Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.
Question on the analysis of contradictory results
Posted 6 feb 2017, 10:55 GMT-5 Fluid & Heat, Results & Visualization, Studies & Solvers Version 5.2a 2 Replies
Please login with a confirmed email address before reporting spam
Hi,
My simulation consists in passing a fluid in a path of variable form. However, for a speed "x", I get a 2D graph with a stationary study. But for the same speed, when I do a time-dependent study, the 2D graph that I get converges to a different solution.
I don't think that's normal. Could you help me understand why this is happening?
Moreover, to which study should I have the most confidence to analyze the results?
Thank you
My simulation consists in passing a fluid in a path of variable form. However, for a speed "x", I get a 2D graph with a stationary study. But for the same speed, when I do a time-dependent study, the 2D graph that I get converges to a different solution.
I don't think that's normal. Could you help me understand why this is happening?
Moreover, to which study should I have the most confidence to analyze the results?
Thank you
2 Replies Last Post 7 feb 2017, 08:41 GMT-5
Please login with a confirmed email address before reporting spam
Posted:
8 years ago
Hello Vincent,
Stationary nonlinear problems can have multiple solutions. When you solve them numerically, which one of those solutions a computer program converges to (if any) is going to depend on the algorithm used and any initial guess (in the COMSOL GUI, "Initial Values") you provide.
Transient problems are even more complex in their behaviors - This is the stuff that textbooks get written about.
If the history of the physical system is captured by your transient model's inputs, you should probably trust your transient model. But as always, you'll want to perform a few "sanity checks" on the results: do they make physical sense? Are the time steps and mesh appropriate to capture the physics, etc? A nonlinear transient problem can also have multiple limit points and small changes in the initial conditions or in the numerical methods used could result in convergence to a different limit (or to no limit).
Jeff
Stationary nonlinear problems can have multiple solutions. When you solve them numerically, which one of those solutions a computer program converges to (if any) is going to depend on the algorithm used and any initial guess (in the COMSOL GUI, "Initial Values") you provide.
Transient problems are even more complex in their behaviors - This is the stuff that textbooks get written about.
If the history of the physical system is captured by your transient model's inputs, you should probably trust your transient model. But as always, you'll want to perform a few "sanity checks" on the results: do they make physical sense? Are the time steps and mesh appropriate to capture the physics, etc? A nonlinear transient problem can also have multiple limit points and small changes in the initial conditions or in the numerical methods used could result in convergence to a different limit (or to no limit).
Jeff
Please login with a confirmed email address before reporting spam
Posted:
8 years ago
Thank you Jeff for this explication !
Vincent
Vincent