Toeplitz-MUSIC algorithm-based optimization scheme for UHV/EHV acoustic source localization

To address the problems of weak anti-interference and large positioning error of the traditional 2D-MUSIC algorithm under low signal-to-noise ratio (SNR) in partial discharge (PD) localization of power transformer bushings, this paper proposes an optimization scheme combining an eight-element cross-shaped array and Toeplitz preprocessing. Innovations: (1) Design an eight-element cross-shaped array balancing 2D spatial sampling and on-site compact deployment; (2) Propose block-wise Toeplitz preprocessing for the covariance matrix to enhance the orthogonality between signal and noise subspaces; (3) Construct a localization system adapted to UHV/EHV scenarios, integrating filtering and synchronous sampling modules. Compared with the traditional 2D-MUSIC and IVMD-MUSIC algorithms, verification shows that within SNR = − 20–20 dB, the root mean square error (RMSE) of the proposed algorithm is ≤ 1.8°, with a maximum correct localization probability of 98%. In the needle-plate discharge experiment, the RMSE at SNR = − 20 dB is 73.1% lower than that of the traditional algorithm and 54.8% lower than that of IVMD-MUSIC, with a single positioning time of only 21.05 ms. This scheme balances accuracy, stability and real-time performance, providing an engineering feasible solution for high-precision PD localization of transformer bushings in complex noise environments.