Distribution Network Fault Recovery Fuzzy Petri Net Algorithms

With the development of daily life and production, the customers require higher quality and reliability of a distribution network. To meet with the customers' requirements and reduce the loss of the faults of a distribution network in the power system, there comes the distribution automation system as a new and efficient means of management and technique. The practice proves that the distribution automation system can improve the working efficiency, the quality and reliability of power supply. This can bring benefits to both the customers and the power companies. One of the most important jobs of the distribution automation system is service restoration when there is fault in a distribution network.Based on available research fruits, this thesis makes deep investigation into service restoration, which includes:Rule-based heuristic inference method is a basic method in fault diagnosis and recovery of power system that also constitutes the basis of expert system, but is rather complex in the inference process. When rules have too many constraint conditions, a great amount of judgments will be involved. Based on Fuzzy Petri Net, this paper investigates knowledge representation under multiple constraint conditions and builds Fuzzy Petri Net model with general inference format. Based on this model, more efficient and simple formalized reasoning algorithm has been developed. This algorithm fully utilizes the parallel processing ability of Fuzzy Petri Net and can be realized with simple matrix operation.After permanent fault occurs to power distribution system, dispatchers need to change the operation status of switches accordingly, segregate permanent fault area and resume power supply to other areas without fault. Heuristic rules are often used in fault restoration scheme for required switch operations, but efficient or suitable scheme is often beyond human intelligence when system is complex. Using heuristic rules in determining restoration scheme by deployment personnel, this paper builds Fuzzy Petri Net model for making restoration scheme and denotes the restoration target as the fuzzy value of Token in Fuzzy Petri Net. The operation result of this Fuzzy Petri Net model will be suitable faultrecovery scheme. Simulation result indicates that when power distribution system is: complex, this model can replace human inference for optimal fault restoration scheme.Expert System that is the best fully-fledged discipline in the field of artificial intelligence has difficulty in establishing a complete knowledge base, and has a low efficiency. The paper has put forward a method that combines Fuzzy Petri Net and Expert System, and uses Fuzzy Petri Net as the reasoning mechanism of expert system. The proposed method has improved efficiency greatly and created conditions for restoring power supply as soon as possible.