A PARAMETER ADAPTIVE CONTROL ALGORITHM FOR DIRECT DIGITAL CONTROL

A parameter adaptive control algorithm incorporating an extended version of the recursive least squares (RLS) parameter estimation method within the Dahlin control algorithm has been developed. The parameter estimation method has been expanded to include the tracking of a time-varying dead time as well as other process parameters. Dead time estimation is achieved by the calculation of process models with different dead times simultaneously, and then employing within the control algorithm that model which best approximates the process dynamics with respect to the sums of the squares of the errors of the different models. A resetting procedure for the parameter estimation method has also been implemented in order to accelerate convergence of the parameter estimates.;The resulting adaptive control algorithm was evaluated based on its ability to provide the satisfactory and consistent control of processes despite uncontrollable process parameter changes. Digital simulations of both linear and nonlinear systems showed that the new algorithm can offer good control. In each experiment, once the desired control response was established, the process parameters were drifted to new steady state values. Controller performance returned to its initial state following automatic returning to account for system changes. Results obtained with the nonlinear systems were nearly as good as those of the linear processes.