialized operations to be operated on geometry or your mesh. So such as one is reflect or scaling up or scaling down something. So such options are always useful and they always come in the tool page. They can always be accessed from here as well. So in the tools, I'll get the same options available in this case, in this drop down list.

So with this, we are pretty much familiar now with the Hyperbox user interface. So next session when we come back, we'll be looking at opening a particular geometry and then cleaning up so that it can be used for further processing. Right in this process, we are going to start doing this particular stage where we first take the CAD data and clean the geometry and then mesh it or create a finite model out of this. So that will be doing right in the next session. So meanwhile, I want to show some shortcuts.

So these are another set of handy shortcuts. If I want to rotate by increments or if I want to simply zoom in or zoom out, there are many buttons available. And then the usual letters will also have some command assigned to it depending on the availability or if you haven't changed anything in the key command mappings. So a good idea or general idea for any newcomer or beginner in HyperMesh is to print this particular screenshot and simply post it on your desk so that you keep looking at the shortcuts all the time.

And over the time, you can get used to them. Once you know the shortcuts, many things become easy, such as now when I press F10, I simply go to this particular panel and then do the operations on my on my model. But apart from that, the other way of doing that is go to some menu menu bar option or go to some panel on the screen and then look for this particular option. And then you can open that. But well, shortcut will always have its own advantages.

So definitely, I recommend printing this page and printing the screenshot or simply search for Hyperbox shortcuts or HyperMesh shortcuts and you'll get to see this whole list. So with this, we have completed the discussion on graphical user interface of Hyperworks. See you in the next session. Until then, thank you and goodbye.

Part 3

Hello everyone. So in last class, we concluded in the HyperMesh introductory session and we were discussing about HyperMesh shortcuts. As I suggested, a beginner should definitely print this page or print this slide and paste it on the wall along the computer so that this can always be referred. So with many buttons on the keyboard, there are many options available and now one by one as we explore some geometry editing features or meshing features, we'll see how useful are these. Apart from that, there are other shortcuts which can help us control the model in the user workspace in the live session.

Now we'll move to our next topic that is geometry editing. So geometry editing, it starts with always the most simplest option or most simplest method. We always start with geometry import. Now HyperMesh is mainly an FEA tool. It is mainly used for FEA preprocessing. So everything before geometry preprocessing such as generating the CAD, all those steps usually happen in a CAD software. And then there are N number of CAD software in the market. So there has to be one single way or there has to be some common surface or file formats that can be used by any of those software and such is the case with HyperMesh.

So with HyperMesh, you can always import STEP files or IGS files and then it also connects with some of the popular CAD softwares and then it can directly import the CAD software files as it is or as they are then converting them into IGS or STEP. So yeah, there are multiple ways or multiple options available on that front, but everything always starts from importing the geometry to HyperMesh and after importing the geometry to HyperMesh, obviously there are no, there is no history or any assembly feature available on that particular geometry, but only a dead geometry, a dead surface will be available in HyperMesh.

So we'll follow these steps and we'll try importing a geometry and we'll see how it goes. So before that, or apart from importing a geometry, if you have saved a previously imported geometry as an HM file, then that can be opened as well. And then HyperMesh file, that is HM file, it can store geometry as well as mesh. So it has dual feature or dual advantages. So if you have imported geometry once, make sure that you save it in HM format so that you save a lot of time when you are trying to open the same geometry.

And then while geometry editing, the one very important term that will keep getting used is topology. Topology, the topology, the meaning of it is only how the surfaces on that particular geometry are connected to each other, to adjacent surfaces of the same component. So that connectivity is controlled by some common surface ages. So in HyperMesh, there is a color coding available for that. So just by looking at our geometry, you can definitely know if those surfaces are properly connected or not. And if they are not connected, what could be the reason for them not being properly connected?

Then, so obviously if a surface is connected with more than one surface or if it is associated with more than one surface, then all these surfaces are considered to be connected together. And then those ages, as I said, depending on the color coding or in fact, the color coding has come from these categories. The surface ages can be categorized as by the number of surfaces they are connecting. So if there is an age which connects to no other surface, but only to one particular surface, then it can be red.

The line which connects two different surfaces, it can be green. And then the line which connects three surfaces can be yellow. So we'll see them as we go along by opening the geometries. So before that, in the previous session, we were discussing about the visualization toolbar. So in the visualization toolbar, the very first button, it gives us an option about geometry visualization. So in this case, while geometry editing stage, that becomes very useful. So by default, many times it is set to auto, which shows the component by its color and the edges of the components are also shown with the same color.

But now in this case, because we want to use that color coding on the geometry, we want to switch it to mixed or 2D topology or bi-topo. So topology being the keyword, we need to switch to any of these available options. And that way, we'll have a better vision of our geometry. Apart from that, depending on what is the current operation going on, we'll always have to switch between wireframe geometry and solid geometry. And then transparency can be managed as well. But if you go on the surfaces one by one, then maybe transparency may not be needed.

So let's quickly switch to HyperMesh. Now, I have this previous file, which I have cleaned. And I'll try importing some mesh or some geometry. So I have different options over here. I have IGS or I have SOLIDWORKS. That means SOLIDWORKS file that can be directly imported. I've also seen Katia files can be imported from the full version. Now for student edition, these are the three options available for now. So I'll stick to IGS and then I'll open an IGS just for the practice purpose. So now I have imported a geometry and now because I have already switched to mixed, if I keep it auto, then I see that the color of the edges has changed.

And here that is because the color of the component is yellow. So if I switch to SolidGeometry or ShadedGeometry, I see that the component is actually in yellow and those edges now are shown in black. But if I switch to mixed or by 2D topo, then I see different color edges. Now we can ignore this yellow line, which is just resting in the, lying on the surfaces. It doesn't have any role in the geometry definition. So to delete those lines, first of all, by topology editing or by geometry editing, what it means is we get rid of unwanted features and keep only what is required.

And that makes the meshing process or preprocessing process very easy, very easy to do and very easy to understand. And then very easy to correct if there is something wrong going on with that. So to delete these lines, which are not required, first of all, I can always go to the delete option over here on the, just besides visualizing toolbar. So if I click on that, I get the option over here. Apart from that, the shortcuts in HyperMesh, F2, that will lead me to the same option.

Now here I have to select an entry type and then I have to select a selection method. So first of all, I will select the entry type. Right now I want to delete lines. So I will select lines, which was third option. Note that the mouse cursor changes its position when you click on that button. Then I click lines and then I can click on lines over here. So when I'm selecting the turn white, while doing this, make sure that you are selecting the correct line.

Otherwise, you may end up deleting the wrong entities in your model. So I have clicked one by one. But apart from that, now to deselect all of these, I can either right click on that or I can simply press shift and create a box by using my right click around that. That will deselect everything. Or I can simply press this button to clear my selection. So to select the lines, I can either select them one by one by mouse click or I can hit displayed, which will select each and every line that is switched on on current display.

So I'll hit displayed and it selects all the lines. But these lines, which are on the surfaces, those are associated with the surfaces. So those won't be deleted. But the free lines, those will be definitely deleted in this case. Apart from that, pressing this button again, resetting my selection. I can hit button all and that will select each and every line that exists in this particular session. Again, there are all the lines selected, but only these lines will get deleted because they are not associated with anything.

So I'll delete. And I see those lines getting cleared. I'll hit return so that I get out. I can always press escape as well to get out of some particular feature. So now we have a particular geometry and now we want to see how it can be edited. Now, as I see, first of all, there is some red line coming here. Red line means always that there is a free edge to this particular surface or this particular shaded surface. That means there is some connection issue between this surface and this surface at this particular edge or this patch of the edge.

So we need to correct that before moving forward. To move forward, first of all, we always use, most of the times, the most important or commonly used feature is quick edit. Apart from that, there is edge edit, point edit, but the key features from these two options have been merged in quick edit option as well. So you go to quick edit and now you have many options available. So off-top that, there is surface to surface, split surface by using nodes, node to node line that can be used.

So the application, The application for this can be if there is a surface and I want to simply split that surface into two parts. So now using this green line, the green line is the one which connects two surfaces all the time. So using this green line, I can see that these two surfaces are getting split on the upper phase only. While from the lower phase, there is no such line. So this green line only splits the upper surface and not the lower surface. Apart from that, split surface using a line and that line can be originated from a node and it can be perpendicular to the line opposite to it.

So I can click this node which can be randomly created and for line, I'll select this and it will create a perfect perpendicular to that line. So by perfect, so that becomes a perpendicular naturally. So I'll hit reject and that line will get cleared. I'll hit reject again and even this line will get cleared. Apart from that, with the latest features in HyperMesh, even the undo button has come active. In the old versions, this undo button used not to be there and that is why the important feature, it can be used all the time.

And it can make your life easier if you end up doing something wrong or end up following a wrong step. So apart from that, there is washer surface, now we'll look at washer surface only after, after certain stage over here. Then unsplit surface, that also is a good feature to do while, while using middle surfaces, that is right after using washer surface. Toggle edge, now toggle edge will do another useful operation. Now toggle edge, if I have a surface which has been, so this fillet and this particular surface has been split using this particular line.

And that is why, because it is connecting these two surfaces, it has turned green. Now if I click on that line, using toggle edge, it gets toggled. What that means is, now this surface and this surface are no more separate, but they are one and the same surface. And if I untoggle this, by using right click or by using reject, I'll get to see that line again and those surfaces are split again. So this toggle line feature can be also very useful when we extract mid-surface.

Then we have filler surface. Filler surface, well, it can be used to fill up holes or something like that, or fill up elongated holes or fill up any geometry or irregularities. And apart from that, delete surface, as the word suggests, it has simply deleted the surface, which I just selected. If I hit reject, that surface will come back. So, very risky feature over here. Usually, it doesn't have to be used if you want to delete the surface. You can simply make a habit of going to the delete panel and delete it than going from the surface edit panel.

Apart from that, there are add or remove point feature, which are always pretty useful. So, I can add points on a random line or random surface, and that will be perfectly fine. Adding point on line is also similar feature. But, one other feature is, after adding the points, now, suppose I want to remove those points. So, while, by clicking, by doing a left click, I added those lines, or I added those points on those lines. By using right click, I can, I can always hide those points or suppress those points.

So, these points, essentially, they come or they define the surface. So, in a geometry feature, or in a topology feature, the shaded region, which we see on the screen, is called the surface. I will switch to mixed representation. So, well, the shaded region is the surface, and what constitutes a surface are its edges and its corner points. So, the corner points over here, they are defining the surface, which is irregularly shaped, irregularly shaped, if we really look at it in depth. So, these points can be switched on or off using this particular button over here.

So, many times, if you are loading a huge structural geometry, and it has lot many geometry features, so those features will always be full of many points. And if you try to switch on all the geometries at a time, it can be a huge or heavy graphical operation. In that case, if you have this button deactivated, that is not to show points, this operation can happen much faster about showing all the surfaces in the file. So, I will switch on the points for now, just to get a better understanding.

So, apart from these points and edges, what constitutes, what is constituted between them is the surface, which we are seeing on the screen. So, after this, in the geometry panel, after quick edit, we have edge edit. So, in the quick edit, we have seen all these features until now, adding filler surfaces, or toggling lines, or deleting surfaces, and then adjusting or setting density, adding or replacing points, and then there are other features. So, well, so, the most commonly used features out of these are add or remove points, or replace points, or filler surface, or washer split, and then splitting the surface into multiple parts.

So, those become very important features, important entities or options to be used. Toggle edge is another very useful feature. Now, while toggling edge, one important thing that needs to be taken care of is the tolerance over here. So, this tolerance, if it is very small, then usually it might give us an error. Now, for an example, we have a couple of red lines over here. That means there is a free edge on both the surfaces. Why don't we connect those surfaces together using that edge itself?

So, if I try to toggle that line, what it will do is it will try to connect these two red lines, and make them into one. So, as of now, I'll hit toggle line, and then I'll click this feature. Now, it will ask me the gap is 0.226858, that is in mm. Do you still wish to toggle? If I hit no, it won't toggle that. And it asks that warning because tolerance for toggling edges is also a prescribed value. If I go higher than that, I might be able to toggle this right away.

And now I see that it has been toggled, and I see those two red lines combining into one green line. And that way, I have a good connection between this surface and this surface. Now, we'll go to the next feature, age edit. But before that, let's quickly look at what exactly is our topology repair operation. So, in topology repair, apart from editing a geometry, the most important thing that needs to be done is creating the mid-surface. So, well, shell elements or a surface with constant thickness will most probably be represented using shell elements, that is 2D elements.

So, these 2D elements are always assigned thickness, which could be constant or depending on the component definition or the way it's defined for that particular solver. Solver, a component can have multiple thicknesses or a component may have to be split into multiple subcomponents to assign separate thickness to them. So, the main point here is that a single component meshed in shells will always have one single thickness value. Now, when you have a constant thickness, and if you can represent that particular component using shell mesh, then to create that shell mesh, you always need a mid-surface, a single surface, which is a median of the two boundary surfaces.

And that is why creating the mid-surface happens to be a very important stage. Now, when I go to HyperMesh, I have this particular file, or this particular geometry opened over here. First of all, before creating an image surface, the important step that I'll follow is that I'll delete all the solids, enclosed solids within these surfaces, if there are any. So, in the entity selection, I'll go to solids, I'll hit. If there is any solid in the model at all. The message window or message bar over here says no change because totally selected is zero. That means there is no solid in this geometry at all. And that way we don't have to worry about it.

So after quick edit, we are going to look at mid-surface. We have auto mid-surface operation and then there are many extraction options and it gives us maximum thickness for which we can find a mid-surface because again mid-surface technology and using a shell mesh using on a mid-surface is also can be a bad modeling practices in some of the load cases. So obviously there has to be some limit on representing or what do we call approximating a surface. So the thickness should not exceed 6 mm ever. So right now that is why the maximum thickness ratio here given is 10 and maximum side to thickness ratio or side of the surface that is length of the surface that is minimum length to thickness ratio can be 20 at the most. I'll go to the automatic algorithm and here it only asks for a surface which is a closed solid or closed volume. So I'll select this surface.

