Study And Application Based On Data Arehouse And Data Mining For Traction Power Supply SCADA System

The tremendous progress has achieved about the electric railway in our country. The length of electric railway has break through 20,000 kilometer to 2005. And at present it is pointing to the direction of high speed and over loading. In order to the security and reliability of the power supply, the requirement is increasingly enhancing to the SCADA of railway traction power supply (TPS). Massive quantities of accumulative date are stored in the dispatching center of the TPS SCADA system in Chinese railway, including real time data and history data. However, nowadays the data of current SCADA system for TPS is only used to fault logs, fault curve records and so on, which is typical operational data. The data are managed by database is mainly used to On-Line Transaction Process (OLTP). It's indispensable for SCADA system to build the integrative analysis system, which can provide the function, such as the analysis and diagnosis to malfunction device, data analysis after event and the evaluation to the protection switch' action.This paper only extends its advantage on basis of the SCADA function in existence and improves its analytical decision-making through the data warehouse and the data mining. Building the data warehouse (DW) in the dispatching center of TPS SCADA system for electric railway and making use of On-Line Analytical Process (OLAP) and data mining (DM), the malfunction state of the equipment of TPS SCADA system can be analyzed deeply, the useful decision-making regulations will be got, which provide ways and means of finding and forecasting the malfunction for us and shorten the check-up and stopping time.At first, this text analyses the deficiency of the current TPS SCADA system and point out the irrationality of the data use. Then it is put forward to improve ability of the decision-making and analysiswith the data warehouse and the data mining.In succession, the whole design on basis of the data mining and data warehouse is given in this text. Moreover, it provides the key steps, such as the establishment and the update of the data warehouse, the choice of the arithmetic and so on. In this section, the ways of data update event-oriented and timing-oriented are adopted;the clustering arithmetic FCM is imported to scatter the sequential data;the data processings such as extraction, cleaning and transformation are also mentioned;and then, the evaluation model is given based on the Association Rules Apriori arithmetic.Afterwards, the mul- dimension data models such as Star Schema and Snowflake Schema are used in this text .the data warehouse by means of Subject-Oriented method, on basis of the electrical devices of transformer substation (Normality — Abnormality — Malfunction — Fault Wave — Protection Action) is given, including the structure of the system, the update of the data, the design of fact table and dimension table. Then it provides the implementation about data warehouse throuhg Microsoft SQL Server 2000.In addition, the text analyzes the deficiency of the arithmetic — Apriori when it is applied in the TPS SCADA system. Then the corresponding improvements are put forward. The classifying of data warehouse, the time sequence pattern, the hiberarchy Apriori arithmetic are brought forward, though which the operation efficiency of the arithmetic is optimized.Finally, the thesis is summarized and the future work is proposed.