基于移动麦克风阵列的换流站异常声源定位方法

An approach for localizing abnormal sound sources in converter stations using mo⁃bile microphone arrays

换流变压器、交流滤波器等电力设备在运行时产生的声学信号中往往包含大量的状态信息,而传统依靠人工听诊判断设备故障位置的方式存在主观性强、效率低下、可靠性不高等问题。为此,提出一种基于移动麦克风阵列的异常声源定位方法。推导了短时傅里叶变换平移后的误差公式,由误差公式定量评估阵列移动速度与采样时间的合理范围;在频域下建立了移动麦克风与声源之间的位置函数关系,以新函数关系为基础优化了传统的波束形成算法;引入三角定位法,解决了波束形成算法只能判断方向的不足,实现了换流站内异常声源的准确定位。依据现场测试数据,搭建仿真模型验证了移动麦克风阵列定位异常声源位置的准确性。

The acoustic signals generated by power equipment such as converter transformers and AC filters tend to contain crucial status information.Traditional methods relying on manual auscultation to locate equipment faults suf⁃fer from subjectivity,low efficiency,and limited reliability.Therefore,a novel approach is proposed,utilizing mo⁃bile microphone arrays for abnormal sound source localization.The derivation of the error formula for the translated short-time Fourier transform(STFT)quantitatively assesses the reasonable range of array movement speed and sam⁃pling time.In the frequency domain,a positional function relationship between the mobile microphone array and the sound source is established.Building upon this new function relationship,the traditional beamforming algorithm is optimized.A triangulation method is introduced to address the limitation of the beamforming algorithm,which can only determine direction,enabling precise localization of abnormal sound sources in converter stations.Based on onsite test data,a simulation model is constructed to validate the accuracy of the approach for localizing abnormal sound sources in converter stations using mobile microphone arrays.