That means all my volume is closed already. That means there are no red lines with no free edges. If that is done then I can hit extract mid-surface and in the message window I can see that it's progress as well. So now we have this mid-surface generated. So now we have this mid-surface generated. In the quick edit button if I want to toggle this edge I can simply go to toggle and I can toggle this edge depending on how it was split. After that we will look at unsplit surface feature. So unsplit surface feature what it does is while toggling an edge we still see that there is some imprint of that edge and using that imprint I can simply right click on it and I can get back the edge which was originally there.

Now unsplit surface what it will do is if I click on unsplit and hit that line. In fact if I click unsplit surface and hit on a feature such as hole then it will simply unsplit that. By unsplit what it means is that now this surface has been bounded by this particular hole edge. Now that edge will be suppressed and that surface will flow forward and you will see that hole can be or has been eliminated from the surface. So I'll click reject and I'll get back that hole.

Filler surface does the similar trick. It's just that filler surface will still leave behind an edge for reference so that a user would know what all things have been filled properly. After this we will go to the next feature that is edge edit. Now in edge edit obviously as I mentioned earlier there are toggle edges feature with same cleanup tolerance or similar cleanup tolerance. And this is very important feature that has been already brought into quick edit panel. Then there is replace feature a line can be replaced with something else or some other lines not just something else some other line.

Lines can be equivalence if there is a group of lines which need to meet another group of lines then equivalence can be one way of doing it. And then unsplit or edge fillets are very similar feature. Edge fillet is a feature which can be used to de-feature from a surface. So right now I have a fillet over here. I can simply select the surface and select the fillet which needs to be deleted and it will be suppressed. So after having this much surface can still go to quick edit panel and add or remove all the nodes that are not required.

I can do that by using this box and now I see that only the require points on the geometry are still in the user workspace. Alright, apart from that there is point edit option. I can add a point on one particular line. So at cursor it will add a point. I'll hit reject. Then on the edge such as this, I have number of points set to be 5. Now if I click on this edge, I'll see that there are 5 equidistance points created between bound to bound for that particular edge.

And then there is surface and nodes another feature on that front. But most commonly features will be the first two. Creating a point quickly at whatever location we want which has been used traditionally. But using an edge if you want to create equal number of points equidistant points then you can make use of this feature. Then suppress, replace or release features are very similar to what we learnt in the quick edit panel. Apart from that project, it will help us project a point to another line which well there could be many reasons to do that and well so this is the way it can be done.

So in general whenever we see any HyperMesh or hyper work software essential thing for everything is that in the first in this panel on the leftmost column or leftmost space we get to see the method at which by which we need to carry out some operation on the screen. And then compute and then this panel it is much like where do you want to apply whatever changes you want to make using this process. So the object for the process or that can be selected over here.

And then the destination component will also help us. So now we already have a mid surface and we have a component that can be meshed. So for this mid surface is pretty much done. Another important feature that I can always use is D feature in fact. So most important or another important feature of that is let's say in this case I don't really want this fillet to be there. I am going to experiment my simulation by using or by removing this fillet and see what happens.

Now to on surface fillet I can go to lines and once I hit remove that feature will be simply removed and you will see a plain edge between the two meeting surfaces. So I hit reject for now and whenever needed we can always use the D feature button. Now surface fillet and edge fillets both can be suppressed in a very simple way or removed from in a simpler way. Here we go. Oh yeah.

Part 4

After getting the mid-surface, the next step that can be done is editing those lines or nodes or surfaces that we have got in mid-surface. So this is one example of mid-surface. There can be many. So I'll open a new example. Now this time, this file can be stored by using this button over here. And then there are multiple options for that, saving session or saving model. Both can suffice. Saving model is usually a better idea because we want to keep the model handy. And the session will be storing many other parts of information, but HyperMesh model happens to be a good enough format for storing the geometry data.

So I can either save the model or I can hit new. Then I can open a model which has been created for you guys. This is essentially the same model, just saved in HM format. So when I open this file, I see that there are a few red lines. As usual, I'll get rid of those red lines by using quick edit panel and toggling those edges using some tolerance. If they're out of tolerance, it will show as an error in this particular box. So in that case, we can increase the tolerance and suppress those edges which are going around.

Now to get a mid-surface for any surface, the most important point is that you have to get a closed surface and that is why you need to first toggle the edges. So to toggle the edges quickly, I'll do the toggling. Now right now, because I had still add or remove feature on, now when I'm trying to click this edge, I'm getting the message current stitch because the cap is too high. In that case, because I still want to connect them together or connect those edges together, I'll put the tolerance value over here as 3.5.

And in the lines, I can now see that there is a proper surface and now it can be used for the meshing. So now, apart from mid-surfacing, in fact, let me quickly toggle those edges. Again, some issue with the gap. I'll get a new value for the gap and then I can make the surface connect with each other. So now I have a closed surface, hopefully. So if I go to mid-surface option, I can select that surface. Hit extract. Now I see the progress over here.

So when I hit mid-surface, I get a new component and it has the mid-surface component out of the current component. Now I have to first hit return button. Then I turn off the geometry and I can keep only middle surface component on. Right-clicking, I can get many options. I can rename that to modified name and apart from that, I can make it current so that whatever changes I make, I'll make it in my version of the surface. Now I see that there are many, many places where we can remove the edges which are not required.

So in this case, I see that this line connecting over here, it may not make much sense highlighting that as a part of AFA study over here. So those lines can be toggled. And that should be perfectly fine with your solver practices as long as it is being captured properly. So after getting the mid-surface, we have a few more features to be explored. So now nodes, node option will help me create nodes anywhere in this particular model. So node can be created with some coordinates given such as if I hit create, so I see that the temporary node or that node has been created at this particular location.

Then if I click on this button as node and then I click on that node and first of all, hit return, then go back to nodes. If I hit create, then it will create a node at the origin, global origin. So from this, key takeaway is that, well, a node can be created using coordinate axis or its XYZ coordinates. And if I click as node and then if I click on some particular node, I'll get to see those coordinates again. In node creation, then there are other ways of doing it.

It is all about the user preferences, how they want their model to be defined. And if that is optimized, then many things can be. So as my next point, this option over here, it will help me create a node at particular point. So if I click this and if I hit create, it will create a node over there on that particular point. Though I don't have any significance for this whole. So right now I'll go to shift F2, which has a shortcut and then I'll hit clear all.

So such all such nodes will be cleared off. Inside nodes, there are other options as well. In this one, as the name suggests, you only have to select an arc and that arc can be selected using some nodes. So this node, now I just created one node over here, but I cleared it. I can still drag my mouse on that edge and I can select that node. Apart from that, node interpolation is another option. I can simply select a line and then I can tell how many nodes I want there.

And it will create those many number of nodes. Let it reject. Node widget will give us an option to associate a node to a surface or moving nodes around using coordinate features such as vector or and many other things in FEA models. Then placing node or remapping nodes. Those are very similar operations. So remapping node, what it will do is. So now this node list, I can say that for line, please, please remap them to this line. And let's see what happens. Apart from that, align node will only align the nodes between the first two selected nodes.

So all nodes can be aligned and then that way many edges can be straightened out. So we'll quickly move to the lines and line shaded. So to create a line for creating any geometrical features. Now, let's say in this case, let's toggle this feature first of all. And then let's get rid of this feature itself. So to create a line, just like nodes, I can simply put coordinates of start and end node. And that will simply give a direction where to create that line. Then this option will allow us to create lines on the go by selecting a pair of nodes.

So let's say if I have this line on selection board or on the selection group, I can hit create and I'll see a line being created over there. The selector has come back to the edge because, well, as the symbol suggests, you can create multiple points there and then I'll hit create. So even I have another row. Then extending a line or offsetting a line from one particular edge. Those are other operations possible using these. And exploring them is very easy. You simply click on that particular picture and you'll see what it is for.

So for now, for example, this is midline. So you just need to, so after creating lines and to select two sets of lines, first of all, what I'll do is I'll select a group of lines and for second line list. All right. After that, other features include creating a line from mesh. So you select a shell mesh or you select some mesh and you select the edge of that mesh and that will create a line. That is another one, not useful right now since we don't have any mesh on the screen.

From there, we can go to surfaces. Now, just like a CAD software, basic surface can be created in many ways. You can enter four points and create a surface between them using this square option. A cylinder, it asks for the bottom center and the normal vector for that particular cylinder. And then the base radius and height. So the first two boxes for locating the cylinder, the base of it and then rest of the features. And then the base radius can be specified and the height can be specified.

So apart from that, other common useful features could be creating a spin or creating or dragging along a vector so that a part gets extruded. And in such cases, so in surfaces, well, this was about only one of the features. So well, creating the surface from mesh is another. So, you have an FEA model and then you select only the nodes and from those nodes it will create a surface and then that surface can be useful for further FEA modeling. So, that's why there is from FE option and I can select elements or I can select nodes and I can that way create this particular surface out of them.

The surface edit feature happens to be one of the very useful features here. This comes up many times after creating a mid surface. We want to trim or we want to edit the surfaces in a way that the geometry turns into the way we want it to. So, as of now what I see is that these points are way too closely located, may not be required to locate them so close. So, for that I'll try to reduce the number of points over here. Now, that means that the surface or the fillet over here may not be properly captured.

But that is fine because as long as I'm able to fit my mesh properly and keep the mesh flow, I should be good. And that's why toggling edges can also be an option now. So, after extracting mid surface, now I'm editing the surface so that I have something to represent my mesh. A surface can be trimmed with nodes or it can be trimmed with a list of nodes. But the most popular options happen to be trimming a surface using a line. So, in here you first select a surface and then you select a free line.

Such are the lines which we deleted earlier while using the delete option. how to delete those three lines. That was one of the things that was deleted. So, we can use those lines to split the surfaces and that could be perfectly fine. Trim with surfaces, another useful feature. A surface can be trimmed with another surface and a surface can also be trimmed with a plane. So, just to get a hang of this plane, what I'll do is I'll select the surface and then I'll try to split it in a trans plane.

So, are in fact surfaces than planes. So, if I want to split the surface using another surface, I'll have to have some surface that is intersecting it in a vertical fashion. Could be a normal or could be some conflicted surfaces. So, in such cases you can use the other surface which is needed for this separation to be different. So, with this I can select a surface that can be used to trim this particular surface. And as we go along and we create or start creating mesh for 2D and 3D components.

We'll keep using these features all time and again and that will give us some good practice to do that. So, for now what I'll try to do is I'll just get this particular bracket ready for meshing. And in that process I'll try to use as many features as possible. So, trimming with surface, then there is untrim option just to undo that option. It is more like an undo for previous 3 options. So, now we have created a surface or mid surface for a particular geometry. And then from this drawing over here and using the distance features that is F4 on the screen, I can select what is the thickness of two nodes or two points.

That thickness says 1.8. So, for a 1.8 mm thickness component, I have got a mid surface. And then if I, so I'll keep both the surfaces switched on at a time. And when I try to read the thickness, it is 1.8 mm. And in the surface edit, I see that offset is not 1.8 but 2 mm. So, I'll keep it 0.9 so that the mid surface can be extracted anywhere. So, surface, now you got major chunk of the surface. I'll hit 0.9 as reference and then I'll, I can show how it offsets that particular screen.

I'll hit reject. I'll see what can be done using different offset values. Ultimately, if you have a mid surface, it's always easy to create the full cut surface out of it by either using meshing or by offsetting surfaces and creating a solid on your own. Apart from that, extending a surface or shrinking a surface are also options available here. So now, in this demo, what I'll try to do is, so in this demo now, I'll try to clean up a few of the surfaces. So, first of all, I'll press F11 and I'll suppress this edge or toggle this edge.

Apart from that, using toggle button, I'll toggle all these lines. Now, if I keep those lines, the element in that region will be very small. And that is something that we don't want right now. If you imagine 2D meshing, it will go until here particular fillet. After that, it becomes flat and then again it becomes circular or it looks like a fillet again. So, the best way to approximate this feature could be toggling all the lines and then let the mesh decide whether it is following the geometry properly or not.

So, in that case, I can stay on the toggle edge button and then I can get rid of unwanted features in that bedroom. So, same is the case with other 2 fillets. I had this where I had suppressed only 2 lines. Now, instead of just 2 lines, I'll try to suppress all of them. Now, I'll go and toggle this edge as well. So, I can simply select the lines by box and toggle from and select toggle mode on my computer or machine. Now, I see that accidentally this edge got toggled.

What I'll do is, I'll right click on that and nothing happens. So, I'll hit unsplit. So, with this approximation, this geometry is pretty much ready for meshing. Apart from that, now you can save this file with some name. But apart from that, the other important or interesting example. Now, in this case, we have seen this example before. We could delete those lines and then we created a mid surface out of it that could be used for further purposes or for the usage. Now, in general, after creating a mid surface, mid surface is first of all cleaned up.

And the cleaned up mid surface will be subject to meshing and only after that, after meshing or focusing on model built. While doing that, getting a mid surface and then editing the surfaces by trimming unwanted lines or toggling unwanted lines and connecting surface becomes essential. So, with this, we can conclude this particular class. In the next class, when we come back, we'll start 2D meshing. So, for 2D meshing, essentially, we'll be starting from geometry stage itself. So, from geometry, we'll quickly get a mid surface and then we'll use surface editing features such as quick edit or edge edit and or de-feature option.

And using that, we'll get the surface ready for 2D meshing and then we'll mesh. So, thank you for joining this class and I'll see you again in the next class. Thank you. Bye.

Part 5

Hello everyone. So after the last lecture on geometry editing, this session will be looking at 2D meshing. The 2D elements are mainly called shell elements in FEA language. Now the shell elements mainly can be of the couple of types. First is four-noded element and other is three-noded element. Four-noded element usually is called a quad element and a three-noded element is usually called a triangular, that is tria element. Shell elements essentially are used to represent a geometry which has uniform thickness and now once you create the shell element on the mid-surface of a geometry, then you can apply a thickness and that thickness gets split into two parts and it gets applied on each side.

