img

QQ群聊

img

官方微信

高级检索

黄金科学技术 ›› 2023, Vol. 31 ›› Issue (4): 624-634.doi: 10.11872/j.issn.1005-2518.2023.04.026

• 采选技术与矿山管理 • 上一篇    下一篇

基于SMA算法优化随机森林的PPV预测模型

邓红卫(),罗亮   

  1. 中南大学资源与安全工程学院,湖南 长沙 410083
  • 收稿日期:2023-02-20 修回日期:2023-04-19 出版日期:2023-08-30 发布日期:2023-09-20
  • 作者简介:邓红卫(1969-),男,湖南岳阳人,博士,教授,从事矿山水害防治与水资源利用、地下水环境评价与水污染修复、金属矿山安全高效开采等研究工作。denghw208@126.com
  • 基金资助:
    校企联合创新项目“浅埋地下超大型石窟岩体工程测试及稳定性检测技术研究”(1053320200341)

PPV Prediction Model Based on Random Forest Optimized by SMA Algorithm

Hongwei DENG(),Liang LUO   

  1. School of Resources and Safety Engineering,Central South University,Changsha 410083,Hunan,China
  • Received:2023-02-20 Revised:2023-04-19 Online:2023-08-30 Published:2023-09-20

摘要:

爆破振动速度峰值(Peak Particle Velocity,PPV)的准确预测是有效控制爆破工程振动危害的前提。为了提高爆破振动速度峰值的预测精度,提出将黏菌算法(Slime Mould Algorithm,SMA)对随机森林(Random Forest,RF)中的树的个数和最小叶子点数2个超参数进行优化。以某露天爆破工程实例中收集的具有4个输入参数(最小抵抗线r、高差H、最大段药量Qmax、水平距离W)和1个输出参数(PPV)的23个样本的数据集为依据,将4种参数组合(Qmax-H-W-r、Qmax-H-r、Qmax-W-r、Qmax-r)作为随机森林算法中的输入参数,确定最优的参数组合。随后对SMA-RF模型、未优化RF模型和国内外常用的6组经验公式的预测结果进行比较,结果表明SMA-RF模型取得了最优的预测效果,因此在工程实践中推荐使用SMA-RF模型预测爆破振动速度峰值。

关键词: 露天爆破, 爆破振动速度峰值, 随机森林算法, 黏菌算法, 预测精度

Abstract:

The vibration caused by blasting is likely to cause instability and failure of facilities such as underground roadways,high and steep slopes in mining areas or ground buildings under dynamic action.Therefore,it is particularly important to predict the intensity of blasting vibration.The accurate prediction of peak particle velocity(PPV) is the premise of effectively controlling the vibration hazard of blasting engineering,but the current empirical formula for predicting the peak particle velocity is not accurate enough.Machine learning has obvious advantages in solving the problem of nonlinear relationship.In order to improve the prediction accuracy of the PPV prediction model,this study proposes to optimize the number of trees and the minimum number of leaf points in the random forest (RF)by slime mould algorithm (SMA) ,which overcomes the inability to obtain the optimal hyperparameters by using a single RF algorithm.Based on a dataset of 23 samples with four input parameters (minimum resistance line-r,height difference-H,maximum segment dose-Qmax,horizontal distance-W) and one output parameter(PPV) collected in an open-pit blasting engineering example,the combination of four parameters of these four parameters (Qmax-H-W-r、Qmax-H-r、Qmax-W-r、Qmax-r) was used as the input parameters in the RF algorithm,and then MAERMSEMEDEA and R2 evaluate the prediction effect of the SMA-RF model for four different input parameters to determine the optimal combination of parameters.In this model,the fitness function in SMA is defined as the root mean square error of the predicted value to enhance the robustness of the RF model.Then,the performance of SMA-RF model and unoptimized RF model and six empirical formulas commonly used in China and abroad were compared.The results show that the SMA-RF model has better prediction accuracy than the RF model,and the SMA-RF model has significantly better prediction effect than the six empirical formulas.In addition,Qmax-H-W-r can train the optimal SMA-RF model in the combination of four parameters,so it is recommended to be used to predict PPV in engineering practice.

