This blog is translated from German with DeepL.

A field report from the perspective of a product developer.

How do we use simulations and FEM analyses in our developments?

It is always the question whether one has enough security with the experience and know-how from other projects, whether one wants to make a rough calculation or whether a precise simulation is necessary. Gimelli Engineering AG is not a calculation office and is not specialized in calculations. You can work with empirical values if you can build in a reserve. We use simple tools for a rough calculation, mostly based on our CADs. They are used more for a qualitative statement, i.e. for a rough assessment or to judge the effect of optimizations. If a precise calculation is needed, we use our partners, e.g. pinPlus. This is the case when customers or regulations require a corresponding proof, or when we are pushing the limits with the design.

Often FEM is already integrated in a parameterized CAD software. Distinctions in the offerings:

It is always a matter of distinguishing whether one wants to obtain a qualitative or quantitative statement.

Qualitative statement: you want to make a very rough calculation to get a feeling where you are with the construction approximately. Or you want to see where or how much changes in the design affect the loads.

Quantitative statement: here you want to know the exact value. A quantitative statement must be prepared professionally, because the quality of the statement is extremely dependent on the preparation of the data, especially the meshing. If you do not choose the right meshing at the right place, you will either get wrong values (too coarse meshing) or you need too much computing power & time (too fine meshing).

Programs for a fast, qualitative statement usually calculate with an automatic meshing. Often the meshing cannot be influenced or can only be influenced globally. I.e. everywhere the meshing is equally fine and one can only influence how fine the global meshing is, which automatically affects the computing time. Thus, one has no possibility to mesh finer in delicate or filigree areas for a precise result and to mesh coarser in hardly loaded or coarse areas to save computing power. This is actually the main problem of the fast variants.

Figure 1: Professional networking of a part for quantitative statement. Source pinPlus: Tip Digimat Case Study

Differences price rails: The more expensive the program, the more precise the calculation. The more computing power and effort you need, respectively the more knowledge you have to have yourself. A colored FEM image is useless if you can’t interpret it yourself and understand what the colors mean. And if you do not know what is important, you will not recognize when it is critical and, above all, how meaningful the result is.

Today there is a tendency: calculating is good, optimizing is the goal. There are more and more offers that not only calculate, but also optimize with as little effort as possible. For example, you can change the topology (critical points) and see immediately what happens. If you can immediately mitigate in the most stressed areas, this brings the most advantages in the development.

What matters most in FEM simulations.

As a developer, it is important to know what you are doing in order to have a rough result quickly. For the calculation specialist, it is important in an FEM analysis to have a result that he can rely on rock solid. For the SLM specialist who produces additively, the most important thing is that he can change the topology (remove material if necessary and add it at critical points).

Figure 2: Example of a simulation of the stresses that occur when a snapper is pressed on. (Simulation with CADFEM Discovery Live)

Figure 3: Example of a simulation of the stresses generated when holding a snapper. (Simulation with CADFEM Discovery Live)

Other simulations, besides the typical strength analysis.

There are very many simulations that can be implemented with appropriate software. In our field, we use simulations in injection molding. For example, for narrow parts, we use a filling simulation to see afterwards whether the part is filled well and where there are flow seams that create weak points in a part. These, in turn, are closely related to strength. So if you want to calculate an injection molded part for strength, you also need to know how it was injected. Otherwise, you can’t make an accurate statement because the injection molding structure has an extreme influence on the strength. We developers are also very interested in warpage. This is an area that cannot be implemented with a simple tool, because warpage also depends not only on the part, but on the mold design and the injection parameters.

E.g., where to inject, at what cross-section, temperature and pressure, what temperature my mold is, how close to my geometry cooling lines can run, etc. Not only the injection molded part is considered, but also the mold and the cooling behavior. Accordingly, you can well imagine that it becomes more complex again. But in many cases, that’s exactly what interests me as a developer: How crooked will my part be? What happens when you add a rib? Does the rib pull or push on my part? 

Parallel to FEM there are simulations with fluids, gas, flow behavior and partly also thermal. CFD software are mostly designed for this. There is also software that can combine FEM and CFD. Nowadays, electronic devices have tight spaces and little air exchange, but at the same time more and more electronic components that generate heat.

Figure 4: Example of flow and temperature simulation of electronic components.
(Simulation with CADFEM Discovery Live)

(Simulation with CADFEM Discovery Live)

One can run a simulation that shows heat generation and what happens when vents are installed. There are variations of FEM that are applied specifically for gearboxes or specific assemblies, for example. FEM itself has a lot of areas of simulation, there are a lot of differences just from materials. Metals, for example, are much more computable than plastics, because the structure and behavior of plastics are not linear in many areas. But even with metals, it is important to know how long you can calculate linearly and when, for example, plastic deformation completely changes the calculation.

With plastics, the aging issue is often added, e.g. steel ages very little, so it is an almost timeless calculation. But if you make a calculation with plastic, you have on the one hand the aging of the starting material and the properties that change in the plastic. On the other hand, you also have the aging of the assembly, or the stress in the assembly. When you clamp something, this tension weakens over time, which means that if you press one part into another, it may lose its tension after years or even fall out.

In addition to aging, the environment also plays a role. With steel, corrosion can cause problems and failure. With plastic, UV radiation or chemicals can make the part brittle and, in the worst case, my part will crumble after a certain time.

More information about FEM simulations in our tip: Digimat Case Study.

Leave a Reply

Your email address will not be published. Required fields are marked *