Optimization of Supply Air Parameters Control Based on Gappy POD Method for Creating Non-Uniform Temperature Fields

Abstract

This paper explores the implementation and application of non-uniform ambient fields from two perspectives: constant heat source and abrupt heat source variation. On the one hand, the proper orthogonal decomposition (POD) method is used to find the optimum air supply parameters for different target temperatures at multiple demand points. The one-parameter, two-parameter and three-parameter cases were considered, respectively, and the parameters obtained from the search for optimisation were verified, yielding mean deviations of 0.405 K, 0.368 K and 0.380 K and mean errors of 1.48%, 1.61% and 1.68%, respectively. The accuracy of the reconstructed results of the POD method and the reliability of the POD method for finding the best results are verified step by step with the help of the experimental platform. The validation results show that the average error between the reconstructed data and the experimental data for the POD method does not exceed 5%; the average errors between the measured and set demand temperatures at the target point are 1.2% and 0.8%. On the other hand, the gappy POD method is used to accurately fill in the elements of the system with arbitrary missing data and to reconstruct the flow field in the presence of missing (gappy) data with a limited number of sensors combined with POD. The errors of the gappy POD method reconstruction are 0.54% and 1.75%. POD and gappy POD methods can better create non-uniform temperature fields in practical scenarios according to actual needs. The results of the study can provide a methodological reference for real-time reconstruction and real-time control of the indoor flow field environment. It also serves as a help and suggestion for the actual project in terms of end-regulation system and reverse design method.