Heat Load Prediction Based On Neural Network And Operation Optimization In Electric Boiler With Heat Accumulator

The electric power storage technology is an important technical measure that can remove peak load and fill valley load, optimize resource allotment and protect ecological environment. As a concrete realization of this technology, the electric boiler system with heat accumulator has been extensively used due to the stimulus of time-of-use electricity price policy. During the heat accumulator supplying to users, the electric boiler usually try to make heat by starting at low price of electric and stopping at high.In general, the system of electric boiler with heat accumulator runs according to experience to utilize the time-of-use electricity price or controlled by the signals of water level or temperature. But the two ways are not ideal to realize the economical operation without full use of the policy of time-of-use electricity price. Actually, known the distribution of electric price and demand of heat load versus time, an optimizing supply curve can be drawn, that is, the concrete economically running policy can be made. But it is difficult to define the heat load in advance. After studying the prediction method and considering the complex, random and nonlinear factors that affect the demand load of heating, the ANN technology is adopted. Different from the general analysis in technology and economy, it is for the first time to combine the prediction in method of artificial neutral network with optimization in use of dynamic planning principle for the running analysis of the electric boiler. This paper tries to establish a CWL (climate-weekday-load) model to predict the heat demand load of users. Based on this and associated with the policy of time-of-use electricity price at present time and its future tendency, a much more economical decision can be made for this system using the optimization method. This paper tries to establish a CWL (climate-weekday-load) model to predict the heat demand load of users. Based on this and associated with the policy of time-sharing charge at present time and its future tendency, a much more economical decision can be made for this system using theoptimization method.In addition, the problem of selecting the neuron number of implicit layer in the network model and the problem of normalization of input-output vectors are discussed. During the operation optimization, the model of dynamic planning is established according to the feature of this problem and the advanced simplex method is used for resolution with the concrete algorithm provided.Finally, Application program design is realized by hybrid programming with the tools of Visual Basic, Access and MATLAB. It is also showed the result helpful for operating control and economical analysis. The result is helpful for the operation optimization of the system of electric boiler with heat accumulator and the popularization of the electric power storage technology, which will bring the achievement of comprehensive profit.