The Ice Storage Air Conditioning System Operation Optimization Research Based On Load Forecasting

In the summer, because of building air conditioning system daily operation of the power users the mall is one of the important factors that cause the fluctuation of power grid peak valley, and ice storage technique is an effective way to solve this problem. However, the existing operation of ice storage air conditioning system is often restricted by the backward operation strategy with a inaccurate realization of peak "move" of ice storage air conditioning, which hardly achieves the purpose of the energy consumption and reduction of operating costs and also can't optimize the ice storage air conditioning system. To solve this problem, this topic of paper is to establish the ice storage air conditioning cooling load prediction model through the control and optimization strategy, scientific allocation unit refrigeration with ice storage any groove cooling capacity to meet the demand of large stores the cooling load and achieve the goal of saving energy and reducing consumption and costs.First of all, based on the xi 'an seg international shopping center, working principle of the ice storage system and the common control strategy are analyzed in order to provides training and testing samples to subsequent cooling load prediction model. For the same aim, the running data of shopping center mall air conditioning load in summer is collected and analyzed.Secondly, according to the characteristics of ice storage air conditioning seg international shopping center, the daily and hourly demand of the ice storage cold air conditioning is forecasted by using the basic principle of extreme learning machine and the network structure to build cooling load prediction model. And the stability of prediction increase through the study of incremental type extreme learning machine(I-ELM) air conditioning load prediction model. Compared with traditional extreme learning machine, the root mean square error of incremental type extreme learning machine is less than 13.6% and the relative error reduced by 11.8%. In addition, the incremental type extreme learning machine cooling load prediction model have a good generalization ability and forecasting accuracy and also can obtain obvious energy saving and economic operation, which provide a new method to the optimization of ice storage air conditioning system cooling mode.Finally, the cooling load and the number of cooling machine running under different load types are calculated depending on the quantity of cold, local electricity price policy and cold load forecasting results, and xi 'an meteorological conditions and operation features of seg international shopping center. According to the data of daily cooling load in xi 'an seg international shopping center from July to November, the most common optimizing schemes of ice storage air conditioning cooling load, which are 100%?75%?40%?30%, are offered and the number of cooling machine working is arranged scientifically, which achieve the most energy save of ice storage air conditioning.