First Principles and Data-Driven Modeling of a Plate Heat Exchanger

Abstract

A plate heat exchanger was measured in counter-flow and parallel-flow arrangement in a wide range of mass-flows. The experimental data was used to estimate parameters of a mathematical model based on first principles. Furthermore, two data-based models were estimated using a nonlinear ARX type and a nonlinear Hammerstein type of model structure. All three models consider the dynamic (transient) behaviour of the system to derive the outflow temperatures of the fluids. A validation on independent measurement data sets shows that all models perform reasonable, each having individual benefits in terms of the considered application.