| Recently considerable interest has developed in the area of designing controllers for non-linear and multi-variable systems. This is due to the occurrence of systems which are non-linear and interacting in the modern petrochemical, pharmaceutical and chemical industries. While numerous techniques are available for designing controllers, very few of them have gained operator acceptance due to the complex mathematics involved. Hence most industrial controllers for non-linear and interacting processes are based on PID routines for approximated or linearized model equations.;This dissertation discusses the development of the "Exact Inverse IMC Controller" algorithm for non-linear processes, "Reference System Decoupling" for multi-variable interacting systems, and an experimental application of "Q-PI/Q-PID" for on-line quality control. These algorithms are very simple and hence easily implementable. The algorithms have been successfully tested on a simulated pH system, a 3 x 3 multi-variable system, and experimental SISO systems respectively. The performance of each system has been compared with the contemporary controller design techniques available in literature. Excellent servo and regulatory behavior has been observed in all cases.;It is believed that the framework presented in this dissertation could have a significant impact in the area of real time process control. |
