The Design And Implementation Of Electric Power Big Data Analysis And Forecast System

As the power companies are making great strides in promoting the use of smart electricity,users are demanding more and more quality requirements on the power delivered by the power system.The power load forecasting is the most important thing in ensuring power quality,directly affecting the macro-control of power grid resources.As well as production operations,it is the guarantee of the safety and stability of the power grid.Accurately,timely and rapidly power load forecasting can not only provide the basis for the power system dispatching department to develop the power generation plan,but also the basis of the scheduling and scheduling plan,the power supply plan and the trading plan in the market environment.Moreover,with the popularity of intelligent electricity collection system and the accuracy improvement of the weather forecast,comprehensive,massive electricity data and accurate weather information make it possible to do an accurate,timely,and rapid power load prediction.Therefore,analyzing and predicting the massive load data generated by smart electricity consumption accurately and timely,via big data technology,is of great significance to maximize the quality of planning.The key to power load forecasting system is the construction of load forecasting model.According to the documents of major journal articles,in the field of power load forecasting,the power load models proposed by experts and scholars from various countries are mature and have been successfully put into operation.However,these models all use a single mathematical algorithm,the model parameters are difficult to adjust in time with the changes in the data,and can not accurately describe the load changes,so the se algorithms are difficult to obtain a high accuracy of the prediction.And with the development of smart grid and the arrival of large data,the amount of forecast data is becoming larger and larger.The original processing platform encounters bottlenecks in processing such a large amount of data.Therefore,the power grid needs a load analysis and prediction system that can process massive data and accurately describe the toad change trend.Based on the above background,this paper deeply studies the application of Hadoop,spark big data processing platform and random forest algorithm in short-term power load.After getting a detailed understanding of the short-term power load forecasting business logic of the power grid,the user power data in the power consumption information collection system and user profile information in the marketing business application system are collected.Combined with external data on macroeconomics such as meteorological information,holiday information,industry recovery rate,industry expansion and installation information and industrial structure changes,a comprehensive analysis has been done on the factors of power load fluctuations load change trends and.various.Meanwhile,it analyses the advantages and disadvantages of the random forest algorithm,proposed a parallel integrated random forest prediction model based on the Spark platform,and tested the model using existing historical power load big data,compared with the original power load forecasting.The algorithm has a high accuracy of prediction and has the ability to process massive data.At the same time,the power load big data analysis and forecasting system is constructed using the spring MVC architecture model,and the results of the parallel random forest forecasting model are applied to the actual power load forecasting business.The system includes data input,data integration processing,system services,and user display.Develop and implement functions such as data acquisition,load analysis,model management,load forecasting,and orderly power usage.Compared with traditional load forecasting systems,operations are easier,data displays are more intuitive,and predictions are more timely,accurate,and faster.