Research On On-line Diagnosis Approach Of Power Grid With Multi Data Sources

Power system is a key national infrastructure and blackouts have great influence on the economy and national security. Blackouts caused by accidents can not be avoided completely. Hence, it is necessary to deeply perform researches on on-line fault diagnosis, analysis and monitoring approach of large-scale interconnection power grid, which is of great significance on improving operation reliability and reducing the probability of cascading failure and blackouts of large-scale interconnection power grid.A deep research was done in this dissertation from the aspects of information processing of multi data source during cascading failure and blackouts, the efficiency of fault diagnosis, novel fault diagnosis method, operation action diagnosis, optimum islanding places during emergency state and the development on-line fault diagnosis, analysis and monitoring system of large-scale interconnection power grid with the on-line fault diagnosis, analysis and monitoring of large-scale interconnection power grid during cascading failure and blackouts and the practical demand of power grid considered. Main research efforts of this dissertation are listed as follows.1. Complex event processing technology and service-oriented architecture were introduced in order to overcome the weak point in information processing of multi data source during normal state of power grid especially in the cascading failure and blackouts. Firstly, a basic event form of power system was presented. Secondly, the generating method of XML form event of power system was proposed based on the analysis of power system event form. Then, a complex event processing method of power system was introduced. Finally, a real-time information processing architecture and solution of multi data source was presented based on complex event processing technology and service-oriented architecture.2. There are several problems such as time-consuming diagnosing and difficulty in on-line expert knowledge maintenance in current power grid fault diagnosis system based on expert system especially in high level dispatching centre. A novel fault diagnosis approach of power grid based on knowledge grid technology was presented in this chapter for solving problems mentioned above. Firstly, knowledge grid technology was introduced to power system fault diagnosis field. Then the knowledge map for power system fault diagnosis was designed according to the background of real power system application. And the generation methods of each part in knowledge map were analysed in detail. Finally, a distinction between fault diagnosis system of power grid based on expert system and knowledge grid technology was contrasted with two practical applications test in power grid.3. Two main problems exist in current on-line fault diagnosis and analysis system for power grid. One is data source problem due to time delay for transferring information and asynchronous time of accepting data. The other is two key blind spots which exist in false-operation and fail-operation diagnosis. A novel fast fault diagnosis and analysis approach of power grid based on WAMS was presented in this chapter for solving problems mentioned above. Firstly, the architecture was designed for data service oriented and developed from multi-agent technology. Then the structure and methods of each agent were analysed in detail. Finally, it is shown by test results of disturbance experiment in a real power grid that this approach is fast and correct in fault diagnosis and analysis result.4. Some operation actions computed off-line are invalid during on-line operating. A novel operation action diagnosis method based on WAMS was proposed aiming at the problem mentioned avobe. Firstly, key point of the invalid operation action was deeply analysed. Secondly, an on-line operation action identifying method was introduced. Then, a novel static security index was presented. Finally, this novel operation action diagnosis method was testified by a practical case.5. Some islanding strategies may be useless during blackouts because they were computed off-line. A novel islanding decision-making approach was presented for solving the problem mentioned above. Firstly, the reason of ill islanding strategy was discussed. Secondly, a slow coherency generator group identifying method was proposed. Then, a novel islanding method based on graph theory was presented. Finally, it is shown by a simulation test with BPA programme that this method is good at islanding place choosing.Based on comprehensive investigation and engineering practice, an on-line fault diagnosis, analysis and monitoring system of large-scale interconnection power grid was presented in this dissertation. The system function scope was enlarged by the study on a fault diagnosis approach of power grid based on knowledge grid technology, a fast fault diagnosis and analysis approach of power grid based on WAMS, an operation action diagnosis method based on WAMS and an islanding decision-making approach, which will enhance the power failure processing performance of the power system.