So that is how the whole thickness will be represented in your FEA model. So as we see in the picture, we have a uniform thickness cell and there is a quad element created with four nodes at four corners and then the thickness has been applied and that is why we see that there is thickness starting around those nodes. The rule of thumb usually is that if your thickness of the component is up to 6 mm then they are very well can be represented using shell elements but for thickness above 6 mm usually the good practice will be to prefer solid elements over shell elements.

Now the limitation of shell elements. Now the limitation of shell elements is that if you are trying to study the normal stresses that arise because of the forces coming perpendicular to this plane then in that case the normal stresses may not be able to be captured. So in case someone is interested in normal stresses then well you have to stick to solid elements in that case but obviously with this limitation there are many other advantages so for example shell elements because the number of nodes is reduced your simulation is going to be much faster and then with shell elements there are many things such as airbag that can be defined and then there are applications where the things can be defined only only using shell elements so airbag is one thing or composite materials can be another thing then if you are trying to define a layered component such as automotive windshield which is a layered component in those cases as well the shell element become very much useful it just that the properties of those elements can be modified here and there but in the end what you have to mesh is shell element is shell element so a huge application for that and if you remember the shape functions that we discussed in the very first lecture so for shell elements there is a bilinear shell function or shape function so that is why you have a shape function going in for x direction as well as y direction and with that combination the results can be found out all over the space and that way the orientation of the elements such as as if there is any angle or other than 90 degree right now between these two lines then that shape function will be altered a bit but in the end that bilinear shape function is going to help us get the results all over the elements than just on the node so quickly looking at the applications of shell elements obviously in the manufacturing world if you are looking at any die component or component to be stamped in those cases shell elements could have a good application in automotive industry all the exterior surfaces which are which are having uniform thicknesses or variable thicknesses but still within 6mm all those components are usually meshed with shell elements because as you can imagine in a car model the number of elements the mere number of elements and node becomes a headache for the solver to solve and that is why shell elements are being preferred in aerospace again in aerospace again windshields can have that application the rest of the aeros airplane body as well it can have shell elements and then airbag becomes another application so for 2d meshing that is shell meshing we always start with mid surface so you have a complete component and you first have to extract the mid surface from it so to extract the mid surface if you remember the steps from the previous session we first make sure that the geometry is closed that means we just make sure that there are no red edges if you if you switch to mixed mode you should not see any red lines if the geometry is closed so as long as we have a close geometry mid surface can be obtained and using this mid surface will be using auto mesh feature in this session to generate the mesh and after that I can take you through other features which can help you create the mesh in other manners or edit the mesh so following these steps in our simulation or on our in our model so now I have couple of components loaded on my screen I'll switch to mixed topology now I see in this particular geometry now I see in this case in this particular geometry now please notice that while rotating this whole model if I want to completely focus on one particular thing in this model and I want the center of rotation to change to some somewhere else somewhere on this particular geometry what I can do is I'll press control first of all I'll zoom in at some location then press control and then click on that particular point so it will catch that point and now the whole rotation will be above that point alright so now I want to focus particularly on this component I see that there are no red lines red edges so that means it is free of all the open edges whereas for this particular component I see there are lot of issues first of all what I'll do is I'll go to the shaded geometry so that I would know better what is going on with the surface now again I need to change the center of rotation so that this becomes easier to handle so I'll hit control and I'll hit this point and now I see that the center of rotation has shifted to that point meanwhile because this is not needed I'll switch it off and now I have everything set up for this particular geometry now I see that there is no surface between these two now I can hit F11 and then I can hit filler surface to check if the filler surface can do the job for me and as we expected it does the job for us so I know that this fillet is complete here I see that the surface the front surface is extended beyond the fillet now I want this surface to be trimmed over here what I'll do I'll go to geometry go to surface edit trim with surfaces or planes check this surface and select the perpendicular surface with that now once I have selected those I'll hit trim now if it helps at all I see that there is a greenish that has just formed that means one of the surface has been trimmed so I'll hit F2 that is for delete that is same as this and now I have to select the surface entity over here and after that I click on the surface I see that only this portion of the surface has been selected I'll hit delete entity and now I see at least the surface is properly matching now I need to take care of this edge I'll hit F11 again toggle this edge with this particular tolerance let's see if we have luck and yeah there we go it worked apart from that we have other edges now I see this is one of those cases where age just doesn't exist at all so what I can do is I can either replace this point with this point using this replace function or I can replace line by line so both options are available I'll do maybe replace point by point for now I'll replace this point so it asks for points to be replaced and then for retaining I want this particular point and I see that this end of the line has been merged into the one which we want to retain here the same case and now I see in that process this this surface has been closed now I see there is another open surface what I'll do first is I'll take care of this open geometry so I'll go to toggle I'll hit this button I see if there is anything possibility I see that there is a gap of huge number there is some issue with this calculation but I definitely know that the tolerance may be maybe I'll do Higher value of tolerance is required here. I'll hit 0.1 and see if that works. And now it worked. Same thing for this edge. There could be very minute difference between those two edges and that is why they were left open.

Now here I see the yellow lines. Yellow lines indicate that these lines are connecting three different surfaces. That means if this is supposed to be a constant thickness surface, there is nowhere else that this surface could go but except for these lines to join only two surfaces. So I'll hit delete and select this surface. Once this is selected, I hit delete. And now I see after deleting that surface, there was another surface underlying that and that surface in fact is properly connected. So the only free edge remaining is this one. First of all, I come here, I see that there is another open surface probably. Is it? Nope. It is not open surface but it is more like an open edge which has to be equivalenced.

So I see the gap is bigger. I'll have to go for 0.1 again and I'll hit this button. Same thing for this case. Same thing for this case. Very similar thing. So I don't see any free edges now. And now we are ready to extract the mid surface. So we have both the components ready for mid surface. What I'll do is, first of all, I'll go for this component. So I'll hide this blue one to hide. I'll hide the mesh as well as the geometry. Now for yellow component, we already have a geometry which can be mid surfaced. So I'll switch on the mesh inside the geometry and while creating mid surface and click, select the closed surface and then hit extract.

Then there are pre-existing plates. We want to review that. Now I'll go to mid surface again and try hitting that. Now this time I see that there is a mid surface already created. So now I see that there is a new middle surface created. I'll hide the previous component and now I have a surface that can be meshed. So to mids for auto mesh, I'll go to 2D and I'll hit this button. Apart from in 2D, so 2D panel will be mainly responsible for 2D mesh that is shell mesh. So all these options will be pretty useful over the time. So these four first options are more like they want to create a redimed volume such as cones or spheres or torus and using that they want to create something out of it.

So they want to create some FEA model. So they have this option. Apart from that rules, spline, skin, drag, spin, all these options are pretty useful while creating any mesh because it saves our work by a lot. Then for now, rest of the options are not really directly useful, but they would have some application here and there. So let's focus on auto mesh. I'll click auto mesh. I'll go to 2D panel. Then I'll go to auto mesh. So here I have surface selector and then I have element size. So element size, most of the times it is already decided based on what are the requirements of that particular load case.

So if you are running a model for crash analysis simulation, the element size is usually preferred to be higher. Otherwise, in case of durability analysis, the element size needs to be lower. And then in other analysis such as NVH, it doesn't really matter what your size is. What matters is how well you have captured the geometry. So with that, now with 6mm element size, what I'll do is I'll select a surface and I'll hit mesh. Alright, let's get out of break. So once I said auto mesh, I got a primary mesh, which is pretty much messed up. That means there are a few corrections still left to be made.

Okay, that which we will do now. Apart from that, there is an issue with these holes that needs to be cleared. So we see that the mesh is still not clean. So now another set of refining operations for geometry. What I'll do is first of all hit F11 that is check elements panel. And that, I'll hit remove points. So in this case, I see that there are two different points. I'll quickly replace them with each other. So if I hit retain, then this issue gets solved. Similarly, I have another edge. After replacing that point, we have other corner.

Check for just to make sure that all of them are on the same page. When zooming out, there is plus minus button that can be used as well. So apart from that, rest of the corners where I can replace the point quickly to get started ahead. Now only the last corner. So after doing that, now I know that there are no leftover edges that need to be taken care of. After that, I'll simply remove the extra points. Then I will also remove the extra points from the circles. So. In auto mesh, I select surface. Then I select the size. I select the algorithms that Elements should go to surface, of surface component, that is middle surface component for now, if we change the middle surface component name to something else, that then also as long as you have the same geometry in the surface, it will try to generate the same mesh then first order or second order that is something that a user needs to decide usually first order elements suffice second order elements just add a few more nodes in your analysis and that few is not really a small quantity but a second order node can really be very heavy to solve and then keep connectivity now when I submit when I mesh the next surface after this you will see the importance of this particular option I'll hit mesh I see that there is good amount of mesh already generated I'll hit abort again now in this time I'll create a washer surface so for washer surface let's say if you want 3mm washer then after that we simply have these holes and they have a washer surface so is the case for these holes but well for our consideration we'll simply ignore the rest of the holes as far as the washer surface goes we want something in these these two holes so I select both the washer surfaces So I have only two washer surfaces selected. I'll mesh them first with a smaller size of elements. Let's go by three and I'll let mesh. Now whenever we do a washer surface, most important thing is to follow, at least on the inner edges follow, inner set or inner edge, always follow even number.

Just like this case where it is already five, I wonder where is six. So I have this mesh created and now to get rid of these triads on the border, I can either reduce the from here. So here because I have an odd number, I'll make it even again. And I see there is better development. Let's take number six. Same thing over here. Number six. And let's not worry about those triads for now. But in actual case, you have to get rid of these triads. So you can actually match the number of nodes on both sides. And that will get rid of all the, all the unwanted triads.

Once this is done, we just work on combine tool. And I see that those elements have been added together. All right. Now let's try doing an auto mesh on that same surface. I'll hit the mesh size. So now I see because of the washer surface, there is a better consideration. So after that, I can either go for more advanced mid mesh hitting or I can keep it as is by hitting return. So in this case density can be changed for one particular edge. In that case, element density or in fact, moving around will help more.

So now with this, I have a mesh. Now to edit that mesh, I'll go to element selection panel. And now I can select individual portions of this mesh. So for auto mesh, there is a shortcut on the keyboard F12. You press that you can simply land in this particular panel. And after this, I'll go to the fillet surface. So for fillets, I can hit surface, can select the fillet. Now for a fillet to be better captured, it's always a good idea to at least have two elements on your fillet if you have a long distance to go in your model.

So we have shuffled it a bit. And now I see that two element rows have been created. Now with this feature, we have a fillet covered as well as the main surface. Now I also see that if I switch to wireframe geometry, these two surfaces are connected with a green line. Now what made these nodes to continue on the other side with the same flow? So this option right here, keep connectivity works in that way. Instead of keep connectivity, you can hit break connectivity and if I mesh this again, these nodes will not be attached with each other.

So I'll hit keep connectivity and go as per what has been going until now. Now coming back to shaded geometry, I need to get rid of these two big holes. I need to get rid of these holes. So what I can do is I can hit a filler surface. I can select those lines and then I can toggle those lines. Keeping unwanted lines on the surface can land some trouble many times very often than not these days. So now I can select the whole surface. I can hit mesh. I can hit abort.

Now first of all, I'll get rid of these points. So F11 add or remove points and then I'll hit right click and make a box around those that only essential point remains. So every geometry needs at least one point on its edge. In case of a circle because there is no edge, there has to be some point to locate that circle. So now I mesh this using the same practice. Now to improve this mesh flow or like what we can do is it elements software remesh. I have this mesh nicely oriented. Now I see very similar problem over here.

Now in this case, instead of editing the geometry, because this has happened very late, what I'll do is I'll press F3 that is known for equivalence of nodes or node connections. So I'll hit the replace point with this point. So I want this to be replaced by this point. LTS and then I'll switch to wireframe. As per how we created or just remove some of the nodes. Now we see that the geometry has moved little inverts from the edge. Now I'll hit F6 again. From shortcuts you will see that it can be used for combine or splitting a 2D mesh. So I'll hit combine. Now I have one particular element over there.

Same thing on the other side. I'll hit F3. Choose the node to be retained. Choose the node which we want to get rid of. Again hit F6. Check those two. It combines. So this way we have created a mesh on one side of this particular mid surface. Now I see that the other side has been perfectly mirrored. So what I can do is probably I can mirror this whole thing here without worrying about these two holes and that might give us a good way to go.

So now we'll keep this panel open for reflect operation or mirror operation. Apart from that we'll look at other problems. So this is how the auto mesh feature is going to work. Now let's use some other feature to create the same or similar mesh. So there is another feature called ruled. In ruled first of all I'll remove all the unwanted elements or unwanted nodes. Now in this case selecting this one. Sorry about this delay. So I'll select the fillet surface first. Uncheck the main surface and I'll hit mesh. Get one element. Now on this side I'll keep the one element and I'll show you what could be the difference between these two.

So I've kept one element and now I want to create a mesh between or on the surface itself. Now to create that first of all what I'll do is first go to surface and auto mesh and see what do we get. Again a correction needs to be done. So this is the one usual operation of combining those two elements. So once this is done we are pretty much free to to see what I want to do. So... So... Now, reflect these on the other side. Now, in this case, let's say that I don't want to mesh around these holes at all. I just want to ignore these holes. So, one option is to suppress those holes in geometry editing or the other operation is simply skipping them by using a planar surface lines or lines of a planar surface.

So, first of all, it starts with creation of the reference mesh on the fillet. Now, with this being done, I can say that these two holes are not needed for me. What I'll do is, at F11, add or remove points. So, for this, we'll go to 2D, we'll go to ruled. And in this case, we just need to select the right entity. So, here I have a list of nodes that need to be matched. And after that, I have lines. So, in lines, I can select this line, this line, this line, and this line.

I'll hit, I'll hit hash, delete service, I'll hit create.

Part 6

So after going for node list, we'll go for line list and hit create and hit reject, reject, abort. Now somehow the previously created mesh went into the other component. So to keep the same component, I'll hit this button over here for setting current component and I'll go back to middle surface. Now I want to move these elements to our original component. So I'll hit shift F11 that is for organizing the HyperMesh, organize between two components. I'll select elements and I'll hit the destination component as middle surface. That way, this is done. Now I need to grab this mesh again, instead of creating surface, collecting surface mesh, mesh with this.

