Aggregate load parameter identification using general regression neural networks

Abstract

In this paper, the GRNN paradigm, as originally presented by Specht is summarised. Next, a G R " implementation of a static load model simulation is described, mnd its performance is d&uucd. A Generic Load Model is presented U a simple dynamic model suitable for wide-& implementation using a GRNN and integrating memsurement-based data. Finally, ongoing and future work are discussed, and preliminary conclusions are stated based on experience to date.The paper is application oriented. Its major contribution k the structuring of the load parameter identiflcation problem into a form conveniently handled by neural networks. Generation of a training set is facilitated by the use of a simplified dynamic equivalent load model whose parameters are easily identifled to form a training set from system measurements using least squarer procedures.Kqwu&-static dynamic load models, general regression neural networks.