ANSA Project 2 - Meshing of suspension assembly and deployment of connections within said assembly

OBJECTIVE

To mesh the given rear suspension model in both 2D and 3D (according to the given quality criteria) after checking for and clearing geometrical errors. Also, to deploy appropriate connections as required between various parts of the assembly after which the model is symmetrically replicated using the reflect tool.

QUALITY CRITERIA

FOR PARTS THAT REQUIRE TETRA MESH:

TYRE & RIM (Trias)

OTHER REGIONS (Trias)

TETRA ELEMENT QUALITY REQUIREMENTS

FOR PARTS THAT REQUIRE 2D MIDSURFACED MESHING:

MODEL IMAGE

PROCEDURE

2D MESHING

1. To start off, the topo tool is used to take care of all the single cons and to ensure there are no free edges. It can be accessed via Faces > Topo in the topo module. In this case, due to the proximity of edges in this assembly, it's a good idea to carry this out separately on each part to avoid triple con formation. All the cons are selected and the algorithm takes care of all the single cons (if any).

For example:

Close up:

In general, this model does not have many single cons (as is evident from the above screenshot, which is actually one of two single cons from the entire model - which are both on the same part.)

2. Then, before we carry out any other actions, we need to carry out a geometry check and fix all abnormalities. That is done via Checks > Geometry where we can execute checks of certain defects that can be selected from the menu and the algorithm looks for said defects. After detection, we can select all the defects and right-click > fix to fix them.

3. We can go ahead and measure thicknesses of various regions in the component using the Measure tool from the top toolbar. This is done to decide the type of meshing we ought to carry out.

As a rule of thumb, parts with uniform thicknesses (throughout) lesser than 5mm will be 2D meshed after generating their midsurfaces and the others will undergo volumetric meshing.

4. Along with taking thickness measurements, it would be easier to assign PIDs to various parts of the model as well. In order to set up some connections, it is important to differentiate components with the help of PIDs. This entails the first step. The entire model will be separated into different PIDs. This requires creating said PIDs and that can be done via Faces > Set PID in the topo deck. The part is then selected and the middle mouse is clicked to bring up the properties window.



Here, the new property can be created and assigned as seen above (Thickness - T in this case). This process is repeated until PIDs have been assigned for every component/part in the model.

While assign PIDs, we can midsurface the parts that will require 2D meshing and then assign PIDs to these midsurfaces as well. For midsurfacing, I used the offset tool (Faces > Offset). When activated, it will ask us to select the surfaces to offset. We can select the surfaces on one side of the part. After that, we can assign the offset value (half the thickness). The offset tool will show the direction of the offset in the form of an arrow. If it is in the opposite direction of the required offset, the offset value needs to be in the negatives. If we proceed, the surface will be offset and we will have the option of keeping the parent surfaces.

5. After this, the surfaces need to be prepared for meshing. This means regions that can result in minimum length failures need to be tended to. In such locations, cons can be suppressed at the expense of features as long as it's minimal. This is done by right-clicking the cons using the Faces > Cut tool.

This process is carried out on all surfaces - midsurface or not, keeping in mind the corresponding minimum lengths for their respective meshes.

6. Before meshing, we need to set the target length (through perimeter > length). With that taken care of, we can assign the other parameters. This is done by accessing mesh parameters and quality criteria from the top toolbar.

The following are parameters and quality criteria for parts (excluding tyre and rim) that are to undergo volumetric meshing (Tria mesh):

Similarly, here are the parameters and criteria for the midsurface 2D meshes:

and finally, the parameters and criteria for tria 2D meshes on the tyre and rim:

7.We can then start meshing each of the parts as per the mesh requirements (2D Mixed mesh using Mesh Generation > Best from the Mesh module and 2D tria mesh using Mesh Generation > Spot). In the following example, I have started meshing the rim starting from the centre, in an outward direction.

After that, we can make use of the reconstruct tool (Shell Mesh > Reconstruct) to generate a better mesh if the initial mesh is not satisfactory. If required, the smooth tool (Shell Mesh > Smooth) can be used as well.

