Week - 4 - Crash Box Simulation

Objective:

• To simulate a crash test for a crash box for which the meshed file has been given.
• A rigid wall has to be created in front of the crash box, and the crash box has to be impacted on this rigid wall with a velocity of 50kmph. The rigid wall can be created using the *RIGID_KEYWORD. 
• The material that the crash box has to be assigned with is elastic material (initially, for visualization purposes) and later steel for the main simulation.
• A thickness of 1.2mm has to be assigned to the crash box and the simulation has to be carried out, and later this thickness has to be changed to 1.5mm and the results have to be compared.

Deliverables:

• Input .k file and output files (d3plot, glstat, sleout, rcforc)
• Animation of the final simulation
• The cross-sectional force generated in the middle of the crash box.
• Acceleration plot of a node in the middle of the crash box.
• Maximum directional stress and strain along the length of the crash box.
• The plot of all energies (total, internal, kinetic, hourglass, sliding)
• Compare the accelerations and stress/strain plots of 1.2/1.5mm crash boxes.

Introduction:

• A crash box is a high-energy-absorbing structure that crashes on the application of loads and reduces the impact on other components nearby. A full-fledged crash box is a highly sophisticated design but in this case, we will go with a rectangular channel which can be thought of as the most simple crash box.
• This crash box has to be impacted on a rigid wall. There are different types of rigid walls available in "LS-DYNA" such as geometric rigid walls (flat, prism, cylinder, sphere), and planar rigid walls. In this challenge, we have to use the planar rigid wall. 
• The rigid wall, as the name suggests, is a non-deformable, rigid body, which can be created without a mesh to represent a real-life rigid structure. Creating a rigid wall keyword, instead of creating an actual mesh and then defining it to be rigid, saves computational time and costs. This makes problem-solving much easier.
• Here, for this challenge, we are using the unit system as 'gm mm ms'. Therefore the velocity of 50kmp with which the crash box has to be hit on the rigid wall has to be converted to mm/ms, which comes out to be 13.88 mm/ms. Using this value, an initial velocity card or entity has ti be created. 
• After completing the case setup, the output requests have to be created. These output requests can later be viewed in the post-processing stage in the 'post' option in the "LSPP".

Procedure:

• The file which was provided in the '.k' format was imported in "LS DYNA" using the 'import keyword file'. It was a pre-meshed rectangular crash box as seen in the image below.
  
  
  
  
• A section keyword was defined for this crash box from the 'keyword manager'. Since the element type used here is shell elements, the keyword selected was *SECTION_SHELL. The thickness of the crash box initially had to be kept as 1.2mm, hence the 'T1' was changed to 1.2. The image below shows the keyword card.
  
  
  
  
• As mentioned in the challenge, the material was initially defined as an elastic material for visualization purposes, later it was changed to 'steel', with the default values of the properties of steel. The material keyword was selected as MAT024 Piecewise linear plasticity, which is a linear material, and was defined from the 'keyword manager' as *MAT_PIECEWISE_LINEAR_PLASTICITY. The image below shows the material definition. All the properties of the material were defined with respect to the gm mm ms unit system.
  
  
  
  
• After defining the section and the material keyword, their IDs were specified in the *PART keyword, in order to assign the material and the section to this specific part. Along with this, the title for the part was also specified as 'Crash box'.
  
  
  
  
• Since the crash box is impacting the rigid wall along its length, there is a high chance that the crash box may deform, though no completely but at least by some minimal amount. The elements of the structure may "touch" each other, and we have to consider that situation in order to get fairly accurate results. For this purpose, we have to define a contact keyword in the analysis. There is only one body present in the analysis, and that may get in contact with itself. Therefore, we will define a 'self contact' keyword, *AUTOMATIC_SINGLE_SURFACE.
  
  
  
• As seen in the image above, a slave set ID (SSID) and master set ID (MSID) have to be defined. But since we are using single surface contact, SSID is sufficient. A node-set has to be created which consists of all the nodes of the crash box. This can be done using the 'create entity' option. Under the 'set data', 'SET_NODE' is selected, and by switching to 'cre' and selecting 'area' all the nodes are picked, as seen in the image below. 
  
  
  
  
• Later, a rigid wall has to be created in front of the crash box on which it has to be impacted. So, a planar rigid wall was created from the 'entity creation' option under the 'rigid wall' tab. In this, the 'planar' option was selected, and in order to specify the direction and orientation of the rigid wall, an element on the 'width' of the crash box was 'picked'. A small gap had to be specified in between the rigid wall and the crash box, just for visualization purposes. This gap of 5mm was specified in the "X" direction with a negative sign, according to the sign convention and the co-ordinate system. This created a rigid wall in front of the crash box as seen in the image below.
  
  
  
  
  
  
• As mentioned in the 'deliverables', a small section in the middle of the crash box along its length had to be analyzed for the sectional forces. Therefore, a node-set consisting of a small strip of nodes was created using the 'create entity' option similar to the previous operation. A section keyword from the 'keyword manager', *DATABASE_CROSS_SECTION_SET was created, and the NSID of the node strip was mentioned in it.
  
   
  
  
  
• Since the crash box had to be impacted on the rigid wall, an initial velocity had to be added to it. This was done by creating an entity from 'entity creation', and selecting the 'initial>velocity'. All the nodes of the crash box were selected since the velocity has to be applied on the whole crash box. The magnitude of the velocity converted in mm/ms was 13.88mm/ms. This value was entered in "X" with a negative sign since the velocity has to be applied towards the rigid wall according to the co-ordinate system. This can also be done by creating a keyword from keyword manager.
  
