Study On Fault Diagnosis And Fault Source Identification For Cable Line In Power Distribution Network

With the continuous development of distribution network(DN) equipments in power system of our country, the distribution plants are required to develop towards land saved, energy saved, compact type, small-sized intelligentialize and unattended operation during the building and rebuilding of urban and rural power grids, therefore it is particularly important to study the online fault monitoring system for the DN. From the data published by the State Grid in recent years, the underground cables in the city are already in a high incidence of failure. It has an extremely important practical significance to study the fault location, fault line identification and fault classification for the safe and stable development of urban economy. The fault location, fault line identification and fault classification methods used in DN are more common with offline fault diagnosis for the moment, which are difficult to make fault prediction for the early intermittent failures. It is likely to cause permanent failure if the intermittent failures are not found timely, which will cause a great threat to the lives and property of customers. Based on this, the distinguishing features of the underground cable in city are intensive studied in this dissertation. According to the problems in the underground cable, the relevant electric quantities are selected reasonably. The related problems of fault location, fault line identification and fault classification are studied technically using advanced processing methods.In the DN fault location, with regard to the problems that how to measure the arrival time of the first fault wavefront accurately and the asynchrony problem for the two-terminal fault location algorithm, a 220V fault location experimental system in DN was built to study the actual loading fault location. The DN fault on-line location was realized with the methods of Empirical Mode Decomposition(EMD) and Wigner Ville Distribution(WVD), which have alos combined with the proposed two-terminal fault location algorithm without GPS. Firstly, the feature vectors are extracted from the transient signal after fault with EMD. Then, WVD is used to structure the instantaneous energy of the feature vectors,which can help to get the arrival time of the first fault wavefront accurately. Finally, according to the nonsynchronous time calculated by the line structure and characteristic parameters of the system, the fault position can be get with the two-terminal fault location algorithm based on the voltage crossing Point. The test and verify of simulations and experiments are made in ATP-EMTP and laboratory respectively, the results show that the algorithm proposed in this dissertation has a higher precision and a well application prospect in the practical engineering.In the DN fault line identification, with regard to the problems that it is hard to extract the weak transient signals after fault in neutral non-effectively grounded system, a fault line identification method was proposed based on EMD and WVD for neutral non-effectively grounded system. The time-frequency characteristics of the zero sequence transient current signal were extracted by EMD, which could realize the abruption between the high frequency transient signal and the low frequency interference signal according to the frequency distribution. The feature vector of the fault line identification could be described by calculating the instantaneous energy of the transient signal with WVD. The fault line identification method was formulated based on the correctness weight through the comparison of phase angle and the combination of zero-sequence current and instantaneous energy. The final fault line identification result will be got through fusing the 11 correctness weights in 2ms after fault. A 380V DN computational model with five feeders was built according to the actual circuit structure. The experimental verification was made in a actual fault feeder, the results of simulations and experiments show that the proposed approach has high accuracy, fast velocity and the correctness weight is defined reasonably.In the DN fault classification, a fault classification technique for neutral non-effectively grounded system using the binary ant colony algorithm(BACA) based fuzzy neural network(FNN), which could be used in neutral non-effectively grounded low-voltage and middle-voltage DN. The high-frequency component of the 2ms’transient signal after the failure was extracted by using EMD. Based on the method of mathematical statistics, the feature vectors of the inference system were structured. Then the change regulation of the feature vectors was researched under different fault conditions. BACA was used to optimize the weights of the FNN, which overcame the shortcomings of slow scouting speed and local minimum. The fault classification result can be got by fusing the feature vectors with BACA based FNN. A computational model was structured in ATP-EMTP based on the physical truth, then the inference system was trained with the computing results. The test samples and the results of field experiments show that the fault classification method proposed in th dissertation has high accuracy and a good adaptability.