Some regions will end up having off elements anyway. Due to feature constraints, they cannot be fixed, as can be seen here:

8. After meshing all regions, we can proceed to the next step of assigning volumes manually to each set of closed-shell elements (for volumetric meshing). This is done by using Volumes > Define > Manual and then selecting the regions that comprise a particular volume. Auto automatically detects and defines closed volumes that are selected whereas the manual option lets us define properties for the volume that is to be defined. Either method works here.

After the volume is defined, the list of defined volumes can be accessed from Volumes > List. The volumes that are unmeshed can be 3D meshed by rt. clicking the particular volume and selecting 'remesh' as shown:

We then have the option of using either one of the tetra rapid or tetra FEM algorithms. Doing so generates the solid mesh elements within each volume.

In addition to that, clicking 'edit' lets us edit the volume properties if required.

An example of 3D generated tetra elements:

This process is repeated individually for every available volume in the model.

9. Next, we need to check the solid elements to see if any fail for the tet collapse criteria of 0.2. Before doing that, we will need to access the solids tab as well on the quality criteria window to input the tet collapse criteria.

Now we need to switch to hidden mode to check for tet collapse errors. There are two ways to fix tet collapse failures. One is by using the Fix Quality tool from the volume mesh module (Improve > Fix Quality). This is an algorithm that works on failing elements. But it may result in protrusions or depressions on an otherwise flat surface.

An example of a protrusion:

To avoid this, the MV free tool can be used (Grids > MV Free) which is available in the mesh module. This lets us manually drag particular nodes of the failing element such that the height of the tetra element can be increased. This helps pass the tet collapse criteria. Understandably, to avoid the same problem as in the Fix Quality tool, care should be taken to not move nodes that are on the surface of the component but the ones that are below the surface.

MESHING THE SPRING COMPONENT

1. For this, we shall be using 1D CBAR elements to represent the part. Firstly, we can generate a middle curve that runs through the entire spring by using the middle tool from the curve options (Curves > Middle) of the topo deck. We will need to select the cons that run through the entire length of the spring that are on opposite sides of the part as shown:

2. We then have to convert the curves into a single one, if it has been broken up into multiple curves. This is done using the Curves > Connect > Multi tool.

3. After that, we can define a COG point for the curve generated from the spring (Points > On C.O.G.).

4. Having done that, we can move to the NASTRAN deck and go to Elements > CBAR and in the window that pops up, 'orientation with nodes' needs to be checked and the selection needs to be 'Curve (box)'. The curve is then selected and we can click the middle mouse button.

5. After that, we are asked to pick the orientation point and we can select the COG for that. On proceeding, we are to specify the element length as well. We can go for 4mm.

6. After that, on clicking the middle mouse button, we are to assign a property to this curve. We can go ahead and create a new PID. In the next window, we are to input the area (A). In that selection box, we can press Shift + ? to open another window where we simply need to specify the cross-sectional radius of the spring. On doing so, the algorithm automatically calcualtes and populates the other values on the PID card as per the given radius.

7. In the next window, we need to define the CBAR element by simply naming it. I named it 'spring'.



8. On proceeding, we will notice that the 1D elements have been created.

9. These can be represented as 3D solids as well. To do that, we simply need to check the 'Lines Element Cross Section' box and select 'as solids'. These options are available in the presentation parameters tab of the Quality Criteria window.

CONNECTIONS

1. Now that all the parts have been meshed, we can move onto deploying the connections.

2. To connect regions where a bolt and nut is required, we can use the Two node RBE2 option.

The RBE2 tool can be accessed via Elements > RBE2 in the NASTRAN deck. It primarily has the 'many nodes' and 'two nodes' options. This procedure involves connecting two sets of RBE2 spiders using a double con. Firstly, the RBE2 component is created separately for each hole using the Elements > RBE2 > Many Nodes option. We can select the nodes surrounding the hole and middle mouse click to form the elements.

Since this is just a representation, we needn't worry about selecting all the nodes involved in the hole. One set on one end of the hole on the component is enough. We can do the same for the hole on the other component as well.

Through this process, we'll have two master nodes (which are to be connected). To connect them, we shall be using the Elements > RBE2 > Two Nodes tool. Then it is just a matter of selecting the two master nodes to form the connection.

This process is repeated for all locations where a bolt connection is to be deployed.

3. The other type of connection we will be deploying in this model are simple two-node RBE2 connections. Not to be confused with the one discussed previously. In this case, we will be providing rudimentary RBE2 connections between two corresponding nodes on different components. This will be used in locations where a shaft like part/cylindrical part is lowered into a hole.

Firstly, we will be selecting Elements > RBE2 > Two Nodes and we will be picking the nodes to be connected as shown:

On proceeding, we can define the RBE2 element to be created:

On clicking the OK button, the connection is created as shown:

We need to make sure there are at least 4 such connections for each location as shown:

4. Another type of connection deployed in this model would be the seam weld connection. In the case of seam welds, the basis is creating the weld locations from curves derived from the particular feature. Therefore, we shall be using the Curves > Cons2Curve option from the topo deck. This will let us select the edge and derive a curve from it as shown.



Then, using the convert tool from the toolbar on the top, the curve can be converted into a seam line as shown (after selecting 'curves' from the dropdown menu):

Now, we can make use of the connections manager (by accessing it from the upper toolbar). The seam line is selected and on proceeding, we get a set of options where we need to tell ANSA what kind of connection we intend on creating. P1 and P2 are selected - To select the PIDs for each, right-click on the option P1 or P2. A box pops up. There, we need to press F1. This will let us select the particular PID. (care must be taken to ensure the right component is selected as the base sheet - the layer on which the weld leg will rest). The FE Rep type here would be Overlap Shell since both layers are parallel to each other.


The root shell option would be 'double row'. Other properties (such as width or thickness) and PID can be assigned for the weld as well if needed. Since the standard length is 10mm and that would affect mesh flow, we can change the width to 4mm to maintain uniformity with the mesh of the connecting parts.

And here is what it looks like:

5. Finally, the spring elements need to be connected to their neighbouring parts as well. For this, we can use cluster RBE2 elements to do so. The process is straightforward. We will need to use the RBE2 tool again. It can be accessed via Elements > RBE2 in the NASTRAN deck. In this case, we deploy the 'many nodes' option to connect the spring and neighbouring part via a single master node. This is merely done by selecting the two sets of nodes - one set from the spring and the other from the part (on the corresponding locations) and letting the algorithm create a master node and form connections. This is also known as a cluster node RBE2.

REFLECTING

After we have meshed the entire model and deployed all possible connections, we can go ahead and reflect it to produce the other half of this rear suspension system. This is done using the reflect option from the Transform tool.

To start off, we need to go to Transform from the upper toolbar. Then select 'Copy'.

Then we need to select the elements that are to be reflected. We currently have all the meshed elements and connections in view right now so we can go ahead and select everything.

On proceeding, we are given the copy options. Here, we can go to the symmetry tab. For the 'Use' selection, we have multiple options, any of them should work provided we give the algorithm the right input. I went ahead and used the 'Mirror 3 Points Plane' option. With this, I will need to define a symmetry plane using 3 points as shown below:

Then we need to click middle-click twice and click on 'finish' on the next window:

And there we have it, our reflected suspension system, now with both halves:

FINAL MODEL IMAGES

2D Mesh Quality

Mesh for midsurfaced parts (mixed)

With 'draw shell as solid' option activated for midsurfaced regions:

Mesh for Tyres and Rims (trias):

The only off elements are these (unavoidable due to feature constraints):

Mesh for remaining regions (trias):



The only off elements are these (again unavoidable due to feature constraints):

3D Mesh Quality

Full suspension model with no tet collapse elements:

A view of the entire model with some of the connections:

LEARNING OUTCOMES

1. Through this project, I was able to familiarise myself further with tetra-meshing and connection deployment processes.

2. Learnt to use 1D CBAR elements to represent spring components.

3. Learnt more about various parts involved in a standard suspension assembly.

RESULT

The given suspension assembly was checked for geometrical errors and fixed accordingly. In addition to that, it was meshed in both 2D and 3D as per the given quality criteria and parameters. Connections were deployed and finally, a symmetrical half was created and reflected to complete the other side of the assembly.