基于信号子空间转换与快速子空间测向算法的局部放电超声阵列信号测向方法

An Orientation Method for Ultrasonic Array Signals of Partials Discharge Based on Signal Subspace Transform and Fast Subspace Estimation of Direction of Arrival

提出一种基于信号子空间转换法(signal subspacetransform,SST)与快速子空间测向算法(fast subspaceestimation of DOA,FDOA)的局放超声阵列信号高精度测向新方法。首先利用SST算法对局放超声阵列信号进行聚焦处理,使得原始信号的宽频空域信息被最大限度地保留,从而实现宽带信号方位信息的累积。然后采用FDOA对聚焦后的窄带信号进行波达方向估计,FDOA无需特征分解,无需估计整个协方差矩阵,可提高运算速度,且具有更高的测向精度。在此基础之上,应用医学宽频超声信号和4 4超声阵列传感器模型,进行局放超声阵列信号测向仿真研究,仿真结果表明,测向误差小于2,验证了该方法的正确性。

A new high-precision method to orient the ultrasonic array signals of partial discharge in power transformer, which is based on the combination of signal subspace transform （SST） with fast subspace estimation of direction-of-arrival （FDOA）, is proposed. Firstly, the SST algorithm is used to focus the ultrasonic array signals of PD to reserve broadband spatial information of original signals to greatest extent to cumulate the position information of broadband signals; then FDOA is used to estimate the direction-of-arrival （DOA） of focused narrowband signals to speed up the calculation because FDOA need not eigen-decomposition, so only sub-matrix is to be estimated but not the whole covariance matrix, and a higher precision of orientation can be achieved. On this basis, the simulative research on orientation of ultrasonic array signals of partial discharge is carried out by ultrasonic wideband signals often used in medical supersonics and 4×4 plane ultrasonic array sensor model. Simulation results show that the orientation error is less than 2°, thus the correctness of the proposed method is verified.