矿山竖井工程围岩质量分级中影响围岩质量分级的因素众多,各个因素间的非线性作用关系复杂,围岩分级过程中人为因素影响大,分级结果的准确性较差。神经网络通过合适的样本学习,能自动建立各个因素与围岩质量分级间的对应关系,能很好的解决类似矿山竖井围岩质量分级评价。从围岩介质特性、环境条件以及工程因素3个方面系统分析了影响岩体质量分级的因素指标,构建了围岩质量分级的神经网络模型,根据工程实例建立学习样本,经过对网络模型的训练与检验,证实神经网络具有较好的收敛性和稳定性,在岩体质量分级中应用具有很好的实用性。 There are many factors that influence the grading of surrounding rock quality in rock mass classification of mine shaft. The nonlinear relationship between each factor is complex. Human factors influence the classification of surrounding rock, and the accuracy of grading results is poor. Neural network through appropriate sample learning, can automatically establish the correspondence between the various factors and the quality of rock mass classification, which can well solve the similar classification of mine shaft rock mass rating. This paper systematically analyzes the factors influencing the grading of rock masses from three aspects of medium characteristics, environmental conditions and engineering factors of surrounding rock, builds a neural network model of grading of surrounding rock quality, establishes learning samples according to engineering examples, and through training of network model With the test, it is proved that the neural network has good convergence and stability. It has good practicability in rock mass grading.