SIMULATING CONDITIONS OF SLOPE FAILURE IN THE LAB WITH SAND AND WATER USING A MULTIMODAL MONITORING DEVICE FOR DATA COLLECTION

Loss of life and property damage can be caused by slope failure. Slope failure is becoming increasingly common along the Great Lakes due to high wave action and increased rainfall caused by climate change. Our main goal is to understand the precursors to slope failure so we can better monitor and predict them. A few benefits of our research include raising public awareness of the dangers of slope failure and protecting communities near shorelines.

We conducted our lab experiments on unsorted sand using a multimodal monitoring device consisting of a strain gauge sensor, four moisture sensors, and eight temperature sensors. The strain gauge sensor, collecting one data point per 0.03 seconds, is run on its own Raspberry Pi platform. The moisture and temperature sensors, each collecting one data point per 10 seconds, are all run by a separate Raspberry Pi platform. We run both platforms simultaneously during our experiments.

Our setup consists of a plastic container to hold sand with a 45-55-degree sloped surface. A buried sandbag (0.05kg) attached to the strain gauge communicates strain information to the Raspberry Pi. Our experiments consist of 10-second intervals of pouring 60 cc of deionized water at the top of the slope. We let the sand settle for one-minute intervals and repeat until failure occurs.

The strain gauge data has been consistent for each test. In the first half, we get negative (push) readings before rapidly rising to positive (pull) readings. The correlation between soil moisture data and strain gauge data is inconclusive, perhaps due to the difference in data collection frequency. Our future work will involve sand layers with different grain sizes to mimic real-life scenarios as much as possible and also to better control water permeability along different layers.

Our project will ultimately lead to designing a system for collecting reliable field measurements for monitoring different triggers of slope failure and remotely transmitting the data in near-real time. This will allow continuous slope monitoring to help protect the public from the aftermaths of slope failure.