Over decades PID controller has proven to be a very useful instrument in industrial sites. The generality of PID controllers allows easier design and tuning compared to other complicated controllers in addition to excellent control performance, and enables field engineers to operate them relatively easily. However, despite these advantages, conventional linear PID controllers display a conflicting relationship: a fast response requires large gains which, in turn, gives rise to a large overshoot. There is a tradeoff between fast response speed and less overshoot in actual applications
This thesis presents a nonlinear PID controller that can enhance the tracking performance of the conventional linear PID controller to achieve a desirable fast response with low overshoot. This is performed by introducing a new type of nonlinearities in the controller gains that are time-varying functions in terms of the error and/or error rate. Then, the parameters of the nonlinear PID controller are optimally tuned by a real-coded genetic algorithms (RCGA) such as the integral of time-weighted absolute error (ITAE) performance index is minimized.
A set of simulation works performed on four systems shows the feasibility of using the proposed method.