Direct guidelines. After this mesh, we're trying to rematch this whole thing. What can be done is first of all, go to mixed set all or map as rectangle aligned. So now we need to understand what is the affection of these nodes with the boundary condition. There shouldn't be much problem. Let's just select one. Now I see there is one unwanted element created over here. So in that case, always delete surface. So now we want to delete that. Now because we don't have that constraint anymore, now I can freely auto mesh the rule version of this mesh.

And now this is much better. From that 2D in 2D panel, we have ruled which we just used, then dragging is mainly used for converting a shell into something else such as solid. Or please call me if you need more help on such features. But usually those features will not be very useful. But in this case, let's just open that and see what happens. So there is a way to drag geometries or drag elements. And if you do them, what will happen is now I can select this geometry and drag it along this line. What I'll get in the end is a 3D mesh. So that is why something that is not required right now, that is why not useful for 3D meshing.

Drag is one of the options that is useful and in 3D mesh that is separately available. So after creating a mesh, first of all, what I realize is that, okay, these elements are not properly equispaced or properly in good shape. That means there must be some quality criteria that means there must be some quality criteria that defines that this kind of element is better than this kind of element or this kind of element. There are a few angles different and then for few elements apart from angle, warpage on many other things. So those are called quality criteria for 2D meshing.

What all that means is that it just measures the deviation from the ideal shape of an element. If our ideal shape of a triangular element is an equilateral triangle, if you have that in any other shape deviating from equilateral triangle, then there is some associated quality criteria that needs to be measured and kept within check for all the elements. Until then, right after meshing, what I can do is hit shift F11 that is smooth and then I can select these particular elements from this surface. Now, I don't need an anchor here, but number of iterations can be higher and I'll hit smooth.

And now I see that some of the mesh got broader, some of the mesh got smaller. That way it is trying to fit the length of 6mm nicely as close as possible. Alright, so we have this and let's quickly do a reflect operation on that. So in tools, I have reflect. Now for reflection, essentially, I need to have for reflecting this particular thing on the other side, I need to have a plane of symmetry. So to get that plane of symmetry, what I can do is shift F4 that will give me the distance between two nodes on the edges.

And if I hit nodes between, first of all, let me clear any previous nodes. Now if I hit F4 and then I hit in between, then I will get to that particular node at the middle. But this division should be based on the geometry and not on the node's placement because node may have differed from the geometry already. That is why in this case, I also realized that these needs to be taken care of. Alright, so what I will do, I will hit F4. That tells me the distance between two points or nodes.

And then I will go for this particular, I will drag my mouse on this particular fillet. I will click on this. And then I will click on the same node on the other side. I will hit node in between. So I have one node over here. Here in this case, the same thing. Hit nodes in between, I get the same type of node. For this, again, you have to be very careful while creating such plane elements or plane nodes because one wrong input can mess up many other things.

So I have three nodes to define the plane for symmetry. Now I can, or plane for mirroring something. So I see on the other side, there are few things that are missing. First of all, the number of elements on the border, that is 3 plus 3 plus 3, 9 plus 3, 12. And the number of elements on the boundary is 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12. So it has 12 elements on both sides. That means I have a green signal to directly reflect the fillet as well.

So to reflect, first of all, I will select the entity. I will select elements. And after selecting all of these elements, I only need to select the plane of symmetry. I will hit 1, 2, 3. And after that, go to elements and hit duplicate button in the selection panel. What duplicate will do is, it will copy these elements on the pallet and when we unload it, that copied elements will be moved to the other end. So now I see that the nodes here may not be connected.

What I will do is, I will hit Shift F3, that is for connectivity. I will put a tolerance over here. And then I can select few elements, just to check their connectivity. Randomly has been chosen. Hit preview equivalence. So it tells me that all these nodes don't have a connection directly to them. I will hit equivalence. Now all those nodes are connected. Apart from that, what other features do we want to see? Apart from this, we just saw 2D meshing using auto mesh panel. We saw smooth, we saw reflect.

Now project, translate and move. Those are few options that we can see quickly. And offset and rule mesh. So rule mesh we have seen already. And element quality is something that will follow right after we complete rest of the tools. So project, translate and move. The base is coming too, that is very good. That is what we'll see now and offset as well. So to project an element onto something is that this project option essentially tells us that the things that are going out of geometry control that needs to be brought back to the geometry and that is why you can select this option.

So for right now let's say in my case these nodes have moved to somewhere. In my case I see that hey there is a discrepancy in my model. I see that this should have been the geometry but the mesh is actually deviating from the edge. In that case what I can do is shift F7 that is project. Now there are many options under project. So entities can be defined as well or entities can be different as well. So right now I'm interested in projecting these nodes back onto this red line.

Apart from that elements, lines, points or any of them can be moved or projected using this feature. Now I'll select nodes and then I want to project it to line. So I'll select a line and then I'll hit line normal. Instead of giving a plane of reference from them. So now I see that it has been properly projected. Many times it requires that after doing some manual iterations on the mesh, your mesh may have been deflected from the original surface. So in that case you may have to project your mesh back to the original surface so that it really follows the mid surface path and that way even after applying thickness your component becomes stressable compared to the actual geometry.

So that is what this project option is all about. The next topic for us was translate and move and offset. So the move option is a very interesting one. So until now I have been pressing F4 for checking distance in between these two nodes or any compare nodes. So two nodes distance is 5.596 and that is well and good. It gives me a little idea about what element size it could be considering for time step calculations or something like that. But at least I have an idea that what exactly is the size of the mesh that is going on. Now by the traditional method if I select or create. So just matching the number of elements on the fillet on both sides.

That's why I switch back to one element on the fillet and that way I am trying to manage that. Now last time we did a reflect operation to connect the mesh or to get the mesh on the other side. This time we will use another operation. So now we have mesh on the fillet and we know that this panel and this panel are exactly alike. Now consider that these meshes were not exactly mirrorable. This plane was in some other orientation. In that case this project option or rather position option becomes very useful.

So now we will look at the position command. Position command lets us put a mesh in a different position. You just need the dimension overall dimensions of the surface to be identical. So to use the position command first of all you go to tools position elements. So here I will select the elements. Now the first thing I have to make sure is that there are seven elements on the H fillet before that. Do I have seven elements over here? Yes I do have. So I can perfectly position this mesh on the other side.

So there is other way available. Reflect is also possible. But just to demonstrate the position command I am using a different feature over here. So for positioning I need a reference plane. For reference plane what I can do is I will take the geometry references. So I will take this point. So take that point first I have to do my left click. I have to drag my mouse on this line. If I keep it pressed then this line will turn white. That means this line has been selected. So then you can click on that particular point to select that as a node.

Now I will select the bottom node as the node 2. And then the third bottom or the third node as other bottom. So I have clicked right now. Once I go on the line I see this symbol. Then I leave the click. And then I click on the edge or the corner. Now I have three point nodes selected for the reference plane. Now because I have to position these as a separate entity. I have to first duplicate these elements. So they have to duplicate it into the same original component. That is the yellow component switched on right now in front of us.

Which again is the current component. So not a worry there. Whether you select current or original. And here I have to select very similar points. So I will click this. Okay now I went wrong here in selection. So I have to be very careful at such stages. So I have to click here again. And select this. Now many times what happens is I might while rotating the model I might lose my focus current completely. And you end up switching model to some unknown location. If you press B it will get me back to my previous view.

And that is one more useful feature. Then I can select the other point. Just like on the other side. One, two, three. Just like that. One, two, three. And I'll hit position. And this is how that node that mesh exactly gets oriented over here. So on this side you see that this tria is located on this corner. On this side as well you see that the same tria is located on this corner. The only thing is that these mesh is not connected to the rest of the mesh. So to do that what I have to do is press shift F3 for equivalence or checking equivalence.

I'll select this set of elements. I'll select this set of elements. And then with correct tolerance. I'll hit preview equivalence. It should show me ideally all the nodes on the line. I'll hit equivalence. So now all these nodes are connected. Also apart from that I also need to project these nodes on these lines so that they'll be considered for connections further. So I'll hit shift F7 that is for project that is again here. I'll select these nodes. And for line. And for line. I have this line over here. If you double click on a line. It gets selected multiple times. I'll deselect all of them. And I'll select that line only once. I'll hit. Using a line normal. I'll hit project. And from there. Now I can go to F12. Use a surface.

This. And mesh this. And mesh this. Then I can have one or two. Each element. As an as an how I like it. I'll hit recalculate. Oh. And then I'll hit return. Now here is where we used. Ruled command earlier. Now for ruled what I need it's only the nodes over here and the lines on this side. So. So. We'll use ruled command again. I'll go to 2D. Ruled. List of nodes. So I can select by box. Apart from list of nodes. I can also select the nodes by path. Now that is another very useful way of selecting nodes.

So I just have to select first node. And then if I skip few nodes and select the next. It will try to find all the nodes on the path between those two. Line list. Line list. I have very similar mesh over here. The good thing about ruled is that it is already connected to these nodes over here. Since they have been used as reference. So. I'll have to just remesh this a bit. Hit mesh. I see there is very similar issue just like before. What I'll do is.

Simply hide those or mask those. So F5 is the mask button. Now I need to get rid of this trash element, which is causing all the trouble. If I do F5 again, unmask all. Now I have a better chance at, so using, by pressing F3, I have done a manual equivalence of these nodes. Now I'll go to F12 and select these elements all over again. I'll hit mesh, abort. There's still one more element left over here, which needs to be taken care of. I'll hit mask, mask. So, this seems to be with one of those elements. So, to rectify this issue, we'll have to focus on few of the elements here. Now I see that element separated. Now I can select that by using a window and hit delete.

And then I'll hit F5, for mask and unmask, I'll unmask all. And now I'm ready to remesh them, to get them into a better shape, which is much more like the shape on the other side, which we created earlier. On this side as well. For the same reason, I can simply use a filler surface over here. Then, completely toggle these lines. And then, do multiple surface auto mesh, which can be risky at times. So, there is some issue with this surface itself. And that is why it doesn't allow a proper mesh to be created. So, in such cases, well ruled is always a better idea, but even while remeshing, it could create some issues.

And, that is when you have to take care of it manually. So, this has been fixed. So, here, if I reduce the size of the mesh, this perhaps could get somewhat better. So, after doing this, go for a higher size bit. So, there are different variations of the mesh coming up all the time. I'll do this option. Go to mesh tile. So, right after you mesh anything, if you are not happy with the way mesh flow is going, go to mesh tile. Then, check this as mixed. Set all. Then, map as a rectangle. That is the ideal shape of any quad element.

And then, go for flow align. And, uncheck the size and skew options. And that could simply fix your flow. And, this is much better flow. Same thing we can do here. With size 5, we can see that this problem could be solved. So, I'll have this flow again. Set all. Set all. And now, this flow has improved by a bit. What I'll do, first of all, mesh by 4. And then, remesh by using 5. This time, it doesn't help. So, what I have to do is only select these many.

Then, remesh. And then, mesh tile. Set all. Set all. And, this is how it gets sorted. Now, to get those smoothened, I can hit shift F12. That is smooth option over here. Select all these elements. Hit smooth. And, they become somewhat properly oriented with respect to each other. So, with this, we have completed the mesh for this particular component. We can go for quality criteria in the very next session. That is, quality index. What exactly does that check for? So, we see that there are few elements failing.

We'll see on what parameter does it fail. And, we'll try to look at the meaning of them. So, for now, we have completed our 2D meshing exercise. And, after this, we'll be looking at element quality criteria. So, thank you and see you in the next session. Bye-bye.

Part 7

Hello, everyone. So, continuing further in the 2D meshing example. Now, until now, we have seen using the auto mesh panel after creating the mid-surface and geometrically enough, we use the auto mesh tool to create the primary mesh. And after that, we also saw a few of the mesh creation tools such as reflect project or even rule meshing or position. Apart from that, there are a few more options available such as offset, move, well, smooth, well, it can be very useful as well. And then in the end, we can also look at check element quality. So, for now, we'll continue with the same example. In the previous example, you must have noticed that we had a washer surface over here. Now, in this new mesh, I have deleted that washer surface and recreated the mesh.

Now, I see that the mesh is pretty much regular. Now, the one important thing that about washer surfaces is that we first need to decide whether we want the washer surface or whether we want to capture that hole at all. Or if that hole needs to be captured, then whether it needs a washer surface. If it doesn't need a washer surface, then what kind of treatment it should have. And in case of washer surface, what should be the element length the mesh can have within the washer surface?

So, right now, because there is no washer surface, the mesh looks pretty much regular. But one important thing about all these holes is that for this particular bracket, these holes could be meaning that, well, there are bolts going through this hole. And so, most of the load that comes on this particular bracket will be coming through the bolts because this bolt will be connected to some other component through these holes and whatever happens to the other component will be happening to these components through these holes because the load is getting transmitted through the bolt.

So, obviously, obviously, modeling of these holes becomes very important. But many times, the usual practice followed in the crash analysis industry is that if you have a hole less than diameter up to 5 mm, then those holes can be ignored. If the hole is less than 6 mm, then it can be ignored. If it is more than 6 mm, then there is a definite number of elements that you should have on the washer surface. And after that, if the size of the hole exceeds way above 20 or 30, then there is a good chance that you may not even need a washer surface.

So, why do we need this washer surface at all? Now, in this particular case, because I don't have any washer surface over here, what I'll try to do is I'll delete this mesh in this particular box and I'll try to recreate that using washer surfaces. And then I'll show you how to adjust that mesh with the rest of the components. But in general practice, whenever we have such components, which are so from a middle surface, they have they are expanding in all directions. We should always start from the middle of the component than starting from the end of the component, because when you start from the middle of the component, you have all the edges free and you can edit the mesh the way you want and then you can follow a particular pattern for the mesh to follow.

And in the end, you can get a much better mesh quality. Now for now, well, I'll hit F2 that is for delete or click this button and here elements I'll select by face. So I click on one element and I'll hit by face. So on this face, whatever the number of elements are there, it will capture those. I'll hit delete entity. Now I see there is an empty surface over there. If I turn on the shaded geometry, I see that surface again. I'll turn off the shaded geometry.

