Research On Thermal Load Forecasting And Optimization Of Integrated Energy System Based On Electro-thermal Energy Storage

With the rapid development of social economy and energy consumption,how to achieve energy efficient,reasonable and pollution-free use has become a major problem faced by many researchers.the operation process of the electric heating integrated energy system is analyzed and an accurate and dynamic thermal load forecasting method is proposed in this thesis.On this basis,from the perspective of system operation economy,the optimization model of the electro-thermal integrated energy system is established,and an improved cuckoo optimization algorithm is used to solve the output of each heat source,so that the heat supply of the heating system is matched with the heat demand of the user,and then realizing the purpose of on-demand heating,avoiding unnecessary energy waste in the operation of the integrated energy system,reducing operating costs.The main work of this paper is as follows:Firstly,aiming at the problem of heat load prediction,a preliminary prediction model of thermal load of integrated energy system based on RBF neural network is established.The model is not only considering the weather factors such as illumination intensity,outdoor temperature and wind speed,but also considering the inertia of the thermal system.The historical heat load value is included in the load influencing factors,which ensures that the prediction method is consistent with the actual physical state.Finally,combined with the actual example,the simulation tests are performed on MATLAB,and the preliminary neural network prediction value is obtained.The result shows that the prediction error is within ±20%,which paves the way for the following error estimation method research.Secondly,aiming at the problem of the prediction error of the thermal load prediction model based on RBF neural network is large,a fine prediction method combining RBF neural network and improved Markov chain is proposed.The method classifies the historical data by the K-means clustering method according to the outdoor temperature and the neural network prediction value,which has a great influence on the prediction accuracy.Each type of prediction error is modeled separately,and then different Markov state transition probability matrices are established.Then,the predicted time error value is obtained through data restoration to correct the RBF neural network prediction result.The simulation results show that the average relative error of the method is 1.32%,and the maximum relative error is 5.73%.Compared with the single RBF neural network method and the basic Markov chain correction method,it has higher prediction accuracy and can play a good reference role for the heat load distribution of the actual integrated energy system.Thirdly,aiming at the optimization control problem of integrated energy system,the optimal scheduling model of electro-thermal integrated energy system is established under the condition of predicted heat load demand value and the heat source output value are equal,including the cost of environmental pollution treatment.Then,the improved cuckoo algorithm is proposed to solve the model.The example shows that the operating cost of the system can be effectively reduced by scheduling the output of each heat source in the integrated energy system.At the same time,the speed of solving the optimization model can be improved significantly by the improved cuckoo algorithm,thus the correctness of the model and the effectiveness of the algorithm are verified.