DEVELOPING A PROTOTYPE MULTIMODAL DEVICE FOR MONITORING SLOPE INSTABILITY USING RASPBERRY PI

Water movement from saturated to unsaturated zones in response to a temperature gradient can make slopes unstable and lead to slope failures in regions where freeze-thaw cycle is common. Understanding the rate of such water migration under different conditions, such as temperature difference, rate of cooling or thawing, initial moisture content, porosity and permeability of rock layers, etc. are therefore important for addressing slope failure hazards in those regions. In order to do this, we have designed a prototype multimodal monitoring device with temperature and moisture sensors using Raspberry Pi platform. Currently we are testing the device under laboratory conditions to monitor water movement through sand in response to a temperature gradient. We are using capacitive moisture sensors and digital temperature sensors in our “phase zero” prototype. Capacitive moisture sensors measure dissolved ions in water instead of directly measuring water content. Therefore, it is necessary to calibrate the sensors to the experimental conditions before we can collect reliable water movement data. We are using tap water and sand grains ranging in size from >2 mm to <0.25 mm for our experiments and sensor calibration. We are using liquid nitrogen to mimic freezing conditions and for creating a temperature gradient, and layers of different sized sand grains to model sediment layers of different porosity and permeability in our sandbox model. We are using strain gauge load cell sensors for vector force data layout to identify any resulting soil creep and/or slope failure from weakening due to water movement.

We plan to expand our system with stackable pi plates and eventually incorporate real-time clock or GPS sensors for accurate date-time information, and ultrasonic sensors to measure displacement in close proximity for monitoring various landslide triggers. Our goal is to find the sensor combinations most suitable for specific locations for collecting reliable landslide monitoring data. Ultimately we aim to deploy our system under field conditions and be able to predict potential slope failures in a timely manner to prevent loss of life. Here we present our experimental setup, calibration methods, and preliminary results from this project.