Now I see there are only two points defined on these circles. What I'll do first of all is align those points with each other so that I'll have a good mesh flow between these. Now mesh flow is very important quantity. If you have if you have some load coming from this end, you always expect that that load will be carried through the bracket to other end. Now for that load to carry the flow of the elements should be aligned to that particular flow of energy that we are expecting.

That is why our rectangular mesh pattern becomes very handy. Now I see that in this case, we see that if there is anything that any shock wave that enters from this end will travel through this and it will go over this surface and then it will enter the surface through this edge. And then it is expected to leave this surface from this edge and go down. Now when I try to go down, I see that the mesh flow over here is pretty much messed up. So that could be an issue.

So we need to take care of the mesh flow over here. Now remember that in the previous session with the example which we discussed, the mesh has been modified just for this example sake. That is why you see that there is the mesh on this side is pretty much regular, much more like an ideal pattern. On the other side, the mesh is pretty much messed up or not so good. So we need to correct that mesh as well. So now focusing first on the middle surface or the central surface.

First, what we will do is press F11 that is quick edit and go to add or remove points and then add a point which is aligned to the other hole. Same on this case. So now I see those points have been aligned. If there are any other points on these circles, then I'll try to remove those by using my right click. Now with this, I'll stay in this quick edit region. I'll go to washer surface and I'll hit the offset value of 3. Now this is more like the width of that washer ring around the hole.

Now I know very well that this hole could be less than 6mm. In fact, if you want to check what we can do is I'll go for three nodes measurement. I'll hit F4. First of all, I'll go for three noded measurement. I'll drag my mouse click on this line red line so that it selects that line. I'll hit three nodes anywhere on the circle. And now this radius will tell me that the radius of this circle is 2.49. That is 2.5. So diameter of this circle is 5mm.

So obviously it is less than 6mm. Ideally it should be, this hole should be ignored. That means I can simply fill the surface up and unsplit the edge so that this hole never existed. But for meshing practice, we'll consider this hole and then rest of the holes which are in the similar diameter region, we can neglect them already. So here once I create the washer surface, I'll hit F12 and I'll get the mesh size of 5mm now. And then I'll select this surface. Whenever I get this window, always make sure that you click this inside of rectangle.

Then with shift button, I'll try to select by window the surface. Now I see that outer surface has also been selected. So with shift and right click, I'll try to deselect that. Now I see only two washer surfaces have been selected. I'll hit mesh. I see that a very crude shape has been generated. I'll hit even number of elements on the inner edge and similar number or the same exact number on the outer edge. Now this number can be changed by clicking on that number and then dragging your mouse up and down.

Or you can click on that number by right click and left click. Left click will increase the number of elements. Right click will decrease the number of elements. Same thing on this side. I'll make it 6 each on both the edges and it'll get mesh so that I get all the quadrilateral shaped elements. Not exactly quadrilateral, but 4 noded elements all over. Now whenever we create a washer surface, always make sure that you have at least 6 elements. on the washer surface that distributes the load evenly in many directions.

And depending on the diameter, you can go for higher number of elements as well. So after that, I'll select the rest of the surface and I'll hit mesh. Now I see that first of all, this mesh is not connected with the rest of the geometry. Then after that, I see that there is some issue here. There is an edge created over here. So that needs to be taken care of. Apart from that, the mesh flow in the middle region is pretty much messy. We have very good mesh flow in this region except for barring these particular elements which can be rectified.

Even this region is pretty good except for this particular tria. Now the tria elements need to be minimized because tria elements compared to quad elements have extra stiffness. Extra bit of stiffness that creates some issues while calculating the right stiffness of the structure. That is why we select this mesh as is and then we carry out some manual cleanup operations. So I'll go here and then in 2D menu, I'll hit replace button. That is, that can be also reached by hitting F3 on your keyboard. Replace this node with this node is the operation that I'll try to do.

That nullifies that quad element into a tria. Now this can be. Combine together to form a nice quad. Same thing on this side. I'll take this node and put it here. Rather, I'll do the other way. I'll do this node on this line. Then what I'll do is I'll first of all switch off the geometry so there are no more confusions in this region. I'll hit B to get back to the same view. Now I know that only these two corners are left. So I'll hit F3 again, take this node, do the equivalence.

Same thing on other side. Now to quickly zoom to one particular region, I can press my middle click and control. Control and then middle click. So it draws a line. So whatever the length of line I draw, it will try to focus around that particular region. Again, I'll use the same operation. And now we have at least nullified those rogue elements. Now next operation would be to connect these with rest of the geometry. So first of all, we'll do our connectivity operation. So we press shift F3. And that is for connectivity. And I'll select element entity.

I'll select all of these and select all these as well. And select all these as well as this is right over here. No matter how many elements are getting selected below that, it won't matter right now. But in other features, I would rather not select them. So I just need to check for these four edges. So I'll try to select only the elements that are concerned to those four edges. Otherwise, I may end up spoiling rest of the elements accidentally. I'll keep the sum tolerance and then I'll hit preview equivalence. So within this tolerance, I see only these nodes can be connected. I'll hit equivalence. Now again, I'll go for higher value. Let's say 0.5 this time.

Preview equivalence. Most of the nodes are now getting covered. If I equivalence them, there are only very few nodes remaining to be connected. I'll hit one for tolerance. Always start from a lower value and then go to higher value so that there is minimum movement of nodes from here or there. And now to check whether we have any free edge, what I'll do is hit find edges. So ideally, because we have selected this much area, if there was any edge between these two layers, I would have got this red line here. But it seems because there is no edge remaining.

They are all connected well together. So I'll hit delete edges. This is very important. Once I create any edge, first delete that and then hit return because if those edges get saved, then it will mess up your rest of the remesh operations. Now I have a connected mesh. I'll hit F6. That is for combining or splitting elements. Middle click does the job many times while pressing enter. And now as I mentioned earlier, we have very good mesh flow on one side. I just need to smoothen those elements.

What I'll do is first of all, I'll isolate these elements. So I'll hit by face mask reverse. So I only these elements to taken to take care of. Now it is important that I chose to connect those nodes before beforehand because now once I hide those, I don't have to worry about rest of the elements. Because these nodes are already connected to them. As long as these nodes are in place, even those nodes will move with these nodes. I'll hit F12 again for meshing and I'll try to get better mesh localized regions. So now this definitely improves my mesh quality by a lot. The mesh flow now is much more regular. Same is the case on this edge or this side.

I see there are two trials which one is going this way, one is going this way. Now they can be nullified and there we go. Now there is the main challenge. Now please see that we are not touching the moisture surface at all by any chance because if I touch that then even that will be considered for remeshing and this element will be moved out of its shape or originality. So as long as I remesh in the correct zones, I should be getting minimum trias. Now I see there is no tria on this surface at all, which is a great sign.

Then I'll hit shift F2 for smooth, which can also be accessed from 2D smooth and for smooth, I'll select the elements again. Now this time I don't want to select the washer surface elements. I'll select only the rest of the surface iteration can be 10 or 15 or whatever number you want. Don't keep it one because well it is the number of iterations. It will try to smoothen the mesh over here. Now I see the mesh flow otherwise is pretty good. I see that there is one node deviation away from the flow.

So that could be corrected. Same thing over here. So once I hit smooth, those nodes will move in a location where mesh flow is not affected at all or by minimum amount. Then to switch on the rest of the elements, I'll shift, I'll press F5 again, which is for masking or unmasking. The same option can be found out here. Unmask adjacent or mask elements. So if I click on this, I see this option again. I'll hit unmask on and that will switch on the rest of the geometry. Now we have surface with the washer surface and much better mesh flow.

At the same time, we see that this particular hole over here, it has been created with washer surface. But apart from that, there is few complicated element orientations over here. Now, if we see if there is any shockwave entering from this point, it should be traveling straight and very small component of it should be going towards the hole. But whereas in this case, because the triads are messing up the mesh flow, what I need to do is I'll select only localized mesh and I'll hit remesh and it might solve my problems by some means.

Same thing here. I'll try to remesh these elements. In fact, I think they are already in a much better shape, but I'll simply try and see what happens. I see that there are still two triads in the region, same as this. So instead, I'll keep them as is and it's good to see with them more there. Apart from that, there is nothing much to improve in this particular area. On this side, because now this is not reflected but created by a manual method, I'll hit mesh and I see that the triads are getting reduced to take care of this issue that mesh is getting distorted too much.

There we go. So we have much better mesh flow now. Same case on the other side, we have much better mesh flow. Now what about this panel over here? Well, for this, we look at the other panel first. Now in this panel, the mesh is pretty much regular. The geometry has been respected and that's why only one or two triads are left. In this case, however, there are many triads. We need to remove them or minimize them to some number. Once I select the full surface, I see that one triad is going this way, other triad is going this way and that is why there is a good chance that they can be nullified with each other.

So I select only the region that can be controlled by these triads and hit mesh. So those triads which are opposing each other will vanish by themselves. Again, I select another local region. I see there hasn't been any effect as such. And now I see that only triads are getting in the corner and rest of them are pretty much well defined and well oriented. I'll hit shift F2 for smooth and I'll hit smooth for final score, final mesh. All right. On the other side, we already have pretty good mesh flow.

So now let's focus on this particular surface over here. Now this time we have a washer surface. Now let's work with that particular washer surface. Whereas the other hole doesn't have it. So take care of this mesh. We can hit F12 again for auto mesh. Then we can select some localized regions. On the other side of the cobra. So let's say choosing search to anchor up here and leave, so let's apply here, as per truck or We can put some cucumber and remove the Rama chips out the power current points.

The two mulch. And this way now I have corrected the mesh much better and what I just had to do is only select the localized regions for auto mesh or remesh by selecting the element entity and then selecting the localized regions. Now I see that the mesh flow is much better compared to earlier. Just like the other side, the occasional trials are fine but the zero trials are ideal. Alright, so we have cleared off the meshing problems over here. So after that, whenever we create any mesh, we have already looked at some of these operations. Smooth, we just saw what exactly it is doing. It is trying to improve the mesh flow by moving the nodes within the selection phase and not on the border of the selection. Now Reflect has been taken care of already. Translate will help So this will help us move the stuff around. So we can move the nodes or elements in whichever direction we want based on a plane or based on global coordinate system. And then offset is something that will help us in case if someone creates some mesh on the surface of the uniform thickness component, then that can be offset to the mid surface.

And then it can be assigned some thickness value so that it gets proper orientation. And if you go there, you can check that out. Questions should be claimed that we have to check that out. Okay, let us know that this time. Now let us check information. If you'll call this then we can get some information.

Part 8

So after creating this mesh, next question that arises is that the best possible mesh ever in the region. So I see that after using smooth feature, the next option that can be used is offset. Now I'll quickly look at what offset is doing or what move or translate are doing, then we'll move to the next topic. So offset can be done by using element offset. What it does is it can create solid layers over your shell element or it can create simply a shell offset. Now here I'll select all these elements on the face and now if I want to offset them in this direction, I'll hit 2mm and now very important I'll hit duplicate.

duplicate to original component. So the element, the message window will say 120 elements duplicated. I'll hit offset plus and now I see that an element layer has been created which is offset by 2mm. If I check the distance. If I check the distance by using F4, that is exactly 2mm in Z direction. So right now because that is not needed, I'll hit F2 that is delete button, select the first any of the element by attached or by face, I can delete those. So I have the original geometry again.

For move option. For move option, move option can be used by shift F4. Otherwise move option can be found out from the panels. The panels we have translate option which is exactly what move is all about. So in this I can select nodes. So in this I can select nodes or elements. So for now I'll select few nodes. I have selected only these three and then I can give some direction using this selection panel. So this selector will help me select a direction. Magnitude can be 1mm. Now if I hit translate plus, these nodes will move in the direction defined by these two nodes.

I'll hit reject and it will come back to its original. If I hit translate minus, it will go in the opposite direction of what has been prompted by those two nodes. I'll hit reject to come back to original. Apart from that elements can be moved as well. So now this translate option can be accessed by shift F4. Apart from this node system, I can also select global y-axis or z-axis. So in z-axis, let's try moving something. So I can select element or as well as full component can be moved in some direction.

Again, experimenting on the same face. In z-axis, if you want to move that up, I'll have to go in the negative z. So I have selected z-axis here. I'll hit translate minus and we'll see that it has been moved by whatever the distance we have specified on each step. So usually the distance over here or the magnitude is less than 1mm depending on at what precision we want to move the nodes or the entities. So apart from elements, one can move components, lines, surfaces and many other entities just by using the simple options.

Now, if I have a third, well, a general plane in which I want to move these elements in the normal direction to that plane and that plane may not be exactly aligned with the global. So in that case, I can use this three node selector again. And first of all, select whatever I want to move. So I'll come back for elements. Hit here, face. Now in this case, if I want to move these in the plane in normal direction to this, this particular geometry, I'll select this selector and I'll form a plane.

So since I have, I have selected them in the clockwise manner, if I hit translate plus just like a screw, it will go down into the, into that particular plane. So I hit translate plus and it will go down into the plane. I'll hit reject to come back to original. Remember, I can also hit translate minus to come back to original, but then if I hit, keep pressing translate minus, it may go well above that. Translate minus, I have to select the elements again. Translate minus will move the elements up.

So with this, we have seen how to use the opposite as well as move or translate options. So right after this, we'll go for element quality checks. Now element quality checks. Essentially, they haven't decided, you know, with a purpose to make sure that the mesh, which is created in an acceptable form by acceptable form. What usually means is the mesh flow should be proper. That means if there is a shock wave entering the component from one end, it should pass through the whole component in the expected or in an expected direction.

And then it should leave the component in an expected direction. That's why mesh flow is very important. After checking for mesh flow, we'll be looking at element free age. This is very similar operation to checking the free ages in the geometry. We just need to make sure that there are no free ages in our mesh as well. If the mesh is neat is to be intact. What we usually do is shift F3. That is for connectivity between nodes and we select all the elements. So we can select that by box or any which method that we want. Tolerance here can be as low as 0.1. But here, instead of looking for equivalence, we'll be looking for free ages.

