Regional power distribution network load demand hybrid prediction method

Abstract

The invention relates to a regional power distribution network load demand hybrid prediction method. The method comprises the steps of data collection, data processing, prediction with different methods and comprehensive calculation of prediction results with different methods. Data preprocessing is to combine with collected data, and abnormal values, missing data and the like existing in data information are processed. The combined prediction method comprises three prediction methods of multiple linear regression, grey prediction and support vector machine prediction, and prediction deviations of different methods are calculated in combination with sample collection. The weight determination of the result of the combined prediction method is to finally determine the weights of different methods according to the prediction deviation by using a variable coefficient method. According to the combination method, interference and influences of irrelevant data and abnormal data are effectively avoided, meanwhile, the weight of the calculation result of each method is determined by combining the combination application of different methods and combining the corresponding prediction error effect, the accuracy degree of the prediction result can be effectively improved, and therefore accurate prediction of the power demand of the regional power distribution network is achieved.

Description

technical field [0001] The invention relates to the application field of load forecasting for regional distribution networks, in particular to a mixed forecasting method for regional distribution network load demand. Background technique [0002] The electric power industry is the basic industry of the national economy, and the scientificity and rationality of the investment strategy of the power grid enterprise is one of the core contents of the enterprise management strategy research. The current economic and social development and changes in the internal and external environment have brought severe challenges to the operation of power grid companies. Therefore, considering the increase in load demand, scientifically and rationally determining the investment scale and optimizing the investment strategy are the top priorities. Through systematic analysis combined with the current situation of regional development and different types of forecasting methods, the accurate dete...

AI Technical Summary

Problems solved by technology

With the introduction and application of deep learning algorithms, it can provide certain support and reference for the accurate prediction of load demand, but the current research depth is insufficient, the practicality is not strong, there are still problems such as low training efficiency and many iterations, and the current situation is not fully considered. The investment inertia of power grid enterprises does not fully consider the changing characteristics of load demand under different situations, and the accuracy of forecasting needs to be further strengthened

Images