Study On The Predition Method Of Multi-stage Load In Multi-scale Spatial Resolution

Spatial load forecasting(SLF)is the basis of urban power network planning and design.Accurate and reliable load forecasting results have important guiding significance for urban power network planning and design.Compared with the previous load forecasting,the SLF prediction results not only include the amplitude of the load,but also cover the spatial distribution of the load,so it can provide more accurate decision-making basis for the planning and operation of the power system.The process of urban power grid planning and operation can be divided into different stages,and different stages of the load forecasting process will often encounter different problems,and if only forecasting under a single load level can not meet all the requirements of power grid planning.The construction of a multi-scale analysis mechanism for power loads is an effective guarantee for the reliability of load forecasting.For this reason,this article aims at the forecasting needs of power load in different load forecasting stages,proposes a corresponding solution,and combines the multi-scale analysis mechanism of power load to improve the accuracy of load forecasting.The main contents are as follows:A method for determining the maximum value of the cell load based on the mixture Gaussian distribution is proposed to avoid the problem that the abnormal data in the cellular load will adversely affect the spatial load forecasting.Based on the analysis of historical load regularity,the actual measured data is analyzed.The statistical characteristics and distribution characteristics,the exploration of various abnormal data manifestations,combined with the comparison of actual load data and typical load data,identification of abnormal data,identification of abnormal data,and separation of abnormalities that adversely affect load prediction The data determines the reasonable maximum value for the development of the performance of the cellular load,and improves the accuracy and reliability of load forecasting.A method for predicting the saturation load of urban power grids based on principal component analysis and multivariate models is proposed,which avoids the problem of multi-variable models used in the load forecasting of urban power grids and the effect of multicollinearity between complex urbanization factors on prediction accuracy.Using the factors that affect the load changes,the principal component analysis method and multivariate model are used to characterize the development trend of power load,and Verhulst model is used to expand the scope of application of this method.This method can fully tap and use the connection between the load and the factors that affect the load change,instead of using the annual maximum data of the original power load to directly predict,and at the same time avoiding theproblem of multicollinearity between load-affecting load change factors and improving The accuracy and stability of the data reduces the adverse effects of the volatility of the annual maximum data of the power load on the forecast.A multi-level coordination method of SLF based on particle swarm optimization is proposed to solve the multi-level coordination problem of multi-scale spatial resolution power load.On the basis of multi-scale analysis of power load,the reasonable use of the maximum cell load is obtained.As a result,the unbalanced results between different load levels in the spatial load forecasting process were considered.Combined with the multi-level characteristics of power load at multi-scale spatial resolution,multi-scale spatial resolution was achieved using the particle swarm algorithm.Under the multi-level load coordination,and based on the results of multi-level coordination,the space load forecast was re-completed,and the more accurate spatial load forecasting results were obtained.