Now, ideally, free age should only arise only where we actually have the part boundary. Now, but where here I see that there is a free age, that means there is a node connectivity issue. Now, these are the issues that I need to fix right now so that there is no error in the later stages. So I'll delete this age first of all, and then I'll select those elements and look for equivalence. I'll increase the tolerance by a bit. And then I'll hit the equivalence button. Now, if I check the whole part again and find edges, I should not see any age in between the mesh.

I should see the red edge only at the free edges of the component and delete the edges. So the second stage of our element quality check is done. Next, duplicate nodes and will be simply go to 2D in that we have all various options available. But whereas if you look at the HyperMesh shortcuts, there is one easy option or one easy shortcut for check elements that is F10. And then in here, if you click on duplicates in the command window over here or in the message window, you should say zero duplicates elements were found in case you have any duplicate elements, such as in this case, now I'm artificially creating some duplicate elements here.

So if I select few, if I hit duplicate, original, and then now because of the edges itself, you can see there are some, there is something wrong over there. And I'll select these elements and I'll hit equivalence. Now, for these same four nodes, there are two different elements connecting them together. And that is why this is more like a duplicate element concept. Now I'll hit F10 and I'll hit duplicates. So it will select exactly the portion that has been duplicated. I'll hit save field. The highlighted elements have been placed in the user mark. And now to retrieve these elements, I'll go to my delete option. The elements I'll go and say retrieve.

Now, once I have selected the elements, I can delete. If I want to add more for deletion, I can always use save and then add more and then retrieve and then do this operation further. So I'll hit right now delete the entity. So I see no more elements over there. Just to check the mesh flow, I'll hit the smooth operation again so that the elements are aligned properly. Now, the next option will be after duplicate nodes and elements, we want to look for limit quality criteria. But before that, Now, the next option is like a manual item, which is the minimum.

Now, the next option to the line of content will be available, and you can see the basic image, So let's complete the rest of the four topics and then we'll look at the element quality criteria together. Shell normals. Now for shell normals, well, very common method. So the concept of shell normal is actually if you have a shell, the line or the vector that that is showing the normal direction of the shell is called the shell normal. Now that is mainly used for a post processor understanding.

whether an element is going undergoing any compression or tension. So obviously the normals exist only for the shell elements. So in tools, if I go for normals, I see display 2D elements. I'll hit display. And now because there is color display, the red color means all the normals for these elements have been aligned in the same direction. On the other side, I see blue color. On the other side, I see blue color. The red color means the normals are outward. So I'll switch to vector display with a vector size of 1 and I'll hit display.

Well, perhaps we should increase the size of the vectors. And that is exactly what we are supposed to look for that all the elements, the normal should be directed outwards. So that whenever this particular component is under compression, the negative Z will show me the compression values and positive Z direction will tell me the tension values. Similarly, if if I had few of the elements going with their normals in the other direction, then on the same component, though the rest of the elements are under tension, I would I could see some compression on this particular element.

And that is really a bad result. So that is why it is important for us to have all the normals directed properly to outwards or inwards, whichever way. Usually if if the load interaction for a particular component is from this side, then towards that side are the normals always directed. I don't see any component going inwards or inside this. That is why this is a good enough way for normals to be defined. So I'll hit return and those normals will get hidden again. After that, the next check quick check will be geometry deviation for geometry deviation. As the name suggests, we need to check how much it has deviated from the main geometry or actual geometry.

So mainly we need to make sure that it is properly projected to the surfaces. All the lines or the edges have been properly covered. So I see there was some gap over here. Perhaps this should have been taken care of during the geometry cleanup itself. So this is not even a edge error. Apart from that, overall, this geometry is not deviating from from the lines as much as I see. But in case I see that, let's say these nodes are not really aligned on this line, what I can do is I can hit Shift F7 that is for project button.

I can select all these nodes on the line. Now to select those nodes, either I can select by window or I can clear my selection and go by path. And then I can select one node and click some node in between. So make sure that all the nodes on the way are getting selected. Then I can hit the end node directly. And after that, I can select two line. I can hit the line list that is this line itself and vector. Instead, I'll switch it to line normal.

So these nodes will be projected to that line in the normal direction to line. Once I hit project, I see all the nodes are properly aligned on that particular line. So this becomes another useful operation. Apart from that, for the project option itself, I can also project the elements to surface or nodes to surface. So I'll select this button and select all the nodes by face. So I'll select these nodes and then I'll hit by face button. So all the nodes are getting selected. Then for surface, I'll select this surface. So I have this surface over here and then I'll switch it to surface normal and hit project.

That just make sure that all those nodes on this particular surface are getting projected to that particular surface or all the nodes that are supposed to be on this mid surface plane or mid surface exactly. Those will be nicely projected. Many times during mesh editing or adjusting the mesh quality, we end up causing trouble on the alignment of surface or the nodes on the surface. Similarly, elements can also be selected and projected on the surface. It almost gets us the similar result. Now, here is a typical case where we are differing from the line by quite a lot. Now, in this case, what I can do is first of all to get it here.

I can either hit shift F3 or rather F3 that is for equivalence. Then I can click this node and I want this node to be here. That is why I'll drag my click on this. Once that line gets selected, I'll click on this point. So it gets that it dragged on this this particular line. Similarly, on the other side, I'll try to follow the same operation. Now, in this case, the the matter is little more complicated. Now, I need to have this node over here and this node over here and in between this node needs to be accommodated.

So I'll click this node. And for this line, I'll click this point. And now it has aligned that one for this power node. I'll have to click this point precisely. So that gets aligned. Same thing over here. And now I see the nodes in between are not aligned properly. To align them, I can either use the project button or I see that these nodes have to be aligned exactly between these two nodes. So what I can do, I can hit F7 that is for node operation again. I can move nodes whichever way or I can place nodes somewhere or I can align nodes.

So here what I'll do is I'll select the first node and the last node in between which I want to align all the nodes. And then I can select all the nodes that need that are needed to be aligned between or with on this line form form by these two previously selected nodes. Hit align and everything gets aligned between them. Same thing. So I'll start over on the other edge. I'll select this and this and I want this one to be aligned exactly between them. Now same thing. One and two node.

So with this, we know how to align the nodes between two elements or two places. All right. Same thing can be done on the other side. But well, so we'll let it go over here. Now next step in quality check will be after geometry deviation will delete free or temporary nodes. And then we'll do renumbering and we also want to check element criteria before checking shell normals as well. So I'll delete the free nodes. Now free nodes or free lines are the lines which are lying around unnecessarily.

So if I check the original geometry, I see that there are unnecessary lines lying over here. There are temporary nodes temporary lines need to get rid of that for temporary nodes. There is an option of shift F2. I'll hit clear all and that will clear all those free nodes. After that, I want to clear these lines. So I'll go to normal delete option. That is why pressing F2 or this button over here instead of elements. I want to select lines. I'll select all the lines. Now that will select these lines as well.

But we don't care about those lines because since they are associated with surface, they won't be deleted. And so this is how we clean up our mesh. After that, what remains is renumbering property and material assignment and element quality criteria. So before this stage, we'll quickly look at what is quality criteria. Quality criteria is based on how your ideal mesh shape should be. So if you are ideal, if you look at a quad element, since it is a 400 element, ideal shape will always be a proper square or proper rectangle.

So square is a special case of rectangle. So rectangle is a better word for that. For triangular elements, ideal shape is always an equilateral triangle with whatever size is specified for the elements. So with that in mind, right now, whenever we match... We always see that the elements are not always in the ideal, ideal shape. So there has to be some amount or some quantity that defines how much it is differing from the ideal shape. And to do that, this quality criteria have been defined. Now, what are those quality criterias?

First of all, minimum and maximum length. So the minimum length or maximum length, they can be measured by various algorithms. So, well, for a quad element, such as in this case, well, length of the side is one way of looking at the size of the element. But apart from that, the minimum height of the element, which is obtained by projecting a normal onto the other side, such as in this case, now this node is having a normal on this line at this point and this line at this point. So, well, so minimum of all these eight lengths or all these lines, that will be the minimum length of this particular element. For triangle element, similar fashion goes, you have minimum height of the element as well as minimum side. Whichever is minimum, that will be considered the minimum length of that particular element. Now, there are various algorithms for that. In one of the algorithms, it will only check for the length of the side.

Now, that may not be the most correct algorithm, but for meshing purposes, perhaps that can be followed by some. The one which I recommend is always the minimum normalize height or minimum height where you actually check for all these lengths and then you come up with the minimum length by comparing these. The whole point here is that because all your FEA analysis is all about shockwave and shockwave is going to come from one end and going to pass on the other end. Now, the element size is very critical in many analysis such as if you are looking at an explicit analysis, there is a concept of time step and this time step depends on your element size itself. If you have two less of an element size, your time step is going to drop significantly and that will slow down your whole simulation.

So, unnecessarily, if you don't need to capture such small elements in some particular region, there is no need to then bring down the whole time step of the whole simulation. That is why it is important to check for the minimum length of the elements. Maximum length, well, that criteria can be suppressed many times, but mainly in the durability analysis cases or fatigue analysis cases or even CFD analysis, maximum length can also be a significant criteria because you don't want two nodes to be located far away from each other.

So, so that you can capture the true behavior of the whole component at each and every minute location for crash and noise vibration purposes or such analysis. Well, as long as your mesh flow is good, you should not have too much of an issue, but you should also make sure that your geometry is properly covered. For NVH, you just need nodes everywhere on the body, whether they are triangular elements or quad elements, that matters little less, but in the end, you just need to follow the basic quality criterias.

Aspect ratio, well, so it is actually the ratio of maximum element age length to minimum element age length. That just makes sure that your element is not too much distorted, much more like a trapezium where one side is only parallel to the other one, but the lengths, the difference between lengths can be huge. So, ideal value is one that is much more like a rectangle or square or even for triangle or an equilateral triangle where your minimum length and maximum length are still the same, but acceptable values can be up to five.

So, many times or in many companies, the acceptable value can go up to five. The ideal value though is always one. So, always go for rectangular mesh or proper quadrilateral mesh than elongated quadrilaterals where you might mess up your aspect ratio. Now, obviously, if you have such a distorted element, you can always imagine what could happen to result when the shock wave enters from this end and leaves at here. So, what kind of transition would happen from this node to this node, how the energy will be parsed and that really decides how accurate the FEA simulation is going to be.

So, aspect ratio happens to be another important criteria. After that, another very important criteria comes is warpage. Warpage is actually an angle between the normals of the split of a quad element. So, obviously, because it is a quad element issue for triangular element, this quality criteria doesn't apply. So, warp angle, essentially what it tells me is if one particular node is out of plane with respect to other three nodes, then the extent by which it is out of plane, that gives us an idea of warpage.

So, if this is out of plane, obviously, the triangular plane over here, that will be aligned at an angle to this particular plane on the other side. Now, as long as this angle is within certain limit, it is good. That limit can be 10 degrees or 15 degrees depending on the application and engineer to engineer. So, in general, if you keep all the nodes within a good reach of the plane angle, then usually, warpage can be solved. Now, I'll try to showcase all these problems in our current example.

But meanwhile, we'll first study what exactly are the quality criterias. So, if warpage is clear, let's move to the next one. Now, Jacobian. Well, Jacobian, again, is a scale factor and that comes from the transformation of coordinate system. If you look at the Jacobian definition from your engineering math, Jacobian is all about transformation matrix from one coordinate system to another. But this coordinate system, well, so, essentially, at every instantaneous moment, the element which you have in a global quantity or on a global coordinate system for the elemental calculations using shape function and the Gaussial integration, all those calculations happen at a local level, at a local element level.

So, for that, a local coordinate system is required for each and every element and that is taken care by the solver itself. But while doing that, if a solver is trying to convert a global coordinate system unit to local coordinate system, as long as all the things are at a considerable distance within reach with respect to global, it is good. But if your element has deviated from its ideal shape by a lot, then it messes up the Jacobian or this local coordinate system as well. And that is why Jacobian happens to be another criteria that tells us about how deformed is your element shape currently with respect to the ideal shape.

So, better the shapes of the element, less the analysis time, because less time it takes for the solver to figure out what exactly or how exactly is this local coordinate system going to work like. And that way, it reduces some solution time. Then comes the skewness. Now, in this case, imagine a quad element, which is much more like a parallelogram. Now, it has already deviated from its ideal shape of an ideal square, but now it has a specialty angle. Now, this angle can many times be harmful to your FE analysis.

That is why to keep this angle in check, even for triangular elements, this skewness has been defined. So, the skew is actually, well, it is the angle that is formed by two opposite ends or two opposite sides of a quadrilateral or one line connecting one of the vertex to opposite face and one line connecting two opposite or two adjacent faces for a triangle, such as how it is shown in this particular case. So, the minimum of those will be alpha and that way, the skew will be calculated.

Obviously, skew has to have a reference value and that skew has to be, so this value usually is 60. So, 90 minus a, the typically followed value is 60. That means your limitation on alpha is only 30 degrees. So, at the most, if you stretch this element further, if this angle becomes 30 degrees, beyond that, this element will start failing in quality. Until then, this element is pretty much fine for the skew. Then comes the included angle or interior angles. Just like in this case, for skew, we are looking at the angle between the line joining the opposite two sides.

For included angles or interior angles, for quads and triads, the ideal values in these regions are 20 to 120 or 40 and 140. So, addition of them becomes critical as well. So, with this, we have kind of completed the element quality criterias. Now, we will quickly look at how they arise in our model or how they can be checked in our model. So, to check that, quickly what we do is go to this panel, check elements. This is exactly where we were looking for duplicates earlier.

