Research Projects

Landslides in Himalayan terrain pose a significant threat to human lives and infrastructure, with the Batseri Rock fall event in Himachal Pradesh last year causing death and injuries. To mitigate these risks, a robust landslide prediction model is needed. Landslide events are linked to slope geometry and geotechnical strength attributes of rock/soil mass, which decrease significantly in water. Most fatal landslides occur during monsoons, and the relationship between strength reduction and water saturation is crucial for stability. The proposed research aims to investigate the Sangla-Batseri-Chitkul road in Himachal Pradesh, combining field data, laboratory tests, and numerical simulation results to create artificial intelligence-based models for cost-effective and reliable landslide prediction. Laboratory testing and numerical simulation will consider varying hydrological conditions, while an empirical relationship will be developed to predict strength reduction based on rainfall and litho-type. Numerous numerical models will be generated for different slopes, binding geometric, geotechnical, and hydrological parameters with safety factors. Deep learning models will be generated using this data, and web scraping will be used to obtain precipitation data from credible online sources. The developed web application will be public domain, and the code can be tested by other users. This is the first attempt to use actual ground data, numerical modelling, and AI tools to reliably predict slope stability. The research will provide valuable insights into landslide failure mechanisms for large slopes, yielding advanced tools and algorithms to tackle them.