Model predictive control of a large temperature difference refrigerating station with ice cold thermal energy storage

Abstract

In this study, a model predictive control (MPC) algorithm is developed to optimize the operation of a large temperature difference refrigerating station with external-melt ice cold thermal energy storage (CTES) for the cooling system of a large office building in Beijing, China. The chillers and ice CTES equipment are connected in series to cool the chilled water so that large temperature difference and ultra low temperature air supply are realized to achieve small flow for energy saving. This study involves two phases: development of the mathematical model of the plant» development and testing of model predictive control algorithms. A co-simulation environment based on Matlab and TRNSYS is developed. The thermal behavior of the ice tank and plate heat ex-changer is analyzed by lumped formulas of the conservation equations in Matlab while models of other equipment such as chillers, pumps and cooling towers are developed in TRNSYS. The ice CTES system is modeled as a hybrid system, where the water phase transitions (solid-melting-liquid) are described by combining continuous and discrete dynamics. The model predictive control algorithm is designed and developed in Matlab to determine the required ice tank outlet water temperature to minimize operating expenses while meeting equipment and thermal comfort constraints under varying loads. The controller is optimized based on a forward neural network dynamic optimization algorithm. The objective function is a constant coefficient quadratic cost function of the ice tank outlet water temperature response and the use of control effort. This paper validates and compares the advantages of the MPC over PID. The findings illustrate that the MPC achieves shorter regulation time and smaller overshoot when compared with PID.