City Power Load Forecasting Technology And Its Application In Tongliao

Load forecasting in power system is significant for power dispatching, planning and corresponding department. Firstly, accurate load forecasting can help to ensure the security and stability of power system, thus, constrain the fluctuation of the load. Secondly, it is conducive to the power management, arranging the power network operation and maintenance plan reasonably, and reduce the unnecessary spinning reserve capacity. On the other hand, the reduction of the cost of power generation and improvement of the economic benefit and social benefit can achieve. Finally, the reconstruction of power grid, to arrange, the reasonable decision of construction and development of power grid can be optimized. Nowadays, the result of load forecasting has become the crucial discipline of the economic dispatch and the implementation of electricity market.In this thesis, the technique of data mining is applied, cluster analysis of the historical load data is performed, the load curve of the sample in the feature extraction, and then the characteristic curve for reference of checking and correcting abnormal load data, effectively avoid the influence of bad data caused by various mechanisms on prediction accuracy. Through the analysis of the industrial structure, economic character of Tongliao area, and combining with the customary routine, the parameter represents the busy degree of electric is obtained by means of the method similar to the analytic hierarchy process. Based on the Tongliao area population, GDP, and the use of intelligent electric busy index regression model by regression analysis and genetic algorithm, the extra consumption of the model predictions to remove the components caused by weather and holiday, the weather and date sensitive components firstly, decision tree classification, prediction of weather and date for sensitive components, the final realization of block fine prediction modeling of load.This thesis carries on the analysis to the load data of Tongliao area in 20141-6 month and realize the prediction of July every integral point load data, the results are fairly agreeable.