Transformer partial discharge location method based on time delay compensation and improved artificial lemming algorithm.

Abstract

To address the large partial discharge (PD) positioning errors caused by the inadequate modeling accuracy of traditional models for the complex internal structures of transformers, as well as the difficulty in solving strongly nonlinear positioning equations, this paper proposes a PD positioning method based on time delay compensation and an Improved Artificial Lemming Algorithm (IALA). First, the proportional correlation between the electromagnetic wave propagation delay and internal geometric features is revealed, and a correction model for the propagation delay is established to modify the positioning equations. Subsequently, Sobol sequence initialization and a dynamic multi-leader collaboration mechanism are introduced to overcome the defects of traditional algorithms, such as premature population convergence and blind search behavior. This enables an efficient solution to the aforementioned strongly nonlinear positioning equations. Furthermore, a complete positioning strategy of "IALA initial positioning-time delay determination and compensation-secondary precise positioning" is proposed. Finally, multi-point positioning experiments conducted on a physical transformer demonstrate that the proposed method effectively eliminates the severe multipath interference caused by the core and windings. Compared with traditional algorithms, the positioning error is reduced by ∼53%, significantly improving the spatial positioning accuracy of PD sources in transformers.