论文部分内容阅读
为克服传统方法确定地下金属矿山岩层移动参数的缺点,提出基于支持向量机(SVM)理论的地采岩层移动角选取方法。选取影响岩层移动的7个主要因素(矿体上、下盘围岩普氏系数、稳固程度,以及开采深度、开采厚度、矿体倾角)作为模型的输入,上、下盘岩层移动角为模型的输出,在收集65组金属矿山开采岩层移动参数的基础上,根据不同开采技术条件,利用SVM强有力的模式识别功能,采用RBF核函数,分别建立了崩落开采和充填回采的岩层移动参数预计模型。为提高预测模型的泛化能力和预测精度,应用遗传算法选择SVM的模型参数。应用该模型预测了三山岛金矿和狮子山铜矿开采岩层移动参数。结果表明:模型选取的因素合理,建立的遗传算法优化SVM回归模型对地采岩层移动角预测效果良好,为岩层移动角评价提供一种新思路。
In order to overcome the shortcomings of the traditional method to determine the movement parameters of underground metal mines, this paper proposes a selection method of ground movement angle based on Support Vector Machine (SVM) theory. The seven main factors influencing the movement of rock strata (the Platts coefficient, the stability degree, the mining depth, the mining thickness and the dip angle of the orebody) are selected as the inputs of the model, and the movement angles of the upper and lower plates are taken as the models Based on the collection of the movement parameters of 65 groups of metal mining rock strata, based on the different mining technology conditions, using the powerful pattern recognition function of SVM and the RBF kernel function, the rock movement parameters of caving mining and filling recovery are established respectively model. In order to improve the generalization ability and prediction accuracy of the prediction model, genetic algorithm is used to select the model parameters of SVM. The model is applied to predict the rock movement parameters of Sanshandao Gold Mine and Shizishan Copper Mine. The results show that the model selection is reasonable. The genetic algorithm optimized SVM regression model has a good effect on prediction of ground movement angle and provides a new idea for the evaluation of rock movement angle.