Who you work with#
The project would involve collaboration with underground mine operators, mining engineers, mine safety officers, geologists, geotechnical engineers, mine surveyors, electrical/IoT engineers, AI/ML developers and GIS specialists. The system would also need inputs from personnel responsible for mine planning and emergency response so that alerts generated by the system can support practical operational decisions.
Where feasible, interaction with communities living near underground mining areas would also help understand the practical impact of subsidence, the warning mechanisms currently available to them, and the type of alerts that would be most useful during a potential ground-instability event.
The situation#
Underground coal extraction can create changes in the stress and stability of the overlying rock and soil layers. In some locations, these changes can eventually result in surface deformation or subsidence, affecting nearby communities, roads, buildings, agricultural land, forests and other infrastructure.
The problem statement identifies that subsidence monitoring in India is still largely dependent on conventional field observations, periodic surveys and post-facto damage assessments, which may not provide sufficiently timely warning before a critical ground-failure event.
A safer approach requires continuous observation of small changes in the ground before they develop into a more serious event. This creates a need for a low-cost system that can collect measurements continuously from multiple locations, analyze the data automatically and communicate warnings to responsible personnel.
What is going wrong#
The main issue is that ground deformation can develop progressively, while monitoring may not always be continuous or spatially dense enough to identify every early-stage change.
Important indicators such as:
- abnormal ground tilt,
- changes in relative distance between monitoring points,
- early crack formation,
- unusual vibration patterns
may provide useful evidence of developing instability. The problem specifically identifies these parameters as signals that the proposed system should continuously monitor.
The lack of a unified real-time system can also make it difficult to combine sensor measurements, historical information, spatial information and AI-based analysis into one operational view.
Why it is not fixed yet#
Mine subsidence is a complex engineering problem because conditions vary from mine to mine depending on geology, mining methods, underground panel layout, depth, ground characteristics and environmental conditions.
A practical solution also needs to satisfy several requirements simultaneously:
low cost + continuous sensing + low power operation + reliable communication + accurate anomaly detection + spatial visualization + actionable alerts.
A system designed for underground coal mines cannot simply depend on a single sensor or a generic machine-learning model. It needs multiple sensing points and should be robust enough to operate under Indian geo-mining conditions, therefore calls for a solution that is robust, low power, scalable and suitable for Indian geo-mining conditions.