Optimal autotuning of nonminimum phase processes under relay control

Nonminimum phase processes with poles and/or zeros in the right half plane are always difficult to estimate and design a regulator assuring the better performance. In this article, exact analytical expressions of the time-domain output for a class of nonminimum phase processes are derived under a relay control system. Process parameters are assumed to be unknown before the test. A single symmetrical relay test is performed to estimate up to five unknown parameters of process model effectively. Based on the estimated process model, new optimum proportional–integral–derivative gains are obtained from a novel technique of evolutionary algorithm. A new control performance index has been defined to achieve a balanced tradeoff between transient performance and actuator preservation. Simulation result has proven the feasibility of the technique by showing the robustness in the presence of static load disturbances and measurement noise