In recent years, with the rapid development of artificial intelligence technology, some artificial intelligence methods are gradually introduced into fault location, and some intelligent ranging methods appear, such as artificial neural network, fuzzy recognition, fuzzy recognition, fuzzy processing, simulated annealing algorithm, etc., among which artificial neural network is the most widely used. Because the neural network has excellent self-learning, self-adaptive ability, and has a certain fault tolerance ability, after a large number of sample training reflecting the fault distance characteristics, the neural network can approach the mathematical relationship between the input and output, and can carry out fault location without establishing accurate mathematical equations. The artificial neural network ranging method is less affected by the transition resistance, and can use a variety of fault types, and can ensure the accuracy of fault location under different operation modes, so the research on this ranging method has great practical value. There are still some problems in this method: firstly, it is difficult to train, because the training samples of neural network need to collect the data of various faults, which can not be obtained by experiments, only the historical data recorded in the case of power system faults can be used. Secondly, the training time of neural network is too long. Because the time that the power system allows to cut off the fault line is very short, the fault location time will not meet the actual requirements of the power grid with the increasing automation of the power system. To sum up, most of the artificial intelligence ranging methods are still in theoretical research, and have not been widely used in practical projects.<br>
正在翻译中..