改进BP算法在开采煤层自燃危险性预测中的应用

为了正确预测开采煤层自然发火的趋势与危险性，提出了一种改进的BP网络模型：采用自适应变步长法和改进模拟退火法（SA法）等措施，以加快BP网络的收敛速度，增强了其跳出局部极小点的能力。实际应用表明，该模型收敛速度快，准确性高，是一种十分有效的开采煤层自燃危险性预测方法，对煤矿安全生产具有重要的指导意义。

The application of modified BP neural network to predict mine coal layer spontaneous combustion

For the purpose of predicting the dangerous state in mine coal layer spontaneous combustion correctly, a kind of modified BP neural network was put forward in this paper.In order to speed up the network convergence speed and avoid falling into local minimum point, a mixed method was adopted, by means of changing network iteration step length and the Simulated Annealing(SA)method.Practical application demonstrates that the prediction model has fast convergence speed and good prediction accuracy, which is also a kind of very efficient prediction method for coal layer spontaneous combustion, and has an important practical meaning for the mine production safety.