Performance evaluation of a standalone PCA-based denoising method for Distributed Acoustic Sensing (DAS) data

This paper experimentally evaluates the effectiveness of Principal Component Analysis (PCA) for denoising distributed acoustic sensing (DAS) data. Experiments were conducted by applying different vibration strengths using a piezo-electric transducer (PZT) at various sensing locations along the sensing fiber. Unlike existing hybrid PCA-based DAS denoising approaches, this work explicitly investigates PCA as a standalone denoising framework, addressing the lack of systematic evaluation of its effectiveness and practical applicability. Results show that PCA improves the signal-to-noise ratio (SNR) by at least 4.7 dB across a range of strain levels. The SNR also shows improvements exceeding 5 dB for sensing fiber lengths up to 5.2 km. For 10.2 km vibration location, PCA still achieved around 2.45 dB of SNR improvement. The PCA algorithm was then compared with traditional denoising algorithms, i.e., Moving Average, Low-Pass Filtering, and Wavelet Denoising, at a fixed sensing fiber length of 3.2 km and 2 Vpp applied to the PZT. PCA outperformed these approaches in noise reduction while maintaining moderate computational cost. Overall, PCA effectively suppresses background noise while preserving the integrity of the vibration signal. These results indicate that standalone PCA is a practical denoising option for DAS applications that require improved SNR at a moderate processing cost.

Text 126981.pdf - Published Version Available under License Creative Commons Attribution. Download (1MB) Official URL or Download Paper: https://teknomekanik.ppj.unp.ac.id/index.php/tekno...

Actions (login required)

View Item