how can we define the fuzzy PID controller rules for a time delay system?
Perhaps the book "Fuzzy Control Systems with Time-Delay and Stochastic Perturbation" by Wu Ligang, Su Xiaojie, and Shi Peng maybe helpful.
Why Fuzzy? Depends on your system. But, assuming you can produce a model of the system I would consider using a PI or other linear control approach combined with a Smith Predictor.
https://en.wikipedia.org/wiki/Smith_predictor
If your system is close to linear, at least around some working points, then fuzzy control obviously does not bring much benefits (and for me is a last choice). I would also recommend getting linearized model in order to make proper decision about control strategy. Try to extract so called normalized dead-time. Such parameter tells you if more complex strategies like Smiths Predictor or simple Model Predictive Control need to be employed. Model predictive control will be superior if your controller is often at saturation limits.
Finally, if there is some possibility to re-design the system you may think about replacing / adding sensors to reduce normalized dead-time. Adding sensors can then open doors to cascade control strategies.
- Badges
- Science topic