Study Of Method For Hierarchical Control And Multi-agent In Air-source Heat Pump Hot-water System

As we know, the energy shortage issue is becoming more and more severe and it's necessary to make the task of energy savings, on this occasion, air-source heat pump interest's lots of businesses widely because of its unique advantages about high efficiency, energy-saving and safety. However, there are still issues in how to fully use air heat source and revise the set temperature automatically according to temperature. On this point, this paper focuses on the research of intelligent control theory, in the meanwhile, the fuzzy control, BP network and the hierarchical Multi-agent control theory are introduced into the air-source heat pump hot water system to resolve these issues and improve the intelligence of the system. The main content of this paper is in the following:According to the detailed circulation characteristics analysis for air-source heat pump, the hot water system was decomposed into sub-systems by hierarchical control structure and each sub-system was implemented by an agent with independent ability to solve the problems, which was to simplify the circuit control and change the complex control system to be simpler.Regarding the water temperature issue, the fuzzy control technology was introduced into the control for the temperature setting of water tank and a fuzzy controller was designed for water temperature setting to adjust control rules automatically by adaptable weighting factor, which the water temperature is set at range of [39℃, 52℃] automatically with the average of temperature and the variable of average temperature as input variables.It was proposed that the start-up time was adjusted moderately on the basis of the rate of hot water yield predicted and water task constraints to resolve the problem that the machine could not work in the best heating state because of unreasonable allocation about start-up time for machines. Based on BP network, predictor about the rate of hot water yield was designed through the research for BP network, and moreover, comparative analysis was accomplished between BP network with momentum factor but not dynamic learning rate, BP network with dynamic learning rate but not momentum factor and BP network with momentum factor and dynamic learning rate. Analysis showed that the BP network with momentum factor and dynamic learning rate had better prediction and it could satisfy the engineering need. The hot water system based on Multi-agent hierarchical control in air source heat pump resolved the present problem that could not make best use of air heat source and made hot water with fixed temperature, which would have better practice and economic utility for energy efficiency.