Research on operation strategy of ice cool storage air conditioning system based on load forecasting

 Taking the ice cool storage air conditioning system of an energy station in Chongqing as the research object, obtains that the actual operating performance deviation of the chiller and the plate heat exchanger is within 5% to meet the design requirements. In a one-week cycle, there are three types of user load—Monday type, Tuesday to Friday type and weekend type. Establishes a load forecasting model of BP neural network optimized by genetic algorithm with Matlab, and trains the model with test samples. The results show that the maximum relative errors are 16.5%, 7.6% and 13.9%, respectively. Through Matlab programming modeling, obtains the optimal operation control strategy at 100%, 75%, 50% and 25% load rates, respectively.