Quantitative evaluation of coal-mining geological condition
With the development of modern coal industry, it is a growing attention to evaluate coal-mining geological condition in the worlds major coal-producing countries. This study proposes an Artificial Neural Networks (ANN) model that was constructed by ten significant factors using back-propagation (BP) algorithm. These seven factors include (1) fault throw, (2) fault density, (3) fault intensity, (4) fracture fractal dimension, (5) coal thickness, (6) abnormalities of coal thickness, (7) coal structure, (8) coal dip, (9) change of floor elevation, (10) combination of rock. The optimizing division method and the inserted-value method were used to establish samples for network training, and the structure of input, hidden and output layers of BP network was optimized. A total of 15 potential cases collected in Dongpo Mine were fed into the ANN model for training and testing. Achievement predicting 27 unknown units demonstrates that the presented ANN model with ten significant factors can provide a stable and reliable result for the prediction of coal-mining geological condition in hazard mitigation and guarding systems. The results show that it is effective in evaluating the unknown units for the trained network.
Coal-mining geological condition artificial neural networks back-propagation training samples quantitative evaluation
Baolong Zhu
School of Civil Engineering and Architecture,Southwest University of Science and Technology,Mianyang 621010,Sichuan Province,China
国际会议
The First International Symposium on Mine Safety Science and Engineering (首届矿山安全科学与工程学术会议)
北京
英文
650-659
2011-10-27（万方平台首次上网日期，不代表论文的发表时间）