Research On Smart Grid Big Data Related Application Problems

Massive data is indispensable for the safe and stable operation of power systems. Withthe rapid informatization and digitization of smart grid, massive data will be produced andaccumulated in the entire process of electric power operation and management, coveringgeneration, transmission, transformation, distribution, consumption, dispatching and utilitymanagement. Data acquisition, transmission, storage, processing and application will cost alot. How to convert data burden to wealth and extract data value to the greatest degree iscrucial to the construction of strong smart grid. This thesis mainly explored and discussed thesmart grid big data related application problems, including power information predictionusing multi-variety data and smart grid data visualization based on XML.Firstly, the thesis briefly introduces the background and significance of smart grid bigdata(SGBD). The important role acted by SGBD in the construction and operation of strongsmart grid is emphasized.Secondly, the sources and features of SGBD are generally summarized. CIM and XML,which are two important data formats in SGBD, are simply outlined. Both the value basis andgeneral processing procedure of SGBD are also summarized.Thirdly, the author discusses the processing technology of SGBD. Heterogeneous dataintegration, unstructured data processing and data extension are considered as keytechnologies of SGBD. The implementation thought for each aspect is discussed.The thesis mainly focuses on two SGBD applications. They are smart grid datavisualization based on XML and electric power information prediction using multi-varietydata respectively. A data visualization and exhibition scheme under the scenario of SGBD isproposed. XML is employed to integrate SGBD resources. A SVG document will begenerated for graphical display, while a HTML document for table display through XMLdevelopment. The feasibility and ease of proposed scheme are testified by the generation anddisplay of power system wiring diagram and power system event data report. Furthermore,the studied cases of both annual power consumption prediction and short-term loadforecasting demonstrate that the full exploitation of data variety and the tolerance of diversitywill efficiently help to improve the accuracy of prediction. It has been proved that even thesimple mathematical model and calculation method can obtain accurate prediction results.