Research On Power Load Forecasting Based On Operation And Control Platform Of Distribution Network

With the continuous development of big data technology,it will be an important topic to use big data technology to forecast power load in the future.Based on the management and control platform of distribution network operation and maintenance,this paper studies the load forecasting system based on big data distributed computing,constructs the corresponding load forecasting model by using the historical load data of users,big data,machine learning,data mining and other technologies,analyzes the power consumption characteristics of users one by one and carries out user load forecasting,and finally accumulates the load forecasting data of users to get the load forecasting of the whole network Result.The results show that the model can improve the speed and accuracy of load forecasting.The main research contents are as follows.1.Based on the hardware foundation of distribution network operation and maintenance management and control platform,a user side load forecasting algorithm based on machine learning and deep learning of distributed computing technology is proposed,which not only achieves fast computing speed,stable and reliable system,but also achieves high accuracy of results,and flexible algorithm debugging.2.Build a big data distributed computing platform based on Hadoop and spark framework,and design an architecture system including data absorption,distributed system,algorithm model,component integration and service application,combining with the analysis of the use scenario,functional requirements and performance requirements of load forecasting.3.Design and implement the load big data prediction system based on deep learning,using the mainstream big data distributed computing technology such as Hadoop and spark,and programming language such as Java and scala to solve the problem of distributed parallel computing of load prediction,and improve the speed and accuracy of load prediction.4.Design and implement the load forecast visualization module based on the distribution network operation and maintenance management and control platform,with convenient user interaction mode,reduce the user's use cost and operation complexity.The module is developed based on B / S structure,using Java Script,HTML,CSS and other programming languages to support multi-user concurrent browser access and remote access.