| In general, the design of the HVAC system is based on the assumption that the system is working with full loads, but in fact the HVAC system is usually working with part of the loads and low efficiency. Refrigerating machines are the biggest energy consumer in the HVAC system, so it is very important to improve efficiency of refrigerator machines, to reduce the HVAC system energy cost for the optimal control of refrigerating machine and to match the refrigeration output with load.This paper designed and implemented an HVAC control system experiment platform based on the principles of large-scale refrigeration system, and it used ADVANTECH series multifunctional data acquisition-cards and control-cards to build up the hardware control system. Its software platform consisted of Real-Time-Target tool-box in MATLAB, which can achieve data acquisition, data control and MATLAB integration.HVAC systems generally have long time delays and nonlinear and time-varying characteristics. In order to solve the difficult multi-variable control problem of HVAC systems, and reduce the energy cost of HVAC systems, this control system offers an experiment platform based on which experiments of advanced control strategies and validations can be conducted.In the refrigeration optimization study, analysis of the impact of the compressor operating efficiency factors, based on BP neural network, this paper established optimal suction pressure model with refrigerant temperature at compressor inlet, the discharged temperature and load as inputs, and suction pressure as the output. We solved the refrigerator optimal control problem under variable loads, by adjusting online the suction pressure on which the refrigerator works. Experiments have shown that the refrigerator can achieve the optimum operating conditions on the optimal suction pressure determined by the model. COP, the performance parameter of refrigerator, is also greatly improved and the energy cost of the refrigerator is reduced. |
PDF Full Text Request