Rapid detection of landslide events using battery-operated long-range wireless tiltmeters

Abstract

Increasingly, infrastructure assets are affected by landslides and slope instability caused by more intense storms and extreme weather events. Future climate and land cover changes will impact landslide susceptibility, increasing risk over large areas. In recent years, long-range low-power wireless monitoring systems have been deployed for the preventive control of slope instabilities, potential landslides and earthworks that could affect infrastructures. The ability to access measurements from several parameters, at the right sampling frequency,
enables analysis of the behaviour and performance of soil, rocks and structures. However, the proper implementation of remedial actions such as evacuations, road closures or allocation of resources, requires immediate alerts. In this paper, we present a method to quickly detect a landslide event through a network of battery-operated, long-range wireless tiltmeters. To analyse data in near-real time and avoid overloading network capacity, the solution is based on edge computing. Tilt measurements collected at 3.9 Hz are compared by a wireless edge sensor with two preset threshold values that define the accepted range for any given axis. This easy-to-deploy solution immediately issues an alert when the tilt exceeds a preset threshold. In the unlikely scenario of multiple tiltmeters detecting an event exactly at the same time, the system-reported latency is under 2 seconds for 10 tiltmeters and under 5 seconds for 25 tiltmeters. The capabilities and limitations of the system are presented, including the results obtained across performance tests and battery life estimations for different scenarios. This paper also outlines an estimated cost of deployment of the wireless solution.