Random forest model for identification of residential structure damage induced by blast vibration
-
摘要: 为快速、准确地评价爆破振动诱发民房结构损伤效应，借鉴随机森林理论并结合工程实际，建立露采爆破振动诱发民房结构损伤识别的随机森林模型；综合考虑爆破参数、爆破振动特征参量及房屋结构动力特性等因素，选取质点峰值振动速度、主频率、主频率持续时间、段药量、爆心距、施工质量参数、场地条件参数、屋盖形式参数、砖墙面积率、民房高度、灰缝强度和圈梁构造柱参数等12个影响因素作为模型输入，将砖混结构建筑物的损害等级作为模型输出；基于多分类器集成的思想，以108组爆破振动实测数据作为学习样本进行训练，建模过程中由多个决策树集成随机森林、用投票的方式实现对民房结构损伤有效识别；用12组现场数据验证模型的有效性；在对样本分类的同时，计算预测变量的重要性值，发现质点峰值振动速度为最重要的评价指标，其后依次为爆心距，主频率持续时间，主频率，圈梁构造柱参数，灰缝强度，屋盖形式参数，民房高度，段药量，施工质量参数，砖墙面积率和场地条件参数。研究结果表明：随机森林模型预测结果学习样本准确度是87.97%，而测试集准确度是91.67%，与实际情况吻合较好，预测精度较高。Abstract: In this work, aiming to the prediction speed and accuracy, we established a random forest model for residential structure damage induced by blast vibration identification on the basis of the random forest (RF) theory. Twelve indexes, i.e. peak particle velocity, dominant frequency, dominant frequency duration, maximum charge per delay, distance, gray joints intensity, rate of brick walls, height of housing, roof structures parameter, beam-column frames parameter, quality parameter of construction and site conditions parameters, were considered as the criterion indices for this kind of damage in the proposed model based on the of analysis of the characteristic parameters of blasting vibration and dynamic characteristics of the housing structure. 108 sets of vibration measured data were investigated to create an RF classifier. RF was a combination of tree predictors, and variable importance was measured by gini importance parameter when the forest grows. A random tree was a combination of decision trees, and each tree is generated depending on the values of random vectors sampled independently, with the same distribution for all trees in the forest. The Gini importance value shows that the peak particle velocity is the most important discrimination indicator, followed by the distance, the dominant frequency duration, the dominant frequency, the beam-column frames parameter, the gray joints intensity, the roof structures parameter, the height of housing, the maximum charge per delay, the quality parameter of construction, the rate of brick walls and the site conditions parameters. Another twelve groups of residential structure damage instances were tested as forecast samples, and the predicted results were identical with the actual situation. Engineering practices indicate that the accuracy of the RF method of learning samples is 87.97%, and the accuracy of the test samples is 91.7%, effectively verifying and supplementing the existing methods for evaluating residential structure damage induced by blast vibration.
-
Key words:
- blasting vibration /
- residential structure damage /
- randomforest /
- peak particle velocity /
- prediction
-
图 2 建立民房结构损伤预测的RF模型
Figure 2. Establishing the RF model of residential structure damage prediction
表 1 M元分类问题混淆矩阵
Table 1. M-ary classification confusion matrix
真实类别 类别预测个数 类别1 类别2 … 类别3 类别1 N11 N12 … N1M 类别2 N21 N22 … N2M … … … … … 类别M NM1 NM2 … NMM 
下载: 导出CSV
表 2 状态参量数据量化建议值
Table 2. Recommended value for quantified input and output parameters
判别因子 取值及其含义 Qc 一般取8，差取6，好取10 Sc 一般取8，差取6，好取10 Rs 木制取3，预制板取4，现浇砼取5 Bcf 无圈无柱取3，有圈无柱取4，有圈有柱取5
下载: 导出CSV
表 3 RF模型学习样本及识别结果
Table 3. Training samples and identification results of RF model
