Data-Driven Approaches for Fuzzy Prediction of Temperature Variations in Heat Exchanger Process

This paper addresses and compares data-based modeling approaches to approximate distributed dynamics of heat exchange process through universal fuzzy approximators. Clustering and swarm optimization methods are used to design nonlinear models to predict hot and cold fluids temperatures over a wide operating range. Experimental data is used to validate the identified fuzzy approximators. The performance of each data-driven fuzzy model is evaluated on both training and testing measurement data.