Simulation-Study of Adaptive First Order with Deadtime Compensators

Model-based deadtime compensators can be used to nullify the effect of deadtime on the closed-loop characteristic equation. However, changing process conditions can cause unstable behavior due to the parametric forcing. Here, single-loop on-line adaptive control is simulated with such compensators using the first order-plus-deadtime model. A least squares equation error formulation is considered for the parameter identification, including deadtime. First and second order sequential (recursive) gradient methods are tested. Thus, the computational burden is such that microprocessor implementation is feasible. Some simulation results are shown and discussed for on-line adaptative control in the face of rather severe combined parameter changes (forcing)