序号 Qmax/kg R/m ν/Hz vppv/(cm·s-1) Δt/ms H/m K/% S/MPa Qc Sc Rs Bcf 损伤类别 实测 RF X1 650 78.69 31.2 1.753 870 2.8 3.28 15 8 8 3 3 V2 V2 X2 650 82.57 18.6 2.714 1 090 2.8 3.28 15 8 8 3 3 V3 V3 X3 780 37.92 38.3 0.457 765 2.8 3.28 15 8 8 3 3 V1 V1 ⋮ ⋮ ⋮ ⋮ ⋮ ⋮ ⋮ ⋮ ⋮ ⋮ ⋮ ⋮ ⋮ ⋮ ⋮ X8 780 89.65 25.3 3.896 820 3.5 2.16 10 6 8 3 3 V3 V3 X9 520 84.39 37.5 1.888 1 215 6.5 2.87 25 8 10 4 4 V2 V2 X10 520 84.39 37.5 0.865 310 6.5 2.87 25 8 10 4 4 V1 V1 X11 780 86.61 39.5 3.215 1 150 6.5 2.87 25 8 10 4 4 V3 V2 X12 650 36.57 38.7 2.799 755 6.5 2.87 25 8 10 4 4 V2 V2 ⋮ ⋮ ⋮ ⋮ ⋮ ⋮ ⋮ ⋮ ⋮ ⋮ ⋮ ⋮ ⋮ ⋮ ⋮ X108 400 81.56 27.5 1.222 255 10.5 2.58 50 10 10 5 5 V1 V1
下载: 导出CSV
表 4 RF模型测试样本及识别结果对比
Table 4. RF model test samples and recognition results
序号 Qmax/kg R/m ν/Hz vppv/(cm·s-1) Δt/ms H/m K/% S/MPa Qc Sc Rs Bcf 损伤类别 实测 神经网络 RF C1 780 121.46 26.6 0.604 215 2.8 3.28 15 8 8 3 3 V1 V1 V1 C2 900 78.25 39.7 2.979 785 2.8 3.28 15 8 8 3 3 V3 V3 V3 C3 780 90.52 17.1 1.497 655 3.5 2.16 10 6 8 3 3 V2 V2 V2 C4 900 51.25 16.3 4.923 385 3.5 2.16 10 6 8 3 3 V3 V3 V3 C5 650 78.64 17.3 1.543 780 6.5 2.87 25 8 10 4 4 V1 V1 V1 C6 900 70.36 29.3 4.193 825 6.5 2.87 25 8 10 4 4 V3 V3 V2 C7 650 33.25 39.6 3.536 1 100 6.5 3.11 50 10 6 5 5 V2 V2 V2 C8 650 70.38 30.9 1.697 880 6.5 3.11 50 10 6 5 5 V1 V1 V1 C9 900 61.43 24.3 3.608 805 6.5 3.53 25 6 8 4 3 V3 V3 V3 C10 650 72.47 23.3 1.589 850 6.5 3.53 25 6 8 4 3 V2 V2 V2 C11 900 48.37 26.3 4.106 865 10.5 2.58 50 10 10 5 5 V2 V2 V2 C12 650 114.81 24.6 0.783 310 10.5 2.58 50 10 10 5 5 V1 V1 V1
下载: 导出CSV
表 5 随机森林的混淆矩阵显示训练集分类误差
Table 5. Confusion matrix drawn from Random forest showing the classification error of training set
真实类别 类别预测个数 分类误差/% V1 V2 V3 V1 56 5 0 0.082 0 V2 7 18 3 0.357 1 V3 0 2 17 0.105 3
下载: 导出CSV
表 6 随机森林的混淆矩阵显示测试集分类误差
Table 6. Confusion matrix drawn from Random forest showing the classification error of test set
真实类别 类别预测个数 V1 V2 V3 V1 4 0 0 V2 0 4 1 V3 0 0 3
下载: 导出CSV
-
[1] Khandelwal M, Singh T N.Evaluation of blast-induced ground vibration predictors[J].Soil Dynamics and Earthquake Engineering, 2007, 27(2):116-125. doi: 10.1016/j.soildyn.2006.06.004 [2] Siskind D E, Stagg M S, Kopp J W, et al. Structure response and damage produced by ground vibration from surface mine blasting[R]. USBM Report of Investigation 8507, 1980: 73-74. i [3] 中华人民共和国国家标准编写组. 爆破安全规程: GB6722-2014[S]. 北京: 中国标准出版社, 2015. [4] 史秀志, 周健, 崔松, 等.露天采矿爆破振动对民房危害预测的DDA模型及应用[J].中南大学学报(自然科学版), 2011, 42(2):441-448. http://d.old.wanfangdata.com.cn/Periodical/zngydxxb201102027Shi Xiuzhi, Zhou Jian, Cui Song, et al.Distance discriminant analysis model and its application for prediction residential house's damage against blasting vibration of open pit mining[J].Journal of Central South University (Science and Technology), 2011, 42(2):441-448. http://d.old.wanfangdata.com.cn/Periodical/zngydxxb201102027 [5] 史秀志, 周健, 杜坤, 等.爆破振动对民房破坏效应预测的BDA模型及应用[J].振动与冲击, 2010, 29(7):60-65. doi: 10.3969/j.issn.1000-3835.2010.07.013Shi Xiuzhi, Zhou Jian, Du Kun, et al.BDA model for predicting destructive effect of blast vibration on housing[J].Journal of Vibration and Shock, 2010, 29(7):60-65. doi: 10.3969/j.issn.1000-3835.2010.07.013 [6] 中国生, 徐国元, 熊正明.基于小波变换的爆破地震信号能量分析法的应用研究[J].爆炸与冲击, 2006, 26(3):222-227. doi: 10.3321/j.issn:1001-1455.2006.03.005Zhong Guosheng, Xu Guoyuan, Xiong Zhengming.Application research of the energy analysis method for blasting seismic signals based on wavelet transform[J].Explosion and Shock Waves, 2006, 26(3):222-227. doi: 10.3321/j.issn:1001-1455.2006.03.005 [7] 史秀志, 周健, 崔松, 等.