I developed the system dynamics model for my problem (optimizaion of heating in a commercial building using a boiler, heat pump and a thermal storage) and the objective function with relative constraints.
I would like to know if there are available toolbox able to handle optimal control problems with non linear dynamics and time varying parameters (weather data: External temperature, Radiation).
Thank you very much for your help
Dear Ettore,
''ACADO Toolkit'' or ''YALMIP Matlab Toolbox'' may be helpful.
You can ask your question in YALMIP google group.
https://groups.google.com/forum/?nomobile=true#!forum/yalmip
http://acado.github.io/
Hope to help you.
No doubt that you can find a toolbox for that. But I would suggest linearizing your system. Thermal modeling of a building can be modelled as a linear system. It still may be time varying, but there are techniques for handling LPV systems. You may run your nonlinear system to take some data (input output) and then use ident toolbox of matlab to fit a linear model to your system (probably at each operating point).
Thank you everybody for your answers! Sorry Hossein Sartipizadeh for the deleyed answer, however the non linear dynamics results from the behavior of the stratification of a water tank thermal storage and the behavior of the air source heat pump COP
Dear Ettore,
I hope you have figured it out by now, but if you think its better to use the nonlinear model, you may take a look to 1) EDMC (extended DMC) structure which linearizes the nonlinear model at each step. 2) "fmin" of matlab that uses random searching to solve nonlinear optimizations.
3) There is a recent work on "successive convexification" that is used for this purpose. At each time instant, it convexifies the non-convex optimization for some iterations until it converges.
I hope it helps
- Badges
- Science topic