Automotive FEA Analysis: Software, Process & Case Study

Why Is FEA Important in Engineering?

Finite Element Analysis (FEA) involves simulating various physical phenomena using the numerical technique known as the Finite Element Method (FEM). Engineers employ FEA software to streamline the design phase, minimizing the need for physical prototypes and experiments, and enhancing product development efficiency while reducing costs.

A comprehensive understanding and quantification of physical phenomena, such as structural or fluid behavior, thermal transport, wave propagation, and the growth of biological cells, require mathematical analysis. Many of these processes are delineated using Partial Differential Equations (PDEs). However, solving these PDEs computationally demands numerical techniques, with Finite Element Analysis emerging as a prominent method over recent decades.

What Is the Meaning of FEA?

Finite element analysis (FEA) provides a way for engineers to test various different load cases or failure modes virtually. This helps different OEMs iterate faster and lead to more efficient designs at a reduced cost. In today’s world FEA is vital to helping engineers get a better understanding of how a system works. However, any computer software is only as good as the operator using it, and engineers must still ensure that all the assumptions they are working with are accurate and true.

What Is FEA Used for in the Automotive Industry?

FEA is a broad term used to cover the use of computational tools to help understand various different load cases in a vehicle. The goal of these virtual models is to discover vulnerabilities in the vehicle design prior to automakers kicking off prototypes or production tools. For example, engineers can use FEA software to run side-impact crash testing (FMVSS 214) on a vehicle prior to ever having built any prototype vehicles. If they see that a vehicle is not meeting the desired crashworthiness targets, they are able to quickly modify the vehicle geometry or change the material type/gauge of the weakest link in the design. This kind of analysis is a core part of how automakers evaluate unibody and body-on-frame structural stiffness long before a single physical part is built. An example of how FEA is used in the automotive industry can be seen below:

Common FEA Software Used in the Automotive Industry

Automakers and suppliers rely on a handful of industry-standard FEA solvers, each with its own strengths:

• Altair HyperWorks (HyperMesh, OptiStruct, HyperView): widely used for pre-processing, structural optimization, and post-processing; this is the toolchain used in the case study below.
• Abaqus: a popular choice for nonlinear structural analysis, including crash and impact simulation.
• ANSYS: a broad multiphysics platform used across structural, thermal, and fluid simulation.
• LS-DYNA: the industry standard for explicit crash and high-speed impact simulation.
• Nastran: commonly used for linear structural and vibration (NVH) analysis.
• SolidWorks Simulation: a more accessible, CAD-integrated option often used earlier in the design process or by smaller suppliers.

The choice of solver usually depends on the type of analysis (linear vs. nonlinear, implicit vs. explicit), the OEM’s existing toolchain, and whether the model needs to integrate with upstream CAD or downstream optimization workflows.

Case Study: FEA Analysis on a Front Motor Compartment Cross Member

This project explores the deflection of a body-in-white (BIW) front motor compartment cross member, which is subject to aerodynamic loading. One function of a front motor compartment cross member is to maintain the hood latch interface to prevent the hood from opening under normal vehicle operating conditions. To simplify the modeling of this problem, the front motor compartment cross rail will be shown as a square 30 mm x 30 mm beam which is fixed at both ends (beam bending problem). The aerodynamic load will be modeled as a 1000N point load applied at the center of the beam. The beam is made from steel with a modulus of elasticity of 210 GPa, a Poisson’s ratio of 0.3, a density of 7850 kg/m³, and a yield strength of 210 MPa. See Figure 1 below: problem description.

Model Development: Loading and Boundary Conditions

The problem was solved using Altair HyperMesh. The front motor compartment cross member was modeled as both a 1D and 2D element. To study the impact of the mesh size, the 2D front motor compartment cross-member mesh was varied from 1mm – 5mm. The 1D element was created by first placing 3 nodes and connecting them with a beam (1D section from the main menu). The properties of steel were created and assigned to the beam using the CBEAM command. In addition to the properties, the beam element cross section was also assigned using the same command. Finally, using the boundary condition tab, two ends of the beam were fixed and a 1000N load was applied to the center. The results can be seen in Figures 2 and 3 below.