Key words: open blasting, blasting vibration velocity peak, random forest algorithm, slime mould algorithm, prediction accuracy

中图分类号: 

  • TD253

图1

随机森林原理图"

表1

随机森林算法中被优化的超参数"

超参数含义范围
ntree树的个数0~300
mtry最小叶子点数0~20

图2

SMA-RF模型流程图"

表2

爆破振动实测数据"

序号Qmax/kgr/mH/mW/m实测值/(cm·s-1
1580.820005.00.763581
2600.8310756.00.489885
3300.82821505.00.007130
4260.860155.01.893658
5280.84682394.50.692248
??????
23260.8110155.01.191707

图3

SMA-RF模型的训练集和测试集分布"

图4

SMA-RF预测模型中适应度值的迭代情况"

图5

SMA-RF预测模型对训练集和测试集的PPV预测"

表3

不同输入参数的SMA-RF预测模型结果评估"

参数组合训练集总分
MAE得分R2得分RMSE得分MEDEA得分
Q-r1.389310.929011.929910.947214
Q-H-r0.876740.969441.250340.5444416
Q-W-r1.246420.956421.497021.105528
Q-W-H-r1.045630.963731.358230.7775312
参数组合测试集总分
MAE得分R2得分RMSE得分MEDEA得分
Q-r1.618330.996242.023041.2771213
Q-H-r1.694210.917612.155221.732415
Q-W-r1.650620.980132.221011.124539
Q-W-H-r1.395840.963622.070230.7131413

表4

不同输入参数的SMA-RF预测模型得分比较"

参数组合训练集得分测试集得分总分排名
Q-r413173
Q-H-r165212
Q-W-r89173
Q-W-H-r1213251

表5

RF预测模型结果评估"

评价指标训练集测试集
MAE1.46051.6951
R20.96750.9045
RMSE2.02012.1923
MEDEA1.06561.6955

表6

6组国内外常用经验公式"

经验方法公式
Amb-HendPPV=Kr/Qmax1/3-B
CMRIPPV=n+Kr/Qmax1/2-1
GeneralPPV=KQmaxAr-B
Indian StandardPPV=Kr/Qmax2/3B
Lang-KihlPPV=K(Qmax/r3/4)B
USBMPPV=Kr/Qmax1/2-B

表7

6组经验公式的常量项及拟合评价指标"

经验公式参数评价指标
KBAnMAER2RMSEMEDEA
USBM13.10270.7479--3.09440.01955.68261.6959
Lang-Kihl0.28671.9959--2.80630.10314.33411.4054
General9.3569e-50.53922.1427-2.29200.26493.78870.1357
Amb-Hend24.93670.6897--3.17930.00794.56761.8497
Indian Standard6.10200.7806--3.02190.03414.55701.7112
CMRIP12.9205--1.13303.15710.00944.58261.6516

图6

6组经验公式的PPV预测"

图7

4个输入参数的MSE平均下降值"

