Identification of large scale distributed parameter systems—some simplifications in the multidimensional Poisson moment functional (MDPMF) approach

Abstract

This paper is concerned with two practically important aspects of identification of large scale distributed parameter systems by the multidimensional Poisson moment functional (MDPMF) method (Saha and Prasada Rao 1980). The existing methods of (i) development and diagrammatic representations of the relations among the MDPMFs of partial derivative terms and those of the original multivariable function, and (ii) elimination of the effects of boundary functions and their partial derivatives from the identification equations, become extremely unwieldy as the order and dimensions of the model increase. This paper suggests alternative techniques in these two situations, leading to a simplification of the algorithm and an enormous reduction of the related analytical and computational burden. In the former situation, the suggested alternative uses the concept of separated variables. In the latter, the boundary function effects are estimated and then inserted into the algorithm to simplify it.