The Fault Location And Analysis Method Of Distribution Network Based On Advanced D-S Evidence Theory

With the development of economy and the improvement of people’s living standards, electricity consumption both in urban and rural is increasing, and user’s demand for power quality and reliability of power supply is enhanced. Against this background, quality and reliability of power supply has been a key factor. Facing user directly, distribution network is a key link in process of controlling and ensuring the quality and reliability of power supply. Therefore, when power failure occurs, how to locate fault quickly, make emergency repair timely, and restore power supply becomes very important.In some distribution networks with high automation level, a large amount of fault information can be obtained when fault occurs because a large number of intelligent monitoring terminals are installed. In this situation, the traditional fault location method can be applied to locate fault accurately. However, in most areas where distribution networks with lower automation level, fault indicators with remote sign function while not FTUs are installed in distribution network. In general case, the fault indicators are installed in outdoor, the working condition of which is rather harsh, which may cause information of fault indicators aberrant or lost. In this situation, the application of traditional fault location method may cause misjudgment, thus maintenance workers cannot locate the fault area quickly, and the speed of repairing and the reliability of power supply are decreased.Aiming at above problem, this paper proposes a novel fault location method based on multi-source information fusion. The multi-source information includes information of fault indicators, distribution transformer and the trouble calls. Firstly, build the hierarchical matrix respectively by multi-kinds fault information including fault indicators information, transformers alarm information and trouble calls information which are caused by distribution network fault. Then, locate fault respectively through hierarchical matrix. Finally, obtain the final location result by fusing the location results of hierarchical matrix through D-S evidence theory. This method is applicable not only to the fault location of the distribution network with lower automation level, but also to the network with high automation level, which combines the information from intelligent monitoring terminals, leading to a more accurate location result. Specific research content is as folio wings:1. Distribution network fault information under fault condition, and relationships between fault information and fault location are analyzed. The problem existing in fault location when the fault information occur identifiable distort and unidentifiable distort is analyzed. Based on above analysis, a novel distribution network fault location method based on hierarchical matrix search and improved D-S evidence theory is then proposed.2. According to the distribution network fault information including information of fault indicators, distribution transformers and trouble calls, and their corresponding relationship with the fault location, a novel fault location method based on hierarchical matrix search is proposed. Firstly, the basic concept of hierarchical matrix and the method of formation of the network hierarchical matrix were put forward. Secondly, a suspicious fault zone library is established. Thirdly, fault areas are obtained by searching the hierarchical matrix and contrasting with suspicious library fault zone. Finally, the method of fault location by information of fault indicators, transformers and trouble calls is discussed in detail under the examples of looped distribution network and radial distribution network.3. According to the independence of the information of fault indicators, transformers and trouble calls, a distribution network fault analysis method based on advanced D-S evidence theory and new decision criteria are proposed. Firstly, take fault indicators, distribution transformer and trouble calls as sensors. Secondly, collect the fault information and obtain fault zone respectively by hierarchical matrix search algorithm. Thirdly, form original evidences according to the fault zone, modify the evidences, and then fuse the new evidences by D-S evidence theory. Finally, set decision criterion, and obtain the final localization results according to the decision criterion. The effectiveness of the location method is verified by four typical examples.This method solves the fault location problem through multiple information fusion when misinformation or information lost happens, and greatly improves the fault tolerance of fault location of distribution network, which makes it easier for maintenance workers locating fault area quickly, and improves the speed of repairing and the reliability of power supply.