Analysis of substation data for knowledge extraction

The quantity and complexity of the data generated by microprocessor based relays have been enhanced beyond the immediate operational requirements. Following a serious incident, the amount of data received in a control center may be excessive and impossible to handle. The network operators have to interpret the data and rapidly apply the most appropriate remedial actions. Due to the emotional and psychological stress, they may not respond to the critical condition adequately. Any overlooks or mistakes can damage expensive power equipment or worse lead to a major blackout. Human operators are unlikely to obtain a thorough overview of an event when the amount of data is significant. What they require is useful information, not raw data. The idea is to enhance the capability of substation informatics and gain insight into the useful information (or knowledge) contained in a large and complex dataset. The paper investigates a new soft computing technique based on rough sets to extract knowledge from the large quantity of data captured by protection IEDs (intelligent electronic devices). The formulated methodology is generic and applicable to any types of transmission and distribution substations. Unlike device based data analysis such as the expert analysis of single disturbance files, the algorithm performs system wide analysis and is capable of handling data from multiple sources.