ABSTRACT
Production activities in underground mines generate respirable dust which impacts worker’s health and productivity. This underscores the importance of accurately predicting dust concentration towards effecting proactive and timely measures of mitigation. We develop an artificial neural network (ANN) model for an underground metal mine that predicts dust concentration using input parameters that are derived from production activities. The model provides fairly good results, with the prospect of yielding better results with improved data collection. The model produces a correlation of 0.70 between the predicted and actual dust concentration. The work in this paper constitutes the first phase of a larger framework that seeks to manage workers’ exposure to respirable dust by incorporating ventilation in short-term production scheduling. In a future work, we seek to incorporate predictions from the ANN model and the impact of conventional dust controls into short-term production schedule optimization as mathematical constraints. This will aid in identifying high dust production activities proactively, and effectively managing available ventilation and dust control measures to enhance miners’ safety.
