摘要
针对线性窄带主动噪声控制(Linear narrowband active noise control,NANC)抑制变压器非线性噪声性能不佳的问题,提出一种基于Volterra和FIR组合滤波的变压器噪声非线性主动控制方法(Volterra and FIR filter-based nonlinear active noise control,VFNLANC)。该方法通过Volterra滤波器的谐波抑制特性处理变压器的非线性噪声,并将通过Volterra滤波器的滤波信号作为FIR滤波器的输入进行二次滤波,以缩小Volterra滤波器的截断误差。考虑变压器声道具有时变特性且存在扰动噪声,在VFNLANC方法中设计次级通道混合辨识架构,通过离线辨识生成初始通道参数并结合在线辨识实时更新参数,可既保证辨识精度又降低计算复杂度。采用实际变压器噪声数据对所提方法进行了仿真实验,结果显示VFNLANC方法的降噪量比NANC方法提升3.2~6 dB。
Aiming at the issue of inadequate performance in suppressing nonlinear noise in transformers using linear narrowband active noise control(NANC),a transformer noise nonlinear active control method based on Volterra and FIR combination filtering(VFNLANC)is proposed.This method utilizes the harmonic suppression characteristics of Volterra filters to handle the transformer′s nonlinear noise.The filtered signal from the Volterra filter is used as the input for a secondary FIR filtering,so as to reduce the truncation error of the Volterra filter.Considering the time-varying nature of the transformer′s acoustic channel and the presence of disturbance noise,the VFNLANC method incorporates a secondary channel hybrid identification structure.Through the offline identification of the structure,the initial channel parameters are generated.And combining it with real-time online identification,the continuous parameter is updated.Thus,the identification accuracy is ensured and the computational complexity is reduced.Simulations were conducted using actual transformer noise data to evaluate the proposed method.The results demonstrate that the noise reduction of the VFNLANC method is improved by 3.2-6 dB in comparison with the NANC method.
作者
李亚俊
孔维健
吴超
LI Yajun;KONG Weijian;WU Chao(College of Information Science and Technology,Donghua University,Shanghai 201620,China)
出处
《噪声与振动控制》
北大核心
2025年第5期41-45,91,共6页
Noise and Vibration Control
基金
国家自然科学基金资助项目(61603088,62173084)。
