I have a model of a hydraulic pump and would like to get a best set of parameters of the components in my model to match the pressure data that I obtained from experiments.
In the optimization module, is there a way I can use my experimentally measured time-varying pressure data (from a spreadsheet, for example) to obtain the best fit model parameters?
Any help is greatly appreciated.
here below a simple example describing how you can use LMS Amesim optimization capabilities to set the parameters of a component with respect to experimental data.
Let us suppose that you want to tune the parameter of a component, here represented by a second order lag (LAG2), in order to match some experimental data read by the dynamic time table (SIGUDA01).
Using the parameter default values, the results are the following:
To set up the optimization, go to Settings --> Export setup.
1. In the first tab, you can define the optimizations input parameters, i.e. the natural frequency, damping ratio and the gain, just by drag and dropping them onto the export setup window. This is the trade-space.
2. The second tab allows you to define the simple output parameters which, in this case, are the results of the table component and of the second order lag.
3. Finally, the third tab enables you to create expressions based on the simple output parameters listed in the previous step. Here you specify the functions that will be minimized during the optimization. For this application, I tried to minimize the in the difference between the experimental and the simulated results, the global minimum and maximum
To run the optimization, go to Analysis --> Design exploration
1. Create a new optimization study
2 In the input tab, select the parameter targeted for the optimization. Provide boundary values if available.3. In the output tab, select the functions to be minimized. Configure the optimization method and run the optimization
5. At the end of the optimization, right click on the study and select “apply best results”
The new results obtained now match the experimental data:
For more detailed information about optimization, please refer to the active suspension demonstrator available in the Siemens PLM website or directly in the LMS Amesim documentation with the followin address (qthelp://lmsimagine.lab/ame_dir/doc/html/manuals/
I hope it helps,