Now, here, first of all, we have for 2D, we have all these operations available. So, for warpage, we have all these operations available. So, for warpage, we have all these operations available. Let's use the limitation of 10. And I see in the message window that there are zero elements failing. So let me artificially or manually fail some of the elements so that I can show you how can one fix the whoppage. All right. Now I have deviated that geometry by a bit. Let's see how much is the whoppage on that particular element. Now I see that the whoppage value. Now to check the whoppage value quickly, what can be done is if I click on whoppage, it will show me all the elements that have failed. But apart from that, if I click on any element here, it shows me all the values, all the quality values on that particular element. As long as all of these are under control, we should be fine. So let's quickly apply all the limits here. skew 60 as I told just now. Core deviation in selkwish is something that we'll ignore. The length of the element, let's keep the minimum at 3.

maximum length, maximum length, could be 10. Jacobian 0.5. Since we have considered this skew, equivalence skew or area skew can be neglected. The most more commonly followed criteria amongst skew is the normal skew. Then minimum angles for quads and trials that needs to be set as well. So now first of all, we need to get rid of this whoppage. Obviously, we can see that this element is way out of plane for rest of the nodes on that plane or of that element. So this can be solved into two ways. First of all, I can take this node and then take this node and then for the plane or for n1, n2, n3, I can select this. So because I have selected the anticlockwise sense translate plus will take this node up.

just like a nut bolt or screwdriver or how the screw works. So this way I'm bringing it up. But apart from that, I could also see that instead of now in complex situation, now this because we have a plane geometry around, we can say that only this node has deviated. But instead of that, well, I might consider sometimes depending on the orientation that, hey, in fact, these three nodes are in perfect plane. And this node is kind of deviating. So in that case, what I would have to do is first of all, check this and then n1 into n3 using these three and then hit translate plus. So now if I check the whoppage of this element, I see that it is 12.99 degrees. So I need to take care of it a little bit more. If I hit F10 and check for that element, it is now under 10 degrees. So this element will no more be flagged for whoppage.

But in the process, I have messed up with rest of the elements. And in the same process, I have also deviated from my original surface. So you always have to be very careful about what or while moving such nodes around. Many times this whoppage comes because we are using two cores of a mesh. So in some regions, if you have too much whoppage, it could be a better idea to use some finer level of mesh in that region. Those regions could be typically some some curved surface or some fillet. And in those regions, it is always a good practice to allow smaller elements to mesh.

Now one quick way of solving this issue here is I align or I'll put all these nodes on on the surface or to the plane. So to plane and I'll select this select this particular plane over here. No need to select any vector. There is no toggle option for normal. Now this way I have already solved the problem. So this is the this is the thing about whoppage. Now we'll go to check element quality panel aspect ratio. Zero of them have failed for skew against zero length. Well, these elements have failed. Obviously, since they are in the washer, they will fail because then the length of the element is just below three.

So usually just to clear these of such issues, what can be done is you select those nodes and then translate them in some direction. So in one and two, if you give only two nodes or two inputs, it doesn't take the plane, but it takes the vector in which it has been shown. Then you can take a very small quantity here for moving. And now I might see that these elements are no more failing for. In fact, the minimum length right now is 2.498, which is not even on this one, but on some other element. Otherwise, I might have seen or if this these two lengths were the bottleneck, we might have seen a solution to that problem.

Now washer surface and though it is flagging the error on the minimum size. Washer surfaces are something that are very important for any analysis around the holes. So washer surfaces are going to ensure that there is no stress concentration in the region or the forces that are coming through these holes are going to be equally distributed or well distributed around the region. And that will contribute to a good analysis result or correct analysis result. And that is why we have to stick to a good smaller size. So at the most, what we can do is scale these elements outwards.

So if we are right now in with the scale constant, what I can do is first of all, take a center of this. So using F4, three nodes. I'll take the center. And then scale 1.05. Even in y-axis is 1.05. And I'll hit scale plus. Uh oh, the one thing that I missed out on is to select the origin. Now with this origin. Now with this origin, I see that the size has been increased by a good margin. So perhaps I should take a smaller scale just to clear out the issue of the lens.

And now this might just solve my problem. So if I hit F10, look for these. Well, the length has improved by a bit. And now this could be because of the height of the element than anything else. So this way I can actually increase the length of the washer surface elements. This is pretty much standard approach to solving such issues. But many times you can even neglect washer surface issues because those are supposed to be very small elements such as of this level. Well, so when I press F10 apart from that, I see minimum maximum angle Jacobian Jacobian can be sorted very easily as well. If you just remesh in that region, you get a better shaped element and that just that might just work for you.

Well, so there is apart from this using check element option. If you go to 2d and hit quality index, this is another way of HyperMesh for checking your quality. So here you check and uncheck all the criteria that you want to consider and then that will be listed over here for how many elements it is failing for. So you go to page 4 of this and then I click this edit criteria button. So target element size is 5 minimum size should be 3 maximum where we don't care about that right now.

Now here minimum size how to calculate that this algorithm can be selected. Now as we suspected that was selected by minimum height. If we go by minimum normalized height or shortest age, the results will be different. So let's go for minimum height and hit apply. Apart from that, what page needs to be changed. So 10 will be a better value than Jacobian skew. All these angles will hit apply. And now we see some results. Now I see that there is. Now I can either go press F10 and check for this element and then move the nodes manually or I can go back to quality index and then look for this cleanup tools button. Now this cleanup tools is a very helpful tool. Now we see that since it is failing for maximum angle quad we see that this angle over here could be the issue.

So we can go for place node and then we can simply drag this node around so that this quality gets cleared. But while doing that the rest of the elements are getting hurt too. So I'll abort the operation but that operation already gets saved. So I'll go back to quality index and move back to origin position so that this error can be demonstrated again. Now in this case I'll click element optimize and this way it has replaced that node to some other location and now the angle is much better.

So wherever I click that element tries to get in a better position. So if I click element optimize on this, it will try to first match with the line or to which it should be pertaining to. So if I click element optimize on this, it will try to do so with this or we know quality index can work as well for us. It just that element optimize is not always the suggested solution to many problems because element optimize can mess up one element or it can fix one element. But while fixing that one, it can mess up five or six different elements.

So to be aware of this element optimize is a very critical tool or it should be used very carefully while clearing out the element quality. So with this, we have completed all the element quality checks. So right after creating mesh or right after getting a geometry, we first of all clear all the free edges in the solid geometry. Then we move to mid surface. We get a mid surface and then we create mesh using auto mesh after auto mesh. We go for getting a better mesh flow for obvious reasons of having the shockwave pass through the whole component easily as expected manner as possible.

After that we select for element free edge. If there are any free edges or any boundaries or well, unequivalence nodes, then those should be fixed. Any duplicate nodes or elements, they can be deleted as well quickly during quality checks. And after that quality criteria should be considered for all the criteria such as minimum, maximum size, Jacobian, skew, warpage and then interior angles. Shell normals again, as I mentioned, if they are not being taken care of, then they can lead to bad result. Geometry deviation is something all about projecting your mesh back to the original geometry and then we delete free or temporary nodes as well as lines and then we renumber everything.

So for renumbering. So for renumbering, what we can do is go to tool, renumber elements start with 100. So right now, so I have these two elements over here. Once I renumber them start with 100 increment by one element. Now I see that those numbers have changed. Then if I hit this button and hit display off, all off, then I can see that that has been cleared off. All right. So with this, we have completed our 2D meshing exercises. There are more examples and 2D meshing being the most common mesh or most common type of mesh being used in automotive industry as well as aerospace.

It is very, very important for everyone to understand how this 2D mesh works. Now to assign a quick property to this, first of all, I'll move all these elements to its original geometry. So I'll hit organize button. Then destination component here should be original geometry elements as displayed and hit move. And here from the tree, I'll switch on the original geometry and mesh. Now here for original geometry, we only need to assign some material of which is as per the requirement or property. Here thickness value should be assigned.

So apart from this, if you simply create a property, you have to assign a thickness. If you hit close, after assigning the thickness, you can go here in the representation panel. Instead of shell, you can see 2D detailed element representation and that way you can see how well it is capturing the thickness of the component. Alright, so we can conclude this session over here and we'll meet again for 3D meshing exercise. Bye.

Part 9

In this class, we will be looking at 3D meshing in HypoMesh. In 3D meshing, there are many other types, but mainly three types of elements that are used quite popularly. First of them is 4-node linear tetrahedron, that is tetra 4 element, so which looks like a pyramid, but if the pyramid base was a triangular base, then how it would have looked, exactly how this tetra 4 element looks. That is why this tetra has 4 nodes. Well, the good thing is each and every node is connected to every other node on the element.

For a brick element, the next type in 3D elements, the brick element is much more like a brick. It has 8 nodes, so it is a classic 8-noded brick element or hexahedron element. So, brick elements are also called as hex, quite popularly. Apart from that, penta 6 elements, where if you stretch a quad element or 2D element, it will become a 3D element which will become brick. Instead of a quad, if you extrude a triangular element, then whatever you get is called penta element or penta 6 element.

These three types, the creation method for such meshes, it is a little different from 2D meshing, obviously. 4-noded tetrahedron elements have so many methods for creation, whereas for 8-node thick shell or brick elements or penta 6 elements, the methods are a little tricky. So, we will be looking at both of them. Essentially, we start from 2D mesh for all these cases. So, in tetra mesh, the flow goes like we first create 2D mesh for the component on the outer surface and not on the mid-surface anymore.

And then we convert that 2D mesh to a tetra mesh. Otherwise, in another method, we can simply select the whole volume of the component and we can convert it into tetra. And then there is third utility browser option available for quick tetra mesh. So, in all these cases, if you go by the first method, there is some initial labor involved for the engineer. But because of that, the whole tetra mesh quality will be very much in check. For volume tetra, yes, the control is at second level and for utility quick tetra mesh option, you just feed the geometry and you don't have much control over this particular mesh, but you only can specify the overall general length of the elements.

For brick, there are different options available for creating and we will be looking at them one by one. As far as the quality goes for tetra elements, the rest of the quality criteria don't really matter, but the size of the element which decides the time step, that matters a lot. And apart from that, one very important feature, it's tet collapse. So, tet collapse is much more like, well, it is a ratio and it indicates the height of that tetrahedron element. If you have a tetrahedron element regularly shaped, then perhaps your tet collapse will be zero.

But if you have your height kind of distorted towards your base of the element, then the tetrahedron element would increase and obviously, if you have such a deformed element already, then when the load gets applied, you can imagine what could happen to that element. So, that is why, tet collapse becomes one important quality criteria. So, in general, tet collapse, it is being tried that tet collapse should be kept within 0.15 of the value. So, whatever the tet collapse is, as long as it is under 0.15, it should be good.

So, now, we will quickly look how the tetrahedron mesh can be created. I have a small assembly over here. I have a hub and I have a cover for the hub. So, I will keep one component at a time. So, for hub, now we want to mesh this. Obviously, because this is not a constant thickness component, meshing this in 2Ds on the middle surface is impossible or very difficult. For 3D mesh, well, fitting up hex elements will be tricky because then innermost hex element will be very fine.

And as you go out, the hex size will increase by a lot. So, we have to go for tetra mesh. Now, in tetra mesh, we have so many options available. But first, what we will look at is creating a 2D and then converting to 3D after that. So, for this, first of all, I will create an auto mesh on the surfaces. Now, this time, in auto mesh, instead of selecting mesh type as mixed, we can select trias. So, the software doesn't have to worry about any mesh flow as such.

But as long as there is a tria, a 5mm 2D mesh. Now, let's convert that to 3D. So, I go to 3D panel and then I go to tetra mesh. And tetra mesh, instead of components, I select elements. Now, essentially, we have mesh button to create this 3D mesh. But before that, we can always set some parameters for this tetra mesh. So, obviously, minimum tetra size may have to be selected because, well, we just selected all the surfaces and meshed them. But we don't know for surfaces like this where there is a chamfer or fillet, we don't know how the element size is and if that can significantly drop the time step of the simulation.

So, that's why minimum tetra size will be maintained at 3. Now, this value could be different in your case. Tetco lapsed, I have set it at 0.15 for now and we'll see how it goes. So, I have selected all the elements and I'll hit mesh. So, in the message window, I see that there are a few nodes and some elements created. Make sure that we are on the right page. I'll hide the mesh on the hub. And the volume 1 is indicating the tetra mesh which got generated.

So, to mask and verify, get mask and you see all the tetra hydra elements getting generated on the cross section. Well, so, to check the quality of this, we go to, we press check elements, tools, check elements and then we go to instead of 2D all this time where we spend our time, we'll go to 3D and I'll check for tetco laps which should be 0.15 so there are 0 elements which have tetco lapsed less than 0.15 in fact the minimum tetco lapsed is 0.278 which is a good sign for this component.

This is how we have created tetra mesh using 2D mesh. Now, for 3D or direct 3D mesh, what we'll do is, we'll look at the cover geometry. Now, in this case again, I have an option of creating 2D and then converting to 3D but this time, let's explore something else. So, in tetra mesh, we can go for volume tetra and then we select enclose volume as surface and 2D type should be tria, 3D type is tetra, element size may be 5. So, once we set up these parameters, we hit mesh, now it will try to create all those mesh elements.

Now, that was quick but once it creates that, we get this message and to cross-verify, we have got all the elements that are needed all over the geometry. Now, I can go into this and go to delete, oh, first of all, let's check some element quality as well. Tetco lapsed if there is anything below 0.15, no, if time step 0.5 microseconds, to select the time step, we always need to assign a material to it, so right now, it is not possible to do that. What we can do is, create a material just for the namesake.

So, with this, now I can go to, here it is telling me that the minimum time step in this model is 0.57 e-4, that is way below what we were expecting. So to manage that, we need to increase the tetra element size. So, for now, I will first of all delete these elements. Now, we have the component. I will go to view, browser, HyperMesh, utility. Inside utility, we go to geometry mesh. There are many other options available. Go to geometry mesh and then, make quick tetra mesh.

Components we select, so once I have selected the component, I will hit proceed. Minimum triangle. Let's put 20. Mesh size should be 5, let's say. Mesh density uniform. Everything else at the default stage and we will hit mesh. Now, we see the constant progress of how the quick tetra tool in the utility browser is working. So, the good thing about this tool is that it does some geometry cleanup before going for 2D mesh and then 3D. So, now, we have tetra mesh created. Check. Now this way, and I can see all the tetras getting created. Now the question is, well now the minimum has come to 6.2 E-5. So time step still is in the similar ballpark. Whereas for Tet Collapse, there are still 0 elements failing in Tet Collapse. In fact, the minimum is right now 0.17.

