5 Life Lessons you Should Apply to your Simulations
Life of a product designer or an engineer can be tough. Deadlines, tight budgets, design constraints and quality assurance are what makes this job so challenging (but also so interesting). One of the great things about working at SimScale is to see every day how our customers meet these challenges by designing extraordinary products across multiple industry verticals.
Simulations can be of great help to handle these challenges by providing early insights into product performance to make informed design decisions with less prototyping and testing overhead. However making the most out of the simulations in a project is sometimes not straightforward – therefore keep the following (arguably “corny”) life lessons in mind.
1. “Life punishes those who delay”
One advice often heard from simulation people is “Simulate early, simulate often”. It’s not only an advice that helps simulation people keep their job, but actually often helps to reach a better result in a design project. Early design changes are cheap, so why not running a few simulations with some rough design concepts to make sure the design project goes into the right direction early on? The image below shows the creation of a design variant of an early cyclone separator design (here done with Onshape).
Simulating these rough concepts is not only significantly easier (fewer details, less CAD de-featuring is needed), but also helps you to explore all possibilities and not to “oversee” a good concept.
2. “Data informed” ≠ “Data driven”
Before performing simulations it’s helpful to ask the question “What would I do differently if I would have the simulation results right now”. This helps to concentrate on what actually is important in the simulation you are planning to do. For example if you are only interested in the numerical value of the pressure drop across a valve you can leave out a lot of the 3D result processing in the end and directly make sure the simulation outputs the actual numerical value. In the end, simulation is not an end to itself but a tool to help you build better products, so it’s important that the result data drives your design and that it’s not just beautiful to look at (even though it oftentimes definitely is…).
3. “Don’t use a sledgehammer to crack the walnut!”
Once you are clear on what kind of information you need from the simulation, try to use the easiest and fastest way that gives you “good enough” results first. At a later point in the design project, running a fully fledged detailed analysis of you application might make sense. But early in the project, it might be more valuable to run fast and iterative smaller simulations.
A good example I often see are cooling applications where a solid body (e.g. a casing) is being cooled via an air or a water flow. Before running a full conjugate heat flow analysis that gives you the exact heat transfer coefficient across the solid-fluid interface, why not starting with a constant heat transfer coefficient from a book and just simulating the heat transfer within the solid? This kind of simulation is by far easier and will help you to get the rough concept right (e.g. number of fins, casing material etc.). Then pick up “the sledgehammer” for the final detailed analysis to be sure the design meets the requirements.
4. “Crawl, Walk, Run”
In case you “need the sledgehammer”, meaning you need to face a more complex simulation, it makes sense to “think big, but start small”. Trying to get the full simulation right in the first step is oftentimes hard to achieve. Start with something simpler e.g. a coarser mesh or even a simpler geometry that gives you faster turn-around times and makes “simulation debugging” easier.
The image above shows a first simulation done for a butterfly valve design project. It has been set up as a steady-state simulation, so just looking at the equilibrium or “final” state of the flow through the valve. This means time-dependent effects such as transient wakes are not captured correctly. But the computing time was just 29 minutes on 8 cores so it’s great for testing if the overall simulation setup including mesh and geometry actually works without using a lot of computing time. Then in the next step, a real transient analysis was set up that took almost 10 times longer to compute, but captures all flow effects which are shown in the image below. One can now clearly see the transient flow behavior downstream of the valve.
5.”Doing the things right” ≠ “Doing the right things”
Translating this phrase to the world of simulation, it would probably be “Solving the equations right” ≠ “Solving the right equations”. So it’s not only important that the simulation tool you are using solves the equations well but also that the user chooses a good strategy to simulate his application. Which simulation approach you are choosing can impact the amount of work and time necessary to get the results as well as their accuracy. To choose a good simulation approach, make sure your application matches the assumptions of the solver you are planning to use (e.g. “No large displacements”, “Mach number below 0.3”, …) and ideally don’t start your simulation setup from scratch, but use another successful simulation as a template. With the SimScale Public Projects, we are trying to make this choice very easy for you.
Let us know how we could help you to make even more out of simulations for your design projects!