Voltage Security Monitoring and Control in Electric Power Transmission Systems

• Main Authors: Chen-Sung Chang, 張振松
• Other Authors: Chin-Wen Lin
• Format: Others
• Language: zh-TW
• Published: 1999
• Online Access: http://ndltd.ncl.edu.tw/handle/89714250356547463214

博士 === 國立臺灣大學 === 電機工程學研究所 === 88 === With the increasing loading of existing power transmission systems, the problems of voltage security have become the major concerns in power networks planning and operation. The real-time voltage security monitoring and control are two main tasks to handle the voltage security problems. First, this dissertation presented a potential algorithm based on Continuation Power Flow (CPFLOW) to compute all the type-1 low-voltage solutions which are used as voltage security index. Type-1 solutions are of a single positive real-part eigenvalue associated with Jacobian of load flow equations and are used in conjunction with energy methods techniques to be as voltage security index for assessing voltage security of power systems. The benefits of the proposed algorithm are the following. First, the algorithm has the potential to find all the type-1 low-voltage solutions by tracing a small number of manifolds. Secondly, traditional methods, which can locate some of the type-1 solutions, suffer from the uncertainty that there might be another type-1 solutions which might be more suitable for voltage security assessment. This uncertainty is eliminated if all the type-1 solutions are located. The proposed algorithm has been tested for two example systems, and encouraging results have been obtained. Secondly, with the advent of phasor measurement units (PMUs), the on-line monitoring of the voltage security has become an possibility. The dissertation used a novel neuro-fuzzy network, Fuzzy Hyperrectangular Composite Neural Network (FHRCNN) for voltage security monitoring (VSM) using synchronized phasor measurement units as the input patterns. In this dissertation, it has demonstrated how neuro-fuzzy networks can be constructed off-line and utilized on-line for monitoring voltage security. The neuro-fuzzy network is tested on 3,000 simulated data form randomly generated operating conditions on the IEEE 30-bus power system to indicate its high classification rate for voltage security monitoring. From simulation results, it is observed that FHRCNN has a high classification rate over 97% and has better performance than traditional feed-forward artificial neural network, and FHRCNN combined with PMUs has better performance than FHRCNN combined with SCADA system. Lastly, this dissertation presents a two-stage computational algorithm for enhancing voltage security. The algorithm consists of the FHRCNN used in the first stage and the Genetic Algorithm (GA) used in the second stage, to monitor voltage security and dispatch reactive power sources under various system conditions. Based on the output of the FHRCNN, if the voltage security is neither very secure nor secure, the voltage security improvement stage is triggered. The GA is used to dispatch the reactive power sources and regulating devices to maintain the specified security level. The reactive power sources used in the proposed genetic algorithm are static capacitors, SVCs, transformer tap changers and generator terminal voltages. The proposed method has been tested on IEEE 30-bus power system with different cases under heavy loading conditions, and successful results have been obtained. The test results have proved that the proposed method can improve the voltage security and prevent the voltage collapse.