Prediction of seismic events in mines using neural networks

A number of fatalities and accidents in hard-rock mining could be avoided if rockbursts could be predicted accurately. Until recently, this goal was elusive because seismicity could not be measured accurately. Recent advances in monitoring technology have improved the accuracy of seismic source parameter measurements. This creates the possibility of modelling the rock dynamics, thus allowing prediction of seismic events. This paper investigates the use of neural networks for modelling seismic time series in hard-rock mines.