Research On Real-time Management Technology Of Data Quality For Power Grid Time-series Data

The grid time series data is an important basis for power plant equipment condition monitoring and fault diagnosis,which requires higher real-time performance.However,the existing methods of data quality management mostly focus on the historical data,which is difficult to meet the requirements of real-time,and the methods and frameworks are difficult to deal with large-scale data sets because of their own problems.In this paper,we focus on the research of real-time data quality control technology for grid time series data,aiming at studying the distributed real-time computing system Storm,and combining it with time sequence analysis and data cleaning to solve the real-time governance for large-scale data sets.First of all,deeply research the principle and method of time series analysis,and then explain.characteristics of ARIMA and the time series data of smart grid;Then analyze the method of data quality control to find the suitable method for this paper,and study the possibility of combining it with time series analysis;Based on the above research,analyze the requirements of real-time data quality control of power grid timing data,and proposes the overall architecture of data quality real-time management framework for grid time series data and the storage mode.Finally,the real-time concurrency control is designed for the massive time-series data source.Firstly,the time series data are forecasted,and the isolated points are identified in real-time.The cause of isolated points is analyzed Finally,three aspects including prediction precision,computing speed and resources occupancy are chosen to verify the effectiveness and practicability of the proposed framework by authentic cases.