Digitalization of asset surveillance through distributed fiber-optic sensing: Geophysics and engineering diagnostics and streaming

Abstract

Fiber-optic distributed acoustic sensing (DAS) can listen to a wide range of signals. This listening takes place at high sampling rates with fine spatial resolution, resulting in large data volumes. Data streaming solutions are available but result in large transmission and storage costs. In this paper, we describe strategies to convert large data streams from DAS interrogator units to diagnostics or processed products. Optimizing DAS systems results in higher signal-to-noise ratio for signals while extracting diagnostic features out of the noise that could be related to production or well engineering. DAS has sensitivity to diverse signals, and the first goal of edge processing is to separate them for consumption by various disciplines. Focusing the processing on specific aspects in the DAS recordings provides data products that are streamed in efficient ways. We show how DAS processing can deploy fast algorithms so that data diagnostics are sent to remote locations. This enables real-time-diagnostics and event-detection tools. By providing the bulk of computing in the field, data upload to remote servers is efficient and targeted. We show how this managed data stream enables digitalization of engineering and geoscience assets.