Intelligent Substation Relay Protection Condition Monitoring Based on Support Vector Machine Algorithm

Aiming at the fuzzy characteristics of the operating state of the relay protection equipment, the fuzzy comprehensive evaluation model in fuzzy mathematics is introduced as a means of quantifying the operation state of the relay protection equipment, so as to achieve an accurate evaluation and formulate a reasonable maintenance strategy. With the continuous improvement of power consumption stability, improving the reliability of the operation of the relay protection system has gradually become a key concern of researchers in the industry, and scientific and effective maintenance methods are one of the keys to ensure the normal operation of the relay protection system.. Whether the power equipment is running well or not will directly affect the reliability of the whole power system. A relay protection state evaluation method based on fuzzy SVM is proposed. Firstly, the historical data of relay protection devices in smart substations are obtained, and the appropriate state evaluation factors are selected. Then, the data of the state evaluation factors are preprocessed, and the percentage remaining life of the devices at each patrol point is calculated according to the state patrol and fault records, and the training sample set is generated. In order to obtain the battery performance online, the online monitoring method of substation battery performance based on Support vector machine (SVM) modeling technology is adopted, which combines the characteristics of AC impedance method and DC discharge method. Simulation results show that by introducing sample weight, the method can distinguish the difference between the sample of this device and the sample of the same model device, so that the algorithm can give consideration to the individuality and family commonality of the device, and has certain anti-noise ability.