运用多层前向人工神经网络-BP神经网络模型对不同支护类型的矿井巷道摩擦阻力系数进行模式识别,并运用改进的贝叶斯正则化方法对网络进行训练,以锚喷支护巷道为例对其摩擦阻力系数进行预测。采用BP神经网络将网络模拟值与实测值进行比较,误差在5%左右,同理对其他支护类型进行模拟预测,误差均在5%以内,预测精度较高。 The multi-layer forward artificial neural network-BP neural network model is used to pattern recognition of the frictional resistance coefficient of mine roadway with different support types. The improved Bayesian regularization method is used to train the network. Example of the friction coefficient of friction prediction. BP neural network is used to compare the network simulation value with the measured value, the error is about 5%. Similarly, the prediction of other types of support is within 5% and the prediction accuracy is high.