Alright, so with this, we have completed our Tetra Meshing exercise. If I switch on all the components by pressing B, I can get the list of all the components. In that, I'll hit all. And hit return. So that shows me all the components that are there in the system right now. We'll conclude the Tetra Meshing exercise and we'll move to other 3D type or 3D meshing exercises such as Hex and Penta Meshing.

Part 10

Another thing about Volume Tetra Mesh is that while creating Tetra Mesh, sometimes it is important to capture some of the features very well, such as in this case on the screen, there is a hole and with normal Tetra Mesh, very few elements have come on the border. But, well, in some cases, we want a few more elements here just for the sake of taking care of stress concentration or analyzing this particular section better because this could be the area of maximum, well, maximum stress. So, in that case, in the Volume Tetra panel itself, we have two features available, Use Proximity and Use Curvature.

If we have Surface Curvature checked, then the holes or such features will be captured much better. So, in that case, the minimum size of the elements will be followed and that is how we will get some additional elements on the border. With Simply Use Proximity, what it will try to do is, wherever there is thin section, it will try to generate as many elements as possible just to keep the uniformity of the section. So, in this case, without selecting any option, we see only one layer, whereas in this case, with Proximity, we see at least two layers getting formed for that particular layer.

Now, when we combine Use Surface Curvature as well as proximity, that time, both the things happen that the curvatures are getting captured properly as well as there are enough number of elements for the thin sections as well. So, obviously, using both the features many times happens to be very beneficial. Well, usage of these should depend only on the application, many times it happens that, well, some curvatures don't really have any significance in the current analysis case. And in that case, if you still have additional elements defined there, then it is just wastage of resources.

So, moving forward, well, after Tetra Mesh, we'll move our attention to creation of hex elements or hexa elements. Now, as we discussed earlier, hex elements are of two types. Now, the brick elements are essentially the eight-noded elements with penta-six being six-noded. The brick element can remind you of a quad element with thickness. And such as in the case of penta-six elements, a tria with thickness will remind you of a penta-six element. That is why thick shells or eight-noded solid elements are always represented using brick elements and penta-six elements.

Now, to create those, there are a few simple tools available in HyperMesh. In 3D panel, we have solid map, linear solid, solid mesh. Basically, the first two columns of the options available, those are all dedicated for 3D hex mesh. In that, well, essentially, all of these will always ask you for some source phase and destination phase. So, such as in the case of solid map, a general solid map feature will always ask for source phase from where it can generate the mesh and the destination for which it is targeting to reach with whatever the density or size of the elements that we have specified.

So, in this case, the source phases can be multiple. You can literally split this into various sections and create the mesh you like. And then the destination phase, however, it has to be one single phase. So, in case you have another split set of surfaces over here, then you can simply toggle the edges to make it a complete one surface. Apart from that, a drag direction to specify the vector in which the surface phase should move in order to capture the geometry and 3D mesh. So, that is important and along faces will only give the measure the idea about what contours are to be followed.

So, if you had any curve on this particular surface, such as this or even the curve along the longitudinal direction, those curves will be followed until the destination phase and that is how the 3D mesh will be created. And then there are linear solids or solid mesh. Some of them are pretty easy to follow. So, we'll be looking at them one by one using this example. So, in this example, the complete solid part has been split into four different sections. The first section is base. Then, if you notice, the second section is curved.

And from what I have seen, it has a constant area of cross section right from the base to the curved end of the curve. From end of the curve until the boss, there is a straight section. But this straight section has a taper along with it. So, the area over here is definitely bigger than the area over here. And then comes the boss, which is connected to this straight region of the 3D solid model. To start the meshing exercise, first of all, we'll have to create a 2D mesh.

Now, for that, what I'll do is I'll hide rest of the components and we'll focus only on the base. I'll press F12 and the size will be 5. In fact, 5 size happens to be very small. So, first of all, I'll go to mixed elements and the size will be switched to 10. Now, here, instead of mixed, since we want to avoid penta-6 elements because they happen to be very unstable element formulation, we'll hit the option of quads only. Now, the quads only will make sure that in case of triads, it will try to include rotating quads so that the mesh will be adjusted accordingly.

Now, since we have this, we are ready for creating the solid mesh for the base component. So, we'll go for, after creation of shell mesh, we'll go for element offset. In this, solid layers happen to be the most useful for us right now. Elements to offset, we'll select whatever the elements we just created. And the geometry to follow will be two surfaces along which it can probably get a direction. Number of layers can be specified. Now, in this case, I'll quickly check for the distance. If you use F4 on the screen, you'll get to this panel.

Now, the distance is 25. What we'll do is, we'll create the mesh with height equal to 5 mm. And then, total thickness is 25 anyway. So, after giving the along geometry surfaces, we'll hit offset plus. That has created the solids along the line which we wanted. Now, we'll switch on arm curve. Now, first of all, we'll make this component as current so that whatever mesh we create in the arm curve will be stored in the same thing. Now, while hiding this particular component, we also want to make sure that we hide only the, only the 3D elements and not 2D.

So, we'll go to mask elements that is F5 on this, on this keyboard. Under elements, we'll select the elements by config. Now, here the config can be anything. Right now, I'll select hexade because I know I have all hexade elements over here. Out of this plate, select entities and mask them. So, I have only 2D elements left over here now. Now, since we have a curved geometry, we can make use of this spin feature. In that, instead of spin geometries, we want to spin elements and that way we are going to obtain the 3D mesh for those along this curve.

So, for 2D elements, the angle should be 90 and on spin, well, we can simply estimate that we want 25 elements. On the other day, it can be, it can be some other number depending on what density we want. But, most important thing of all, it is asking for a normal direction to the plane, or plane definition, and some axis. So, to do that, first of all, I'll need a center point of this circle. So, I'll hit F4 and in 3-noded measurement, what I'll do is, I'll drag my mouse click on one of the surface lines.

And, I'll take 3 points randomly, and I'll hit circle center. So, it gives me a node which is right at the center of this particular arc. Now, I have the same element selected in the spin elements feature. You have angle specified, 25, and then, well, for the plane of rotation, the x-axis indicates the axis, which is, the global axis, to which the plane is normal, or the global x-axis is going to be normal to this particular plane. So, that is why when we create any rotation, we want this to be normal to x.

For base point, we'll select the point or the node which we just created. Once selecting that, we'll hit spin plus, and it is creating the mesh on the other side. So, in that case, we'll hit reject, and we'll hit spin minus, and it's creating the mesh on the other side now. So, so far, we have created two components of the mesh. What is left is only the last two components. For that, first of all, in this case, we'll be using a linear solid because the mesh on this side and this side would be exactly same. It's just that their cross-section is decreasing in area. So in linear solid, what we are going to need is first of all, elements. So those have to be 2D elements.

So for that, we need the faces of, faces of these elements in this second component. So that we can get the face of the elements over here, that is 2D version of these elements. And that is something we can make the linear solid out of. Now when I hit find faces, it creates a faces component. I'll hide the arm curve for now. Even the base is not needed anymore. We also want, don't want the boss to be here yet. So now we have a mesh over here and we want to propagate that until this particular surface on the curvy, curvy region.

So we'll go to linear solid in 3D. For creating linear solid, we want a source element group and the destination element group as well. But we don't have the destination element group. So we'll press F2 and here we'll try to create the same thing. So we select only the region and now it is essential for us in order to avoid any penta 6 elements. It is essential to match the number of elements. Now I see that this particular section is coming from this region of the curvy bar.

That means we need 3 elements on this particular region. Apart from that, yes, below this point, we have only 4. So right now, since we have 4, we don't need to worry yet. And for this side, we have 3 plus 4, 7. 3 are already here. So 7 is good. And then here we want 4 elements to be there on the curvy region. So now we have a properly represented mesh in 2D world. So going back to our linear solid command, from, what I'll do for from is select the source elements for 2 elements.

Now we have already created the destination jaw elements. And now alignment. Now this is something we need to understand very clearly. For alignment, we need to select 3 nodes on this side and we have to select exact same equivalent nodes on this side. So those 3 nodes would represent this position in the new system. So now in order to do that carefully, we have to go for alignment of first set of nodes and then alignment of second set of nodes. So I'll click here. So I'll click here to switch the selection method. And once this is done, the density. Well, if this region could cover 25 with satisfactory quality, what we'll do here is keep the half of it. So 12 happens to be a good enough number for now.

I've set the density. We hit solids. Now the number of elements in each group is not equal. That is the error. That means we need to make sure that the number of elements are matching on both the sides. So just to give a quick check, we have 3 plus 4, 7 plus 3, 10 elements in a row. That is true. But from that 4 elements, perhaps we need to reselect the model.

Part 11

So after creating 2D mesh on this surface, on the curved surface, and getting the faces out of the arm curve, we'll create the linear solid. For linear solid, we'll first select the source elements. Now to select this plane, what I'll do is select one of the elements and then select by face, so it will select only those. Same thing on this side, I can either go for a box selection method, and under alignment, I'll have to select three nodes on this side. And then the very same three nodes on the destination surface as well, whatever the size of the element they might be creating.

And to do this clearly, what we should do ideally is, in the first box, we select only the nodes for this, and then we click on the second box, and we select those nodes. Now the density along the line happens to be good enough, because if we have covered this in 25 elements, almost half of it could be a good deal for this particular bracket. So I'll hit solid, and now since my current component was set to arm curve, it has stored those elements in the arm curve.

So what I'll do now is, first of all, set the current component to arm street, and then hit shift F11, that is for organize. And then, I'll select the elements by, by the box selection, I will move them to this region. Now once this is done, what I'll do is, go to faces again, that is under tools, and I'll hit delete faces. Now I know there are no 2D elements anymore in this model, which we created using faces. Now there could be 2D elements on the base, that needs to be taken care of.

Now after having these faces, the next thing that remains to be meshed is the boss. Now the boss, for that, the most important thing, or the feature we are going to use is solid mesh. So for that, we always need a source geometry and destination geometry. For source geometry, since we have already defined mesh in one region, the lines over here have been created exactly at the point normal to that particular node, where, where the last mesh left. Same thing on the other side. Now when I hit F12, to select all these surfaces at a time, and then I'll hit mesh.

Now in this case, what we need to make sure is, if, if the, if this particular feature over here is capturing 4 elements, we need to make sure that we have 4 elements meeting that, in this, on this end. Same thing here, about that particular section. Same thing on this side, we need 4 elements, so I have to switch this to 4. On the other side, though, alright, so, we have these many elements, and then, what we'll do is, to feed this whole path, to solid map, and to use these nodes as its guideline, we'll have to create one node on the base, which, which represents a true normal.

So I'll press shift F7, that is for project. Instead of a plane, I'll hit surface, I'll select a node, duplicate it, and then, for surface, I'll hit project. And for a long vector, I'll hit normal surface, so, once we hit project, it will create a temporary node over here. So, now, the next case happens to be, creating the solid mesh, or solid map, for, for completion of this assignment. So, in solid map, what I'll do is, follow the general method, so, under source geometry, usually, you'll see this, but, if you toggle that, to nodes, you can select nodes, for, instead of node path, we can actually select, none.

And for destination geometry, we have this surface over here. Elements to drag will be the elements on the top surface, and then, for along geometry, usually, it could live on surface, or lines, or nodes. Here, we'll hit mixed, and then, for the elements, or for the node path, we'll select all the nodes. The element size should be 10, and, now, we'll try to see what problem, it might give us. For, along geometry, we'll also select a line, just to give it a better idea, and then, I'll hit mesh.

So, it has created a mesh over here. I'll hit escape, and, again, because we had a legitimate mesh on this side, we can see, it will be properly connected. Before that, I'll hit delete button, and elements by config, I'll delete all the 2D elements, that were created during the, along the way. So, quad4, so, elements by config, quad4, select entities, and delete. Well, so, what we can do, in this case is, since this has already been deleted, we don't see any, entities anymore. So, if I switch on the rest of the components, I'll make the boss as current, and the solid map mesh, has been created in the solid map component, which needs to be moved, to its original component.

So, I'll hide the, bar, these elements. With this boss, we have completed our 3D meshing example. Hope you like it, and, just to summarize the whole example, what we started with is, 2D mesh on this geometry, and then we tracked this, using element offset, along, along the surface. Now, to take care of such features, which have gone wrong, what I'll do is, press project, then select these nodes. Instead, we should project it to a line, along the line normal, and we see, it has been created.

So, we have the complete example. We started with a shell mesh, and then, with element offset, we completed the base part. Then, using the spin feature, the 2D mesh, in this base of this, was converted into 3D, and, was brought up to 90 degree level. After that, a linear solid concept was used, with this as, well, this side as, source mesh, and then, using that, we also created the destination mesh. And, we created this 3 mesh solids, which, using linear solid. Now, for the boss, the challenge was, that we can create some mesh here, and, simply, go on, creating the whole component.

But, since we have, this parallel component, connecting a little bit, then on the face. So, we have to follow the nodal path, until the end. And, that is how we get to the result. So once you are done, always never forget to put the card image as part and then hit export. And with this, you can create the 3D mesh model. So we are complete 3D meshing exercise and we'll be looking at 1D element meshing very soon and after that analysis and various other tools.

Want the rest of Crashworthiness Analysis using HyperMesh and Radioss?

Get the full course syllabus and fees.

The full course is paid (₹40,000). We'll send the syllabus and call you back within one working day.

Talk to a career counsellor

Free 20-minute call. We'll map the right course to where you want to work.

By submitting you agree to our Terms and Privacy Policy.

Full course · paid

Unlock the full Crashworthiness Analysis using HyperMesh and Radioss course

The remaining 7 lessons, mentor-reviewed projects and a certificate come with the paid course (₹40,000). Leave your details: we'll send the syllabus and fees and call you back.

Get the syllabus and fees

A counsellor calls you back within one working day.

By submitting you agree to our Terms and Privacy Policy.

Grow by Skill-Lync · free lesson

Start watching

Free, no payment. Leave your details once and every free lesson on Grow plays straight away. A Skill-Lync counsellor may call you about the full Crashworthiness Analysis using HyperMesh and Radioss course.

By submitting you agree to our Terms and Privacy Policy.

Talk to a career counsellor

Talk to a career counsellor

Free 20-minute call. We'll map the right course to where you want to work.

By submitting you agree to our Terms and Privacy Policy.