Prediction of Water Consumption Based on PSO-BP Model in Mining Face

Abstract

The prediction of water consumption is not only the important basis for development plan, the planning of water, dust pipe network layout and choice of water-saving measures of mine, but also the foundation and prerequisite for optimization design and reliability evaluation of water supply pipe network. In this paper, the water consumption of mining face is as the forecast object. Particle swarm optimization (PSO) is used for optimizing the weights of BP neural network, and then the optimal BP neural network is built for predicting the water consumption of mining face in the future for a period of time. In this paper, lin-nan mine is object of study and the water consumption of the week ahead is forecasted. The result shows that the error is within 10% of the predicted values and the real values, which indicating that the PSO-BP method used to predict the dust water supply network water is reliable.