摘要
为减少变电站噪声污染,针对变压器噪声控制问题,提出一种基于遗传小波神经网络的变电站内变压器噪声自适应抑制方法。首先,将变压器噪声进行小波神经网络建模,比较变压器实际噪声信号和模型输出噪声信号的大小。其次,根据残余噪声信号幅值绝对值,自适应选择遗传算法或者梯度下降算法作为小波神经网络中参数迭代的优化方法。最后,利用一种降噪综合性能评价策略,确定模型隐含层最优结构。通过3种不同模型的仿真,结果表明遗传小波神经网络模型对变压器附近的噪声信号有较好的抑制效果。
In order to reduce the noise pollution of substation, an adaptive noise control method based on genetic wavelet neural network is proposed. Firstly, the transformer noise is modeled by wavelet neural network, and the actual noise signal and the output noise signal of the model are compared. Secondly, according to the absolute value of residual noise signal amplitude, the genetic algorithm or gradient descent algorithm is adaptively selected as the optimization method of parameter iteration in wavelet neural network. Finally, a comprehensive performance evaluation strategy of noise reduction is used to determine the optimal structure of the hidden layer of the model. Through the simulation of three different models, the results show that the genetic wavelet neural network model has a good suppression effect on the noise signal near the transformer.
作者
姜鸿羽
刘松
周健
朱官健
王珂
施志强
JIANG Hongyu;LIU Song;ZHOU Jian;ZHU Guanjian;WANG Ke;SHI Zhiqiang(Huaian Power Supply Company,State Grid Jiangsu Electric Power Co.,Ltd.,Huaian 223022,China)
出处
《电力科学与工程》
2020年第4期25-31,共7页
Electric Power Science and Engineering
关键词
变压器噪声
遗传算法
小波神经网络
降噪综合性能评价
自适应抑制
transformer noise
genetic algorithm
radial basis wavelet neural network
comprehensive performance evaluation of noise reduction system
adaptive suppression
