Prediction of dust concentration in open cast coal mine using artificial neural network

Abstract

Coal dust is a major pollutant in the ambient air of coal mining areas. The pollution due to open cast mining is more severe than pollution due to underground mining. Prediction of ambient concentration of pollutants should be known to implement any control techniques to reduce their concentrations. In this paper, three models were developed to predict the concentration of dust particles at various locations away from the source of pollution. These models are developed using Multilayer Perception Network and learning is done by back–propagation algorithm. The data for training and testing the network is collected from the field work done in North Karanpura Coal Mine in Jharkhand, India, which is an open cast mine. The meteorological data (wind velocity, dispersion coefficients, rain fall, cloud cover and temperature), geographical data (distance of the receptor point from the source in the direction of wind and distance of the receptor from source in the direction perpendicular to wind direction) and emission rate are used as inputs in the formation of models. The number of inputs for Model 1, Model 2, and Model 3 are six, seven, and nine, respectively. The output (dust concentration) is same for all the three models. The performance of the developed models was evaluated on the basis index of agreement and other statistical parameters i.e., the mean and the deviations of the observed and predicted concentrations, root mean square error, maximum deviation and minimum deviation, normalized mean square error, model bias and fractional bias. It was seen that the overall performance of Model 3 was better than Models 1 and 2. Artificial neural network (ANN) based dust concentration prediction model yielded a better performance than the Gaussian–Plume model.