Hybrid TDOA/DOA Multiple PD Localization Using Two Arrays With KPCA-mnc-FastICA and Sparrow Search Algorithm

Abstract

Time difference of arrival (TDOA) and direction of arrival (DOA) are, respectively, applied to detect partial discharge (PD) target with ultrasonic signal. The reason why these two methods cannot be applied at the same time is that the required parameters, such as arrival time differences and directions, cannot be extracted simultaneously. This article solves the problem of calculating these two parameters simultaneously, and then proposes the hybrid TDOA/DOA localization with two $2\times 2$ sensor arrays in PD localization first. This innovation can help PD positioning to obtain better accuracy and lower hardware requirement. Our previous research, which is kernel principal component analysis modified noncircular fast independent component analysis (KPCA-mnc-FastICA), is further studied to calculate arrival time difference and direction of multiple sources simultaneously. To tackle the problem of slow calculation speed, this article explores a combined method of hybrid TDOA/DOA positioning model and sparrow search algorithm (SSA) with two arrays which is sufficient to meet the requirement of multiple PD localization. The spatial accuracy of the discharge model measurement in the laboratory and the numerous PD signals mixed simulation experiment are supported by empirical evidence. The outcomes demonstrate the suggested method’s superior performance in terms of hardware reduction and multiple target localization.