Field-deployable acoustic monitoring for high-voltage circuit breakers using equivalent STI-based measurement-region selection

Mechanical failures of high-voltage circuit breakers (HVCBs) can lead to substantial operational and economic risks. Acoustic monitoring has attracted growing attention recently as a non-contact approach capable of capturing mechanical-acoustic information during switching operations. However, in substations, background noise, reverberation, and unstable field acquisition conditions often result in low-quality and low-clarity acoustic signals, impacting the effectiveness of subsequent diagnostic analysis. To address this challenge, this paper proposes a field-deployable microphone-placement evaluation method based on a blind estimation of an equivalent Speech Transmission Index (STI), eliminating the need for impulse-response measurements that are impractical for HVCB transient operating sounds. First, near-field acoustic radiation during breaker operations is simulated to analyze frequency-dependent sound-pressure distribution, directivity and attenuation, and the surrounding space is partitioned into 6 representative regions. Second, an equivalent modulation-transfer representation is reconstructed directly from measured opening and closing waveforms to obtain estimated STI values under realistic field conditions. A ray-based acoustic simulation model is further constructed to obtain theoretical STI values and validate the blind estimation. The estimated STI values are then used to identify the recommended acoustic measurement region. Finally, measurements collected from the STI-recommended region show improved signal separability under blind source separation (BSS) and consistently higher fault-classification accuracy across multiple classifiers. The proposed method provides an objective and physically interpretable basis for sensor deployment in substation acoustic monitoring of HVCBs, improving the robustness of downstream diagnostic analytics.