  
  
  
• In order to avoid the generation of hourglass energy, an hourglass keyword has to be created. This is done by creating *HOURGLASS from the keyword, and the value of 'ihq' was set as 2, while all the other values were kept constant. As seen in the image below this hourglass control is based on the formulation of Belytschko and Tsay. 
  
  
  
  
• After all the initial case setup, the termination time was defined as 2ms. For this a *CONTROL_TERMINATION was created and the 'ENDTIM' was specified as 2.000ms, keeping all the other values default.
  
  
  
  
• Along with this, a 'd3plot' was also created to define the time step value for the simulation and for the database for the entire model. The DT was set as 0.1 as seen in the image below.
  
  
  
  
• Later for the output requests, different ASCII plots were created like the glstat, matsum, secforc, rwforc, rcforc, sleout, etc. from the *DATABASE_ASCII .
  
  
  
  
• After ensuring all the input variables and output requests were defined properly, the keyword was saved. The simulation was then run from the "LS RUN", and the termination was normal as seen below.
  
  
  
  
• The animation below shows the Von-Misses contour for the crash box. The maximum value of stress reached in the crash box upon impact was observed as 3.72e2 MPa.
  
  
  
  
• As per the requirement, the thickness of the crash box was changed from the *SECTION_SHELL, and was updated to 1.5mm. The keyword was saved separately and the simulation was again run.
  
  
  
  
• The animation below shows the Von-Misses stress contour for the crash box having a thickness of 1.5mm. Upon comparing the two simulations, it can be seen that there is a  variation in the maximum values of the Von-Misses stress. The maximum value of the Von-Misses stress for the thickness of 1.5mm crash box is around 3.731e2 MPa
  .
  
  
  Further below is the comparison of other deliverables for the two thicknesses 1.2mm and 1.5mm respectively.
  
  Cross-sectional force:
  
   
  
  
• As seen in the images above it is evident that the crash box of thickness 1.5mm is experiencing a higher sectional for of a little more than 0.2e6 N, whereas the maximum sectional force on the crash box of thickness 1.2mm is around 0.18e6 N. The sectional forces are zero until the impact happens, and later there is a steep surge in the forces upon impact.
  
  Acceleration Plot of middle node:
  
   
  
• The acceleration of the middle node with ID 5149 was plotted for both the thicknesses. Initially, until the impact the acceleration was zero. But after the impact, the maximum acceleration in the 1.5mm thick crash box was more than its thinner counterpart. The maximum value noted in the 1.2mm and 1.5mm thick crash box was around 3.12e3 mm/ms^2, and 3.25e3 mm/ms^2.
  
  Maximum Directional Stress:
   
   
  
  
• This plot shows the variation of effective stress for element no. 6182 which was at the front of the crash box for both the thicknesses. Though the maximum value of the effective stress upon impact for both the thicknesses was somewhat similar, there was a notable difference in the plot after the 0.5ms mark.
  
  Effective Plastic Strain:
  
   
  
• The images above show the effective plastic strain for both the crash boxes for a specific element no. 6182. There was no difference between both plots.
  
  Energy Plots:
  
  
  
  
  
• The images above compare the plots for the different types of energies viz. kinetic energy, internal energy, total energy, hourglass energy, and sliding energy. There is a notable difference between both the plots. Upon impact with the rigid wall, the kinetic energy of both the crash boxes reduces drastically since the box slows down. The kinetic energies, right after the impact at 0.5ms, for both 1.2mm and 1.5mm thick crash boxes, was recorded as 1.18e4 and 1.553e4 Nmm.
• The maximum internal energy for the 1.2mm thick model was 9.54e4 Nmm, and 1.19e5 Nmm for its 1.5mm counterpart. The total energy for the thicker model was also high at a magnitude of 1.35e5 Nmm, while the 1.2mm thick has a total energy of 1.08e5 Nmm. Apart from these energies, the hourglass energy was very low, and the sliding energy was 0 throughout the simulation.

Conclusion:

• In this challenge, the crash box was simulated for an impact on a rigid wall successfully. Initially, the material for the crash box was defined as an elastic material for visualization and understanding purposes, which was later changed to linear material, steel, with the MAT 024 *MAT_PIECEWISE_LINEAR_PLASTICITY keyword. 
• The main purpose of this challenge was to give an overview or introduction to crash analysis of complex models like vehicles. The model provided in the challenge was already meshed, and the only objective was to do the case setup and post-processing. The thickness for the crash box was defined as 1.2mm, which later, had to be increased to 1.5mm for the comparison of the results.
• A rigid wall was created in front of the crash box at a distance of 5mm from the face for visualization purposes. The type of the rigid wall was set as planar as all of the 4 sides of the face of the crash box had to be impacted on it at the same instance. The output requests of the crash box included ASCII files of glstat, matsum, sleout, rcforc, rwforc, etc.
• The parameters that had to be checked and plotted for both the thicknesses included stress and strain plots, energy plots, crash animation, force plots, acceleration plots, etc. There were some notable differences in the results of both the crash boxes. The only reason behind this was the thickness of the section. 
• The crash box with a thicker section had higher total energy values, higher stress values, and higher cross-sectional force values, as compared to its thinner counterpart.
  
  

GDrive link for all the related files:  Crash box files