Co-simulation – Art or Science?

Practitioners of co-simulation engaging in conversation at a conference over their experiences and expertise could easily end up in a heated debate and complete disagreement over proposed methodologies. One thinks theirs is the best invention since sliced bread,while the other believes in their own. What is the reason for differing opinions and methodologies of co-simulation?

Let me get a bit deeper into the problem with an interesting anecdote. A research project wanted to analyze combustion instabilities in gas turbine engines by building a co-simulation between a CFD software and a combustion chemistry code. More than a year and million Euros spent later, a well-crafted co-simulator was available. However, the results did not make sense. Someone finally asked: can high-frequency combustion instabilities be analyzed by co-simulation? The answer: no, because cosimulation changes the stability properties of the physics.

In this article, I will uncover the reasons behind these expensive errors and different judgments, and give some guidance for best practices to get started on a successful use of co-simulation.