Distributed dynamics approximation through lumping and fuzzy clustering : reliability and validity issues

Abstract

This paper presents a fuzzy reasoning based approach for modeling the distributed dynamics of a parallel flow heat exchanger. Distributed-parameter and lumped-parameter models of the heat exchanger are first derived. The validity of such model representations is discussed and limitations regarding their use for practical purposes are presented. The alternative modeling approach presented in this work aims at developing a moderately complex model with interpretable structure for the parallel flow heat exchanger which is the main component of a pilot thermal plant. The proposed multivariable fuzzy rule-based model achieves simultaneous prediction of the air and water temperatures. The experimental study conducted on the real plant shows the performance of the fuzzy model in capturing the key dynamical properties of the physical plant over a wide operating range and under varying dynamics. Copyright © (2011) by International Operational Modal Analysis Conference (IOMAC) All rights reserved