Research On Power Quality Data Analysis Based On Big Data Platform

With the development of smart grid and the complexity of power grid environment,the power quality problem is becoming more and more serious,so the demand for power quality data processing is also increasing.This paper mainly focuses on the following two questions:first,as the quantity of power quality data monitoring terminal increasing,and the sampling frequency speeding up,this produces a huge amount of power quality data,so the RDBMS has been unable to,deal with the storage and computing needs;Second,with the development of information technology,the value of the power quality data is reflected,it can not only reflect the operation situation of power grid in the past a period of time,more importantly,we can analyze and predict the future of power quality.For the above two questions,this paper sets up a big data platform for processing massive power quality data,and dedicated to analyzing and predicting the mass data of power quality in the background of electrified railway.The main research work is as follows:1.Aiming at the problem of large amount of power quality data,this paper build a power quality analysis system,implement the collection of the history and real-time data of massive power quality,and design a reasonable storage structure,provide support for the analysis of power quality.For the problem data,data preprocessing is carried out before the data is stored in HBase,and the method of missing data and outlier data is proposed,implement effective processing of problem data.2.Aiming at the business requirements of electrified railway,this paper puts forward the separation method of main and standby monitoring points by means of KMeans clustering algorithm,and the accuracy of the method is 100%.3.Aiming at different forecast demand of power quality data,this paper put forward two kinds of forecasting methods,firstly based on the time sequence correlation,using ARIMA trend prediction algorithm to implement the prediction of multiple time points;Secondly based on the correlation between active power and other indicator,using decision tree regression algorithm to implement the prediction of single time points..4.The prediction experiments in the background of electrified railway are carried out,and the experimental results show that the proposed method has good accuracy.