ANN-PSO-GA模型在湿喷混凝土强度预测及配合比优化中的应用  

Application of ANN-PSO-GA model in UCS prediction and mix proportion optimization of wet shotcrete

作　　者：韩斌[1] 吉坤 胡亚飞 姚松 HAN Bin;JI Kun;HU Yafei;YAO Song(School of Civil and Resource Engineering,University of Science and Technology Beijing,Beijing 100083,China;NFC Africa Mining PLC,Kitwe,Copperbelt 22592,Zambia)

机构地区：[1]北京科技大学土木与资源工程学院,北京100083 [2]中色非洲矿业有限公司,铜带省基特韦22592

出　　处：《采矿与安全工程学报》2021年第3期584-591,共8页Journal of Mining & Safety Engineering

基　　金：国家自然科学基金项目(51374034,51304011);国家重点研发计划项目(2018YFC1900603,2018YFC0604604)。

摘　　要：为了实现对湿喷混凝土强度的高精度预测及其配合比的智能精细化选择,本文建立了一种新型的ANN-PSO-GA模型。首先采用粒子群算法(PSO)对人工神经网络(ANN)进行优化,实现其对湿喷混凝土强度的高精度预测。然后将训练好的ANN-PSO模块联合给定工程条件作为目标函数,利用遗传算法(GA)进行寻优,即可得到满足该目标下的最优配合比。研究结果表明:以金川二矿区支护为例,该模型对该矿湿喷混凝土强度预测的平均相对误差MRE为2.755%,可决系数R^(2)为0.980。通过联合矿山精准支护需求,仅用时约4分11秒就寻优得到符合C15,C20,C25,C30强度标准的湿喷混凝土最佳配合比,并且经过强度实验与流动性实验检验,均符合该矿标准。本模型的成功应用不仅极大地提高了该矿湿喷混凝土强度确定的效率,还使之前依靠实验摸索的配合比寻优工作变得更加智能化、精准化。In order to achieve highly accurate prediction of uniaxial compressive strength(UCS) of wet shotcrete and intelligently precise selection of its mix proportion, a new ANN-PSO-GA model was established. Firstly, ANN was optimized through particle swarm optimization(PSO) to precisely predict strength of wet shotcrete. Subsequently, with well-trained ANN-PSO module, combined with given engineering conditions, taken as objective function, the optimal mix proportion in the target condition was obtained through genetic algorithm(GA) in optimization. The results have shown that in the practice of support work for Jinchuan mine, the mean relative error(MRE) of the model in predicting strength of wet shotcrete is 2.755%, and the coefficient of determination(R^(2)) is 0.980. Combined with accurate support requirement, it took about 4 minutes 11 seconds to obtain the optimal mix proportion of wet shotcrete in accordance with strength standards of C15, C20, C25 and C30. Moreover, it passed the strength and fluidity test and corresponded to the specification of the mining area. The successful application of this model not only improves the efficiency in selecting the strength of wet shotcrete in this mine, but also makes the optimization work of mix proportion that previously relied on experiments become more intelligent and precise.

关 键 词：湿喷混凝土 配合比 强度预测 智能优化 算法 

分 类 号：TD853[矿业工程—金属矿开采]