基于随机森林回归方法的爆破块度预测模型研究

Study on blasting fragmentation prediction model based on random forest regression method

在筑坝材料爆破开采过程中,块度控制是确保筑坝质量最为重要的手段之一。现阶段关于爆破料块度预测的研究中,存在模型预测精度低、模型泛化能力差等问题,难于准确控制堆石料块度、符合爆破开采堆石料上坝条件。针对目前爆破预测模型存在的不足,并有效控制堆石坝料爆破块度,基于随机森林回归方法建立了爆破块度预测模型。通过交叉验证法,将随机森林模型与其他预测模型进行了对比分析,体现了该模型的优越性。在爆破块度预测系统上,结合某工程实际,验证了该模型可行性,为堆石坝爆破施工管理与控制提供了科学指导。

During blasting quarry of dam materials,fragmentation control is a key means to ensure dam construction quality.In previous studies of the prediction of blasting material fragmentation,several problems are left unsolved such as low prediction accuracy and poor generalization ability,and difficulty still remains in accurate control on the fragmentation of rockfill dam materials to meet the requirement of blasting quarry.This study develops a blasting fragmentation prediction model based on a random forest(RF)regression method to overcome the shortcomings of previous blasting prediction models and improve the control on blasting fragmentation.Cross-verification and comparison with other RF prediction models shows this RF model is superior.We verify its calculations and applicability using a blasting fragmentation prediction system of a practical project,and demonstrate its usefulness in management and control of the blasting construction of rockfill dams.