论文部分内容阅读
在分析了重采样方法和重采样滤波器应满足的频域特性后,提出了利用神经网络法设计有限冲激响应(FIR)重采样低通滤波器的新方法。这种方法能从频域设计具有任意幅值、相位的FIR滤波器,从而能根据前置低通滤波器的特性,对差动电流误差进行补偿,并减小这一误差,提高差动保护的可靠性。最后,建立了电力系统变压器差动保护的简化仿真模型,验证了所述理论的正确性。
After analyzing the frequency domain characteristics that should be satisfied by the resampling method and the resampling filter, a new method of designing finite impulse response (FIR) resampling low-pass filter using neural network is proposed. This method can be designed from the frequency domain with arbitrary amplitude and phase of the FIR filter, which can be based on the characteristics of the pre-low-pass filter to compensate for the differential current error and reduce this error and improve the differential protection The reliability. Finally, a simplified simulation model of differential protection of power system transformer is established, which verifies the correctness of the theory.