Medium-long Term Power Load Forecasting Of A Region

Load forecasting is an important content of power system planning and operation researching, the premise to ensure reliable power supply and economic operation of power system, the basis for power system planning and construction. Accurate load forecasting is helpful to improve security and stability of power grid, reduce the cost of power generation effectively, satisfy electicity demand, and enhance the power supply reliability, so as to improve power system economic and social benefits.This paper analyzed the electric power system load forecasting method, classification, characteristics and influenced factors. A large number of data about A city urban area has been collected, such as electric power consumption, maximum load, regional gross domestic product etc.We have also investigated the urban geographical situation, natural conditions, the‘twelfth-five' plan, social development and power supply. Through the transverse and longitudinal analysis to the data, and comparison of the precision of forecast models and the actual results, this paper avoided the subjective factors by using quantitative forecasting technology, looked for proper load forecast model for the situation of A city urban area, and forecast its power load objectively and actually. Finally it was decided to adopt the regression analysis method, the neural network method and the elastic coefficient method for the medium-term load forecast of A city urban area. The forecasting results of the elastic coefficient method are similar with the forecasting results of the regression analysis method and the neural network method, which testified these two analyses can be used as the load forecasting method of A city urban area and the forecasting results have the certain accuracy. Then through these methods, we got three prediction schemes(i.e. high medium and low quality schemes) of electric power consumption and maximum load of A city urban area between 2011 and2015.The load density method for the long-term load forecast was adopted. In order to analyze it visually, we use the Auto CAD software to draw the load density distribution of A city urban area. Then we analyze the power and energy balance of A city urban area. Based on these analysis results, we pointed out that A city urban area needs to improve the structure of the power network and the reliability of the grid and the power supply. The results of this paper can be used as the reference of the transformation of the A city urban area distribution network, the electric power dispatching, the improvement of power supply quality, the maintenance plan and the generation scheduling.The results show that recursive right combination forecasting model is better than average right combination forecasting model and single model in terms ofimproving model accuracy, and applicable to medium-long term load forecasting.Finally, the article carries out programming calculation on modeling process, which greatly improve the prediction speed, and predicts the total load of A city region in 2015-2018 by using the above methods, which provide the basis for network planning of A city region.