Abstract:
Background Solid-state pulsed power sources serve as indispensable front-end driving units in high-power microwave systems, and their core component—high-voltage nanosecond pulsed power generators operating under harsh high-voltage and ultrafast-pulse conditions—urgently requires a reliable, safe, and convenient testing approach to mitigate system failure risks and ensure stable performance of the entire microwave system.
Purpose To address this demand, this study establishes the physical mapping relationship between the fault modes of such generators and their corresponding acoustic characteristics, and proposes a non-contact fault diagnosis method based on audio features for safe, non-invasive condition evaluation.
Methods Acoustic signals emitted by the pulsed power generator under various working states (including normal operation and typical fault scenarios) are collected to construct a dedicated experimental dataset, and feature filtering and fault classification are performed on the extracted acoustic features to achieve precise fault identification without direct physical contact with the high-voltage hardware.
Results Experimental validation confirms the effectiveness of the proposed method, with performance testing demonstrating a fault classification accuracy of over 92% across all targeted fault categories.
Conclusions This study provides a novel technical pathway for fault detection of high-voltage nanosecond pulsed power generators, and lays a solid foundation for further research on acoustic intelligent diagnosis of solid-state pulsed power systems.
下载: