Simulation Inefficiencies: Quick Fixes for Common Slowdowns

October 30, 2024 Krystian Link

We often focus on reducing inefficiencies in product development through simulation. Whether it’s speeding up lengthy physical testing, minimizing design rework with strong conceptualization, or enhancing the quality of on-time launches.  One may wonder, are there inefficiencies in the simulation process too?  If so, what impact do they play on the simulation process, and what can we do to eliminate them?  Read on and find out about some quick and easy ways to fix them. 

Human Error.  Let’s think about how many button clicks there are to assign boundary conditions on over 100 HVAC system inlets. What if every other system inlet has a different temperature conditionWere all of the gas properties correctly? Are all 600 customer-required contours available and oriented correctly?  If that’s giving you some heart palpitations or you are saying to yourself “glad I only have three boundary conditions to worry about for my work”, there’s lots of opportunity to reduce risk (and human error) without resorting to late-night simulation double-checking.  A lot of Ansys software procedures can be automated: 

Do you have 1,000 different inlets that need to be changed to one new value, all at once?  Create a variable for it.  Change the variable value once, and it can be updated across your 1,000 inlets with a few clicks. 

Need to output three different contour plots every ten timesteps for 300 timestepsUse a journal file or leverage the Calculation Activities menu within the solver to automate various inputs and post-processing commands. 

Need to create a parameterized design of experiments (DOE)?  Utilize Ansys Workbench to automate the simulations as a batch. 

Resource Management.  Teams often have limited resources that must be shared, and the simulation team is no exception.  High-performance computing (HPC) resources, like powerful remote workstations or cloud clusters, are frequently shared across multiple physics disciplines Just ask any CFD or FEA engineer, and I’m sure they can name that one person who “is always hogging the resources”.  Joking aside, bottlenecks can ensue and even unhealthy rivalries can develop, causing a pile up of work and lack of team cohesion.  Here’s some ideas to get those tackled: 

Running overnight.  If you aren’t queueing simulations to run overnight, you are wasting resources that are sitting idle.  This becomes especially beneficial when your simulations take a solid 6-12 hours to run anyway.  Use the daylight hours to set up models, post process old models, and brainstorm new simulations to conduct, and use the evening hours to let the workstations and clusters to do the heavy lifting for you.  Spoiler: Many times, the demand of HPC resources is a LOT less in the evenings, so you could even be a little greedy and grab an extra 10, 20, or 50+ cores to run your simulation faster since no one else is using them.   

Communication.  If your team is not communicating with the remaining simulation teams (or team members), no one is going to have full visibility to who needs what resources when.  Visibility is key to maximizing your HPC resources, and that takes time and communication. 

Share.  You are not five years old, this is not the grade school playgroundYour coworkers are adults, and everyone needs to learn to play in the same sandboxTake some time to build camaraderie within your teamsAfter-work events, team building sessions, and learning about your teammates is critical to team cohesion.  Start seeing things from other peoples’ perspectivesWhen you know and respect the people you work with, you will be more inclined to tell yourself  “you know, my coworker was going through a tough time on that project and needed some extra resources to get it done, and that’s ok” versus emotionally reacting that “they stole my 40 cores”.  

Data Storage EtiquetteAll simulation files (from CAD to reports) should be stored in one place that is accessible to everyone.  This makes legacy knowledge more accessible to everyone and can potentially cut down on simulations in the future.  Consider data storage applications like Ansys Minerva to facilitate the end-to-end simulation data storage. 

Inconsistency.  Is your whole team performing the same simulation, using the same best practices, consistently the same way?  If your answer is “yes”, are you really sureWe are all human, meaning we can do the same thing in different waysThis can be problematic in simulation because those differences can materialize in the form of poorly converged models, incorrect solution setups, or incomplete data in final reports.  However, there’s a lot of ways these differences can be harnessed and leveraged internally for more robust simulation practices: 

Lessons-Learned or Closed Project Reviews.  In a fast-paced product development environment, these activities are often tabled or neglected as “they don’t add bottom-line value”.  Quite the contrary, in that it gives your team the time to learn from each other.  If there were hangups or errors that occurred during the project, this environment allows engineers to not only learn from their past mistakes, but also teaches less-experienced engineers ways in which to avoid common pitfalls.  Additionally, this environment allows the differences in your team members’ thought processes to really shine.  There actually might be a way to do XYZ simulation faster because your coworker did it in a former position, or you collectively attack a common problem to create a faster standard work procedure. 

Standard Work Procedures.  If you don’t have them, make them.  More likely than not, you will be doing the same simulation more than once, and you will have new hires needing to learn how to do that simulation.  With standard work procedures, you take the guesswork out of meshing and turn simulation setup into a point-and-click affair.  Everyone is doing it the same way, everyone is getting the same result.  For a deeper dive into standard work procedures, read more here. 

While these aren’t an exhaustive list of simulation inefficiencies, they are the most common and effective trouble areas to remedy.  If you aren’t sure where to start or are looking for some coaching on where else you can pull some inefficiencies from your simulation process, reach out to us.  We can scope custom engagements and mentoring opportunities to help address what you’re looking to improve. 

About the Author

Krystian Link

Krystian is a CFD application engineer at RandSim with over 10 years of product development experience in the automotive and manufacturing industries. His simulation experience focuses on vehicle thermal management, external aerodynamics simulations, and HVAC systems, including a publication in SAE’s Journal of Commercial Vehicles ("CFD Windshield Deicing Simulations for Commercial Vehicle Applications"). After completing his MBA at Indiana University’s Kelley School of Business, Krystian became even more passionate about building and implementing strategic solutions that not only address customers’ simulation needs, but also their business goals.

Follow on Linkedin More Content by Krystian Link

No Previous Articles

Next Flipbook
Rand SIM Enablement Services
Rand SIM Enablement Services

Rand Simulation has a variety of ways to help your team get going with analysis software and processes.

×

Have Questions?
Just Ask.

First Name:
Last Name:
Organization:
Country
Comments:
Thank you!
Error - something went wrong!