To model the beam as a 2D element the first step was to create the 4 nodes which will make up the square end cross-section of the beam. These nodes are then connected using the geometry tab to create a square. Then the square is then extruded, and a beam is formed. It’s important to assign the correct properties (beam thickness) and material to this beam to ensure the correct result, see Figure 4.

Unlike a 1D element, a 2D element requires you to mesh the beam. Whenever you are meshing any part, it is important to make sure the elements are as close to square as possible, since this part has no round edges which is not much of an issue. When applying the boundary conditions, it is important to select all the nodes at each of the ends of the beam. The smaller the mesh size, the more nodes you will have to select. Make sure each of these nodes at the ends is constrained by all 6 degrees of freedom to represent it being clamped down. Since you are applying a point load to the beam, you first need to create a reinforcement (RBE3 under the 1D menu) to ensure your load is being distributed among an area in the center. Failure to do so will skew your results since the top of the beam subject to the load will deform much more than the opposite end. See the Figure 5 model setup below.

Just as before, once the model has been created, the boundary conditions applied, and the RBE3 reinforcement in the center of the model, you can run your model using OptiStruct. Once the HyperWorks solver has finished running your model, you can then view your results and perform post-processing using HyperView. See Figures 6 and 7 to view the results from the 2D model below.

Results and Discussion

When comparing the effects of varying the size of the mesh, the results were very interesting and not what I expected. Even though the displacement did not vary much (-1.73 mm to -2.08mm), the stress results varied drastically from 235 MPa to 1967 MPa. Of the three trials conducted, the most realistic appears to be trial 2 with the mesh size of 3mm. See Table 1: 2D displacement/stress results below.

When comparing the effects of running the model 1D vs. 2D, the results were less surprising. The displacement on the 1D model was less than on the 2D model. Since the 1D model had the fewest nodes, it is expected that the result would be the least accurate. Unfortunately, I was not able to get stress results for the 1D model; the only logical reasoning behind this is that since it was modeled as a line, it has no theoretical cross-sectional area, making it impossible to calculate stress. Further trials would need to continue to investigate. See Table 2: 1D vs. 2D displacement/stress results below.

Sample HyperWorks solver output can be seen below, with calculated displacement for the 1D and 2D models.

Conclusions and Recommendations

In conclusion, to receive better results I would recommend modeling the load as a distributed load along the majority if not all of the beam. This will help spread out the load among a greater area and will increase the accuracy of the results. The reason I believe my stress was so high for the 2D model (1mm mesh & 5mm mesh) was that HyperWorks seems to treat my load more as a point load and would not distribute it amongst the full beam as I had hoped. See the image below. If I were to continue with a point load again, I would recommend increasing the area of the RBE3 in hopes of improving the results of the 1mm and 5mm mesh models. For a related look at how mesh and structural choices affect a full vehicle rather than a single beam, see our guide to unibody vs. body-on-frame FEA rigidity data, and for how these structural components are actually manufactured, see our guide to sheet metal hydroforming.

Frequently Asked Questions

What FEA software is used in the automotive industry?

Common automotive FEA solvers include Altair HyperWorks (HyperMesh, OptiStruct, HyperView), Abaqus, ANSYS, LS-DYNA for crash simulation, Nastran for linear and NVH analysis, and SolidWorks Simulation for CAD-integrated analysis.

What is FEA analysis used for in cars?

Automakers use FEA to virtually test load cases like crash impacts, structural stiffness, and vibration before building physical prototypes, allowing them to catch design vulnerabilities and iterate faster and more cheaply.

What is the difference between 1D and 2D FEA elements?

A 1D element models a part as a line with assigned cross-sectional properties, which is faster to set up but cannot calculate stress directly. A 2D element models the actual cross-sectional geometry and mesh, giving more accurate displacement and stress results at the cost of additional setup and computation time.

Why does mesh size affect FEA stress results so much?

Stress results are highly sensitive to mesh size, especially near concentrated loads, because a finer mesh captures sharper local stress gradients. Displacement results are far less sensitive to mesh size, which is why displacement can vary only slightly while stress varies drastically across mesh sizes.