A New Model Reduction Technique for the Design of Controller by Using Moment Matching Algorithm

Abstract

A new system abatement scheme is proposed for the abatement of the complexity of the large-scale continuous time dynamic plants. In this abatement technique, a generalized version of the standard pole clustering process is applied for the computation of reduced denominator coefficients of the abated model, and the coefficients of the abated numerator polynomial are calculated by using the improved Padé approximation method (IPAM). The generalized pole clustering algorithm (GPCA) ensures the retention of important characteristics of the large-scale systems such as stability and dominant poles in the abated system. The various performance error indices are computed to validate the proposed algorithm. The least performance error indices values indicate the superiority of the algorithm. The proposed scheme is implemented on several standard systems, and it is giving the least error indices values. After computing the abated model of the large-scale system, the transfer function of the abated system is used for the construction of PID and lead/lag compensators by applying a moment matching method. The controller constructed with the help of the abated model is applied to the actual higher-order plant, and it shows that the response of the closed-loop system of the actual model is completely overlapping with the response of the given reference plant. The proposed technique is compared with some standard and recent methods of reduced-order modeling. Further, the time responses are plotted and time-domain specifications are computed for showing the usefulness of the proposed scheme.