Martinsson, J, Törnman, W & Mozaffari, S 2024, 'Innovative bayesian based seismic anomaly detection', in Daniel Johansson & Håkan Schunnesson (eds), MassMin 2024: Proceedings of the International Conference & Exhibition on Mass Mining, Luleå University of Technology, Luleå, pp. 1013-1027.
(https://papers.acg.uwa.edu.au/p/2435_G-13/)
Abstract: Detecting anomalies in mining-induced seismicity and inferring probable causes is an important part of mining engineers’ work in seismically active mines. Fast and reliable detection of anomalies is crucial for both safety and productivity in dynamic mining and seismic environments.  

We present a novel anomaly detection method, using a hypothesis test based on the robust framework of Bayesian statistics, with the ability to set false alarm probabilities. By providing robust, automatic and real-time detection of statistically significant anomalies, our proposed solution takes the burden of making subjective assessments away from mining engineers, allowing them to focus on high-impact decision making and action.  

Implemented as an online web application, the solution provides actionable, real-time seismic insights accessible on multiple devices, ranging from control room displays to smartphones available in the engineers’ pocket. It includes a dynamic interface that visualizes indicators of changes in seismic activity and energy release in real-time alongside time-series of current and historical data for informed decision-making.