How to get started with FEA analysis

Finite Element Analysis (FEA), a popular tool in the engineering world, is a numerical method employed for solving complex engineering problems. It excels in tackling engineering riddles that cannot be unraveled easily by standard analytical methods. At times FEA can feel like magic, solutions created by some sorcery, if you may. These codes are complex, but with the right move of the wand, FEA can serve as a powerful tool for engineering by boosting productivity and reducing time and cost.

A popular practice of late is to employ commercially available out-of-the-box FEA code to solve problems. This approach works well for most applications as there is no need to reinvent the wheel. However, for a tailored solution, many organizations and capable engineers are developing their own code, to combat high costs, implementation complications and other limitations while using commercially available code. For the purpose of this discussion, we shall term the end user group using commercially available code as the “FEA user” and those developing their own code as the “FEA developer”.

If you are part of the “FEA developer” group, then be prepared for a lot of math. I am not going to do a deep dive on this topic but in essence your job is to develop the engine for solving complex problems. This requires deep theoretical knowledge of different mathematical models for solving linear and non-linear fluid and solid mechanics problems, constraints, contacts, boundary conditions and more. It also demands a fair knowledge of math, especially for solving differential equations. A very good resource for learning more about this is a book by Klaus-Jurgen Bathe called Finite Element Procedures.

For common applications, as an FEA user, one does not have to be proficient in the mathematics behind the code. The commercial codes available today are extensively tested and have been validated for most of the common applications. While it is important to be aware of the mathematics behind your specific application, you may be required to deal with math only for complex problems such as modeling fracture and fragmentation or when trying to model materials that do not have well defined models. If not, you should be able to work smoothly with the tools readily available in commercial code.

As a beginner, it is easy to get confused and frustrated when trying to create your first model. Most of the time we get lost in trying to figure out the “how”. Finding the options available in the preprocessor and understanding what they do, can be quite challenging. And yet, with the right approach and fitting mindset, no mountain is too high to climb! The key to tackling an FEA tool is not tackling the tool at all. As contradictory as it may sound, at the end of the day, FEA is just a complicated calculator. This calculator, like any, will produce results if the required inputs are provided, right or wrong. Your job, therefore, is not to worry about the processing functionality of the tool, but to approach your problem like an engineer and make use of the tool for solving problems. In essence, this means that to become good at FEA, one ought to first be a good engineer. I believe that following the steps highlighted here would be the best way to start learning FEA. Also, along the way you will learn where to find all the options and functionality in the preprocessor as well!

1. The first and the most important thing is to understand is the engineering problem you are trying to solve. Think about it in a practical sense. Can you do any hand calculations to give you an idea about the problem? Can you figure out what magnitude of loads to expect? What would be a realistic deformation to expect? What are your boundary conditions?
2. Find an example problem, download the input files and study them. For example, if you are trying to learn LS-DYNA, they have a lot of example files available at LS-Dyna Examples One of my favorite examples is a static tensile test simulated with shell elements. This example is designed to learn implicit methods in LS-DYNA. The reasons as to why this example is a great way to learn will become obvious as you read on. This example can be found at the link 🡪
3. Next step is to download the input file deck and run it in your system. Check if the run completed successfully or if there was any error. Scan the output files to identify where you can find important information about your model. For example, the model mass is usually reported in the output files and this is a good way to double check the mass of your model.
4. Once you are through with that, the next step is to take a deeper dive into the simulation. This is why I love this example. It comes with several questions in the description and I encourage everyone to think about them and answer them. Here is a short list of some of them:
   
   1. Which shell element formulation is used?
   2. How many steps are used to apply the load?
   3. Is the simulation linear or nonlinear?
   4. What types of nonlinearity exist in this problem? (material, geometric, contact?)
5. Next the example problem also instructs the user to make changes to the input file and run a new simulation. The new results are then compared to the original analysis and there are more questions to answer such as:
   
   1. Modify the *CONTROL_IMPLICIT_SOLUTION keyword to perform a linear simulation and record the results below for the large applied load. What is the reaction force at the fixed end of the beam? Why does the reaction force not equal the applied load?
   2. Modify the input deck to use the linear element formulation type 21 and repeat the simulation. Does the reaction force now match the applied load?
6. Further you can also experiment with the file yourself. You could try exploring different mesh sizes and element types etc.

Not all examples come with similar questions and learning objectives. However, it is important to follow the same process and formulate your own questions. Once you understand the “why” of the problem and the choices made, it is very easy to find out the “how”. As such, I believe this will be useful information for aspiring engineers trying to unravel the FEA complexities. Keep in mind that it might be a bit frustrating in the beginning but once you start figuring it out you will start enjoying it. Happy engineering and keep yourselves engaged with feedback and comments in the section below!