Breiman L,2001.Random forests[J].Machine Learning,45(1):5-32.
Chen Yibing, Li Tianyi, Li Xinyan,et al,2022.Research on the relationship between typhoon precipitation cloud spectrum and precipitation based on random forest and remote sensing[J].Remote Sensing Technology and Application,37(5):1277-1288.
Davies B, Farmer I W, Attewell P B,1964.Ground vibration from shallow sub-surface blasts[J].Engineer,217:553-559.
Fan Yong, Pei Yong, Yang Guangdong,et al,2022.Prediction of blasting vibration velocity peak based on an improved PSO-BP neural network[J].Journal of Vibration and Shock,41(16):194-203,302.
Guo Qinpeng, Yang Shijiao, Zhu Zhonghua,et al,2020.Predition of blasting vibration velocity using GA-BP neural network[J].Blasting,37(3):148-152.
Guo H, Zhou J, Koopialipoor M,et al,2021.Deep neural network and whale optimization algorithm to assess flyrock induced by blasting[J].Engineering with Computers,37:173-186 .
Hu X, Qu S,2018.A new approach for predicting bench blasting-induced ground vibrations:A case study[J].Journal of the Southern African Institute of Mining and Metallurgy,118(5):531-538.
Jiang Nan, Zhou Chuanbo, Ping Wen,et al,2014.Altitude effect of blasting vibration velocity in rock slopes[J].Journal of Central South University(Science and Technology),45(1):237-243.
Lee S L A, Kouzani A Z, Hu E J,2010.Random forest based lung nodule classification aided by clustering[J].Computerized Medical Imaging and Graphics,34(7):535-542.
Li S, Chen H, Wang M,et al,2020.Slime mould algorithm:A new method for stochastic optimization[J].Future Generation Computer Systems,111:300-323.
Li Xiaohan, Liu Kewei, Yang Jiacai,et al,2019.Analysis of blasting vibration effects under different ground stress[J].Gold Science and Technology,27(2):241-248.
Lin H P, Ahmadianfar I, Amiri Golilarz N,et al,2022.Adaptive slime mould algorithm for optimal design of photovoltaic models[J].Energy Science and Engineering,10(7):2035-2064.
Liu Qiang, Li Xibing, Liang Weizhang,2018.PCA-RF model for the classification of rock mass quality and its application[J].Gold Science and Technology,26(1):49-55.
Luo Xiaofeng, Qiu Wei, Huang Wenlong,et al,2020. Correction of blasting vibration propagation attenuation formula under complex terrain based on dimensional theory[C]// Engineering Construction Collection of Pumped Storage Power Station.Beijing:China Water Resources and Hydropower Publishing House.
Roy P P,1993.Putting ground vibration predictions into practice[J].Colliery Guardian,241(2):63-67.
Siskind D E, Stagg M S, Kopp J W,et al,1980.Structure response and damage produced by ground vibration from surface mine blasting[R]. Washington:United States,Bureau of Mines.
Tan Wenhui, Qu Shijie, Mao Shilong,et al,2010.Altitude effect of blasting vibration in slopes[J]. Chinese Journal of Geotechnical Engineering,32(4):619-623.
Yang Lianbing, Chen Chunbo, Zheng Hongwei,et al,2021.Retrieval of soil salinity content based on random forests regression optimized by Bayesian optimization algorithm and gentic algorithm[J].Journal of Geo-information Science,23(9):1662-1674.
Yang Youfa, Cui Bo,2009.Prediction of peak blasting velocity[J].Journal of Vibration and Shock,28(10):195-198,234-235.
Zhang P, Yin Z Y, Jin Y F,et al,2020.A novel hybrid surrogate intelligent model for creep index prediction based on particle swarm optimization and random forest[J].Engineering Geology,265:105328.
Zhang Yan, Wang Pengpeng,2022.Blasting vibration velocity prediction model based on RVM[J].Blasting,39(1):168-174.
Zhao Huabing, Long Yuan, Song Kejian,et al,2012.Predictive methods and influence factors of blasting vibration velocity[J].Engineering Blasting,18(1):24-27.
Zhou You, Chen Zuobin, Wang Jing,et al,2016.Effects of minimum burden on deep-hole rock blasting block size[J]. Engineering Blasting,22(6):70-74.
陈绎冰,李天依,李欣艳,等,2022.基于随机森林和遥感的台风降水云光谱与降水关系研究[J].遥感技术与应用,37(5):1277-1288.
范勇,裴勇,杨广栋,等,2022.基于改进PSO-BP神经网络的爆破振动速度峰值预测[J].振动与冲击,41(16):194-203,302.
郭钦鹏,杨仕教,朱忠华,等,2020.运用GA-BP神经网络对爆破振动速度预测[J].爆破,37(3):148-152.
蒋楠,周传波,平雯,等,2014.岩质边坡爆破振动速度高程效应[J].中南大学学报(自然科学版),45(1):237-243.
李萧翰,刘科伟,杨家彩,等,2019.不同地应力下爆破振动效应分析[J].黄金科学技术,27(2):241-248.
刘强,李夕兵,梁伟章,2018.岩体质量分类的PCA-RF模型及应用[J].黄金科学技术,26(1):49-55.
骆晓锋,邱伟,黄文龙,等,2020.基于量纲理论的复杂地形下爆破振动传播衰减公式修正[C]//抽水蓄能电站工程建设文集.北京:中国水利水电出版社.
谭文辉,璩世杰,毛市龙,等,2010.边坡爆破振动高程效应分析[J].岩土工程学报,32(4):619-623.
杨练兵,陈春波,郑宏伟,等,2021.基于优化随机森林回归模型的土壤盐渍化反演[J].地球信息科学学报,23(9):1662-1674.
杨佑发,崔波,2009.爆破振动速度峰值的预测[J].振动与冲击,28(10):195-198,234-235.
张研,王鹏鹏,2022.基于RVM的爆破振动速度预测模型[J].爆破,39(1):168-174.
赵华兵,龙源,宋克健,等,2012.爆破振动速度预测方法及其影响因素[J].工程爆破,18(1):24-27.
周游,陈作彬,王静,等,2016.最小抵抗线对深孔岩石爆破块度的影响[J].工程爆破,22(6):70-74.
[1] 谢饶青, 陈建宏, 肖文丰. 基于NPCA-GA-BP神经网络的采场稳定性预测方法[J]. 黄金科学技术, 2022, 30(2): 272-281.
[2] 王梅,陈建宏,杨珊. 基于等维动态马尔科夫SCGM(1,1)C模型的黄金价格预测[J]. 黄金科学技术, 2020, 28(1): 158-166.
[3] 李夕兵,朱玮,刘伟军,张德明. 基于主成分分析法与RBF神经网络的岩体可爆性研究[J]. 黄金科学技术, 2015, 23(6): 58-63.
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!