A Physical and Numerical Based Model for Early Prediction of Landslides Using Wireless Sensor Network

Abstract

Landslides are a frequent and recurrent problem in hilly regions of India and predicting them is always a challenging task. In this paper, an attempt was made to deal with this problem using advanced physical and numerical modeling methods. Detailed understanding of the initial slope failures is very interesting, and challenging at the same time, in the design and development of wireless sensor network based on early warning of landslide monitoring. A small scale physical model was developed to assess the instability through a sensor network with variable rain fall intensity. This was achieved by increasing the simulated rain water flow intensity in different time spans (dry condition, at t=0 to t= 30 min, 0.5 mm/min at t=30 to t= 60 min, 0.75 mm/min at t=60 to t= 91 min and 1 mm/min at t=91 to t= 120 min). The water level and movement in the slope was recorded by rainfall sensor, vibration sensor, soil moisture sensor and a digital camera. The following changes were observed during the slope failure: a) movement of small particles at top of the slope; b) initial failure of medium size soil particle; c) scouring of soil mass; d) whole slope collapse. The obtained results clearly indicated the superiority and effectiveness of the proposed system in providing a factor of safety for the progressive slope.