Integral-based algorithm for parameter identification of the heat exchanger

Abstract

This paper develops an integral based method for the system identification and modeling of the heat exchanger. The proposed method formulates the model through the use of integrals, thus rendering a nonconvex optimization problem into simple linear optimization which can be solved through an ordinary least squares method. The convection heat exchanger model is firstly assumed, whereby all the inlet and outlet temperatures are measured or known. Two experiments were conducted. The first involved a series of step responses with incremental heating input, whilst the second involved step responses taken only after the system has cooled down to room temperature. Results show that the proposed method allowed a more complicated model of the heat exchanger to be constructed, whilst still being robust to Gaussian and quantization as well as encoder noises.