← Back to Sessions

Strategies for Passive Das Data Analysis at the Edge

Description: 

A significant advantage of distributed acoustic sensing (DAS) is that only a single power source needs to be maintained, and an interrogator can be connected to a fiber-optic cable to continuously collect data for months or years with little human intervention. This makes DAS an appealing tool for seismic data acquisition in remote areas, particularly for studying fine-scale processes across long distances. However, remote areas often have limited data transfer capabilities, sometimes orders of magnitude smaller than DAS data collection rates. These data transfer limits have pushed us to increase the amount of automated analysis and quality control carried out by limited computing resources in the field, particularly for long-duration experiments in remote areas. To achieve this goal, we are investigating the use of lossy compression of passive DAS data, increased use of data products for initial analysis of passive seismic data, and algorithms that directly rely on these compressed data and data products. For multiple types and levels of compression, we have quantified the lossy compression errors through multiple metrics. We also quantify errors as they propagate through common workflows for small seismic event detection and cross-correlation. Additionally, we have released open-source software for real-time data product computation that can be carried out continuously on the computer of a DAS interrogator. The methodology developed in this research is intended to impact cost-effective DAS monitoring in a wide range of applications.

Session: Understanding Earth Systems with Fiber-optic Cables

Type: Oral

Date: 4/19/2023

Presentation Time: 10:30 AM (local time)

Presenting Author: Eileen R. Martin

Student Presenter: No

Invited Presentation: Yes

Authors