Can't find a trimmed operating point to a Simulink Model

4 views (last 30 days)
Hi there,
I'm using the Linear System Analysis toolbox to linearize a 7DOF Simulink vehicle model. The model inputs are steer angle and throttle, and the model outputs 4 wheels speeds with long, lat and yaw velocities.
When I try to trim the model to achieve a certain long speed, the optimisation routine struggles to find a trimmed state to within an acceptable level of accuracy.
Is there any way to diagnose which parts of the model are causing the problems?
I have removed all triggered subsystems and any other discontinuous features which may cause problems...
AJ

Answers (1)

Arkadiy Turevskiy
Arkadiy Turevskiy on 8 May 2013
A couple of ideas:
1. When you do the trimming, you click "Specifications" to specify which model states should be at steady state. By default the tool assumes you want all model states to be at steady state (and so all the boxes in the "steady state" column are checked). Often times, you want to uncheck some of the states. For example if your model has a state for a vehicle position, you don't want it to be at steady state (you want constant speed, not position). So recheck that.
2. Sometimes you get better results with a different optimization algorithm. Click on "Optimization Options" and try a few different optimization methods and algorithms. Maybe one of them will converge.
3. Another way to trim a model is to take a simulation snapshot. If you can run your Simulink model to a large simulation time where it comes to a steady state, you can compute a simulation snapshot at a large simulation time. You can either use it as is, or you can initialize the model with it and try trimming again- as you will be starting from a much closer initial condition, this time it may converge.
Hope this helps.
Arkadiy

Categories

Find more on Linearization in Help Center and File Exchange

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!