露天采矿爆破振动对民房危害预测的DDA模型及应用[J].中南大学学报(自然科学版), 2011, 42(2):441-448. http://d.old.wanfangdata.com.cn/Periodical/zngydxxb201102027Shi Xiuzhi, Zhou Jian, Cui Song, et al.Distance discriminant analysis model and its application for prediction residential house's damage against blasting vibration of open pit mining[J].Journal of Central South University (Science and Technology), 2011, 42(2):441-448. http://d.old.wanfangdata.com.cn/Periodical/zngydxxb201102027 [8] 史秀志, 周健, 杜坤, 等.爆破振动对民房破坏效应预测的BDA模型及应用[J].振动与冲击, 2010, 29(7):60-65. doi: 10.3969/j.issn.1000-3835.2010.07.013Shi Xiuzhi, Zhou Jian, Du Kun, et al.BDA model for predicting destructive effect of blast vibration on housing[J].Journal of Vibration and Shock, 2010, 29(7):60-65. doi: 10.3969/j.issn.1000-3835.2010.07.013 [9] 董陇军, 李夕兵, 赵国彦, 等.露天采矿爆破振动对砌体结构破坏效应预测的Fisher判别模型及应用[J].岩石力学与工程学报, 2009, 28(4):750-756. doi: 10.3321/j.issn:1000-6915.2009.04.013Dong Longjun, Li Xibing, Zhao Guoyan, et al.Fisher discriminant analysis model and its application to predicting destructive effect of masonry structure under blasting vibration of open-pit mine[J].Chinese Journal of Rock Mechanics and Engineering, 2009, 28(4):750-756. doi: 10.3321/j.issn:1000-6915.2009.04.013 [10] Zhou J, Shi X, Li X.Utilizing gradient boosted machine for the prediction of damage to residential structures owing to blasting vibrations of open pit mining[J].Journal of Vibration and Control, 2015, 22(19).DOI: 10.1177/1077546314568172. [11] 史秀志. 爆破振动信号时频分析与爆破振动预测研究[D]. 长沙: 中南大学, 2007. [12] Melih I, Mahmut Y, Hakan A.Prediction of ground vibrations resulting from the blasting operations in an open-pit mine by adaptive neuro-fuzzy inference system[J].Environmental Geology, 2008, 56(11):97-107. doi: 10.1007/s00254-007-1143-6 [13] Shi X Z, Zhou J.Prediction residential house's damage effect near openpit against blasting vibration based on svm with grid searching method/genetic algorithm[J].Advanced Science Letters, 2012, 11(1):238-243. http://www.wanfangdata.com.cn/details/detail.do?_type=perio&id=66b331ff4b09ecb8641a35da2efeceeb [14] Breiman L.Random forests[J].Machine Learning, 2001, 45:5-32. doi: 10.1023/A:1010933404324 [15] 方匡南, 吴见彬, 朱建平等.随机森林方法研究综述[J].统计与信息论坛, 2011, 26(3):32-38. doi: 10.3969/j.issn.1007-3116.2011.03.006Fang Kuangnan, Wu Jianbin, Zhu Jianping, et al.A review of technologies on random forests[J].Statistics & Information Forum, 2011, 26(3):32-38. doi: 10.3969/j.issn.1007-3116.2011.03.006 [16] 秦锋, 杨波, 程泽凯.分类器性能评价标准研究[J].计算机技术与发展, 2006, 16(10):85-88. doi: 10.3969/j.issn.1673-629X.2006.10.029Qin Feng, Yang Bo, Chen Zekai.Research on measure criteria in evaluating classification performance[J].Computer Technology and Development, 2006, 16(10):85-88. doi: 10.3969/j.issn.1673-629X.2006.10.029 [17] Congalton R G.A review of assessing the accuracy of classifications of remotely sensed data[J].Working Pages, 1991, 119(2):270-279. http://www.sciencedirect.com/science/article/pii/003442579190048B [18] Landis J, Koch G.The measurement of observer agreement for categorical data[J].Biometrics, 1977, 33(1):159-174. doi: 10.2307/2529310 -
图(4) / 表(6)
计量
- 文章访问数: 4270
- HTML全文浏览量: 1486
- PDF下载量: 313
- 被引次数: 0