Validating a Product Costs Money and Time…Surprise.
Saying that simulation has come a long way is an understatement. My first encounter with Ansys Fluent was as an intern in 2010, and I remember hearing how some complex flow simulations were taking a day to run. By the time I was in the chair full-time doing analyses myself in 2013, the same simulation took four hours. More computing horsepower, between GPUs, CPUs, server clusters, and cloud computing, made extremely long duration simulations in 2012 the span of a coffee break today. So, the new question becomes, if the product’s validation plan only calls for a simple steady state simulation, why not include a transient case (or five) as well? If I have all this computation power, why not use it? The answer to this question is simple: Every simulation has a cost/benefit analysis.
Business Thermodynamics: What You Put Into a Simulation Does Not Equal What You Get Out Of It
Simulation engineers provide tremendous value when they have a fundamental understanding of their simulation’s impact on the product and its business implications. For instance, all products should have a design roadmap with the associated costs, activities, and timing of various milestones documented. Naturally, analysis should be a milestone in this plan. Unfortunately, the timeline is often either compressed or rushed, so the analysis team must maximize the ROI (return on investment) their simulations provide. For example, investing a few hours into a steady state simulation can be a “quick and dirty” way to get some sound design guidance, but it won’t be enough to ensure a product can pass an industry standard operating condition such as a cyclic heating/cooling cycle. Conversely, a transient simulation can address these unique time-dependent simulation conditions at the cost of days or weeks of runtime.
Worst of all, this operating condition may have a less than a 5% chance of being seen in the “real world” by a product. Is it prudent to simulate the steady state case, the transient case, or both? That’s a question you, your team, and the project manager should be discussing. It is the simulation engineer’s responsibility to utilize the computational horsepower they have appropriately. Don’t get me wrong, I love running high-core count simulations in thirty minutes as much as the next person, but just because you can run a simulation, doesn’t mean you should.
Each simulation costs time and money, and you need to ensure it answers the design question you need to understand to launch a successful product. Not doing so wastes company resources and risks additional project costs and launch delays. Choose the right simulations that answer your core design needs, and you can save tremendous amounts of development cost and time. The first and simplest decision in this process is determining if the simulation should be steady state or transient. In this article, we can break down what the difference between a steady state or transient simulation is, why it’s important, and how you can optimize your computation time for your specific case.
Steady State Simulations: Design Guidance for 80% of the Problems
Steady state simulations are not time dependent. This means that your CFD application is solving your system’s equilibrium point. You may remember this from your undergraduate studies as the First Law of Thermodynamics (“Stuff in equals stuff out” as my classmate would say). These simulations are faster to run on fewer computational resources, consequently reducing overall costs of the run. Most customers develop and grow their simulation skillsets conducting steady state simulations because the learning curve is comparatively low to transient simulations.
On top of that, steady state simulations offer a sizable amount of design insights when developing your product in a small amount of time. Whether it’s a clean sheet design or a refreshed design, doing steady state simulations builds confidence in your product and your own engineering capabilities. You develop your engineering intuition and begin to “know” how future products might behave because of these simulations and may not need to run as many simulations next time. Some common questions that are perfect for steady state simulations are:
- What does the flow or temperature field look like?
- Does the product exceed a certain temperature or flow rate?
- How can I get a rough idea on how this product behaves?
Notice that these questions aren’t necessarily “high resolution”. They fundamentally address if a product design is worth pursuing. At the beginning of the product development process, many design engineers have a handful of designs they think will perform but need to reduce that list down to one or two that will become their primary designs. Steady state simulations can rule out the poor designs quickly and focus company resources (you) on the most important design(s) that will launch the highest quality product in the shortest amount of time. But what it can’t do is answer how your product behaves over a certain duration of time.
Transient Simulations: When You Care About the Journey More Than the Destination
After running a few steady state simulations, maybe you’ve determined the best design path. However, your product still may need to meet certain certifications or standards dictated by your industry. These may take the form of ensuring certain components don’t fail after a certain amount of time or ensuring your product performs under various loading conditions. Transient simulations are time dependent. This means any of the following:
- Inputs (boundary conditions) are changing with time
- Inputs (boundary conditions) are constant, and my simulation runs for a set duration
- Inputs (boundary conditions) are changing with time and my simulation runs for a set duration
This is important to understand because steady state simulations are all about the “destination” (what things look like when everything is balanced). Transient simulations are about the “journey” of your product over time. How a product performs over a period of time can be drastically different than how it performs in a steady state operating condition. You may be able to determine your product survives but may need to understand how long it can run beyond that survival point.
Many industries have standards or certifications that require a product to operate in a specific manner over a certain period of time, which is where transient simulations can help. However, because of the time-intensive nature of these simulations, they can take two to fifty times as long as steady state simulations. Transient simulations also require a greater understanding of the physics involved and ensuring the simulation setup represents this behavior appropriately. This is a sizable investment to answer a very specific question.
If you are just trying to determine how your design might perform at a high level, you should not be using a transient simulation. If your validation plan or design failure mode effects analysis (DFMEA) does not have any time dependent criteria, and you/your team don’t anticipate time-related failures in your product, spend those computational resources elsewhere on products that can benefit from additional analysis.
So how do you know you need transient simulations? Here are some questions that would drive a need for them:
- Does my component fail after X minutes?
- How quickly can my product achieve a certain outcome?
- Will my product survive cyclic testing or various loading scenarios?
- Does the certification or internal design criteria call for any time-dependent behavior?
In these cases, notice the questions are narrow in scope compared to the steady state simulation prompts above. The product must do something over a certain period of time. Things are changing in these simulations. One key point to remember is that if you run a transient simulation long enough, it could very well end up looking like a steady-state simulation. That’s why understanding what data you need to get out of the simulation and the physics you are trying to simulate is critical. The launch of your product, while meeting both timing and budget constraints, depends on it.
Make sure to Contact Us today if you're looking for guidance on your next simulation project or if you're interested in how to take your team to the next level!
About the AuthorFollow on Linkedin More Content by Krystian Link