Short-term Load Forecasting Of Typical Low-voltage Station Based On BP Neural Network

With the development of the national economy and the increasing requirements of residents for the reliability of electricity consumption,the vigorous promotion of the construction and transformation of the distribution network has also exposed some weak links in the distribution low-voltage station area,and it is impossible to accurately and effectively grasp the stage.The law of load change in the area has greatly slowed down the upgrading of power distribution planning and construction,which has seriously affected the development of the power market.Therefore,it is necessary to develop the load forecast of the typical low-voltage station area of the distribution network.In this paper,the short-term load forecasting model and algorithm of low-voltage station area are established by analyzing the load characteristics of typical low-voltage station area,and a set of low-voltage distribution network planning auxiliary software system is developed.The work of the thesis mainly includes:(1)The load characteristics of the low-voltage platform area of distribution network considering many factors are analyzed.Firstly,the basic concept of the low-voltage platform area of distribution network is expounded,and the influence of climatic factors,economic factors,time factors,voltage factors and other factors on the load characteristics of low-voltage platform area is analyzed.Taking into account that meteorological conditions are the main factors influencing short-term load forecasting in low-pressure distribution networks,further analysis of the load characteristics of low-pressure stations under complex weather conditions,mainly analysis of the effects of temperature and rainfall on load characteristics,Finally,the typical load characteristics of low pressure platform in years,months and days are analyzed.(2)The load prediction algorithm of typical low pressure platform is studied.Considering that there are some errors in load collection in the low pressure station area,data preprocessing of load data in distribution network is carried out first.Then we study the shortcomings of traditional BP neural networks in the load forecasting of low-pressure stations,and put forward the method of bacterial chemotaxis to train and optimize the initial value of BP neural networks,and establish a model of BP neural network load prediction algorithm based on bacterial chemotaxis optimization.A load forecasting model of low pressure platform in distribution network based on improved BP is constructed.(3)The short-term load forecasting of typical low-voltage stations is studied.Firstly,the research on the evaluation criteria of the load forecasting results of the low-voltage station area,including the error analysis of the load forecasting results,the univariate analysis and the qualification rate analysis;then the short-term load forecasting analysis of the low-voltage station area of the distribution network is carried out,respectively,under working days and rest days.The low-voltage station area is used for load forecasting.According to the prediction results,the advantages and disadvantages of traditional and improved BP neural network algorithms are analyzed.The effectiveness of the improved algorithm is verified.Finally,the application of load forecasting in low-voltage distribution network is analyzed.(4)Based on the analysis of the load characteristics of the low-voltage distribution network and the research and results analysis of the load forecasting algorithm,the development of the auxiliary decision-making system during the planning of the low-voltage distribution network was carried out.Firstly,the overall design scheme of the assistant decision-making system is analyzed,and the key technologies in the development process are briefly introduced.Then the system development is carried out to build the overall architecture of the system,including the hardware and software architecture.Finally,the interface and function description of the results system are explained.To demonstrate its application and characteristics in practice,indicating the effectiveness of the auxiliary decision system.