基于GEP算法的膏体充填体强度预测  

Strength Prediction of Paste Filling Body by GEP

作　　者：潘虎[1] 丁函[1] PAN Hu DING Han(Hubei Liberal Arts College, Xiangyang, Hubei 441053, China)

机构地区：[1]湖北文理学院,湖北襄阳441053

出　　处：《矿业研究与开发》2017年第3期70-73,共4页Mining Research and Development

基　　金：湖北省教育厅科学技术研究项目(Q20142607)

摘　　要：国内某锡矿采用膏体充填,针对其充填体强度预测问题,构建了充填体强度的GEP算法(基因表达式编程)预测模型。以灰砂比、充填体质量浓度、密度和养护龄期作为输入因子,充填体强度为输出因子,采用输出预测公式对膏体充填体强度进行预测,得到GEP模型预测结果平均相对误差为7.74%,对比BP神经网络预测模型(平均误差21.01%),GEP模型对充填体强度预测的精度更高,可以作为充填体强度预测的新方法。A tin mine in China uses paste filling. According to the strength prediction problem of paste filling, the strength prediction model of filling body was built by GEP. The cement-sand ratio, the curing period, the mass concentration and the density of filling body were selected as the input factors, and the strength of filling body was used as the output factor. Then, the strength of paste filling body was predicted by output prediction formula. It obtained that the average relative error of prediction result of GEP model was only 7.74%. Compared to the prediction model of BP neural network with the average error of 21.01%, the GEP model had a better accuracy in the strength prediction of filling body, which could be used as a new method for the strength prediction of filling body.

关 键 词：充填体强度 膏体充填 GEP BP神经网络 预测 

分 类 号：TD853.34[矿业工程—金属矿开采]