I am planned to use nonlinear pid controller instead of linear, could anyone help me in this regard.
What type of nonlinearity you wont to introduce to PID?
Some nonliearities are already included in PID block of the Simulink, like output saturation and anti-windup.
If it is some variable gain than common blocks can be used, e.g. for Kp_var*error you could use multiplication block where one input would be error and other Kp_var. Kp_var than would be output of some other block that specifies how that gain will vary and which signals will it depend on. It can be output from some Fuzzy network and as result you would have Fuzzy-PID, or from Neural Network resultin in NN-PID, ore from scheduler with result in gain-scheduled-PID.
Same way can be done with other terms, where instead of error at multiplier input would go derivative of error (error -> d/dt block ->) or integral of error (error -> 1/s block ->).
If you have a state space representation of your controller you could use the Differential Equation Editor (DEE) to enter the equation directly. You can get the DEE block by typing "dee" in the Matlab command terminal.
If you are doing gain-scheduling PID you can certainly do that using the normal PID block plus the gain-scheduling variables. An easy way for gain-scheduling is to use the local controller network proposed by Hunt et al.
You can use also a S-function to enter the differentiel equation directely.
- Badges
- Science topic
- Similar topics
- Engineering
- Control Theory