基于神经网络的串联谐振耐压试验方法研究

Research on the Method of Series Resonance Withstand Voltage Test Based on Neural Network

采用变频串联谐振设备对各类大中型变压器进行交流耐压试验可以有效解决传统耐压设备额定电流不足的问题。但串联谐振试验需要选择合适的电抗器和补偿电容,以保证试验在工频条件下进行。采用神经网络的计算方法,根据试品的电容自动计算出电抗器和补偿电容的组合方案,可以有效减少计算量,提高计算精度,为试验提供便利。

Using variable frequency series resonant equipment to conduct AC withstand voltage tests on various large and medium-sized transformers can effectively solve the problem of insufficient rated current of traditional voltage equipment.But the series resonance test needs to choose a suitable reactor and compensation capacitor to ensure that the test is carried out under power frequency conditions.This article will use the neural network calculation method to automatically calculate the combination of reactor and compensation capacitor based on the capacitance of the sample,effectively reducing the amount of calculation,improving the calculation accuracy,and providing convenience for the test.