基于数据挖掘技术的地下工程目标毁伤效应计算方法

张磊; 吴昊; 赵强; 王幸; 任新见; 王继民; 孔德锋

doi:10.11883/bzycj-2020-0114

基于数据挖掘技术的地下工程目标毁伤效应计算方法

doi: 10.11883/bzycj-2020-0114

张磊^1,,
吴昊²,
赵强¹,
王幸¹,
任新见¹,
王继民³,
孔德锋^1, ,

1.
军事科学院国防工程研究院工程防护研究所，河南洛阳 471023
2.
军事科学院，北京 100071
3.
河海大学计算机与信息学院，江苏南京 211100

详细信息

作者简介:
张　磊（1974－　），男，博士，副研究员，ustczhanglei@163.com

通讯作者:
孔德锋（1989－　），男，硕士，工程师，kdf35@126.com

中图分类号: O381
计量
- 文章访问数: 588
- HTML全文浏览量: 206
- PDF下载量: 166
- 被引次数: 0
出版历程
- 收稿日期: 2020-04-17
- 修回日期: 2020-11-26
- 网络出版日期: 2021-03-05
- 刊出日期: 2021-03-10

Calculation method of damage effects of underground engineering objectives based on data mining technology

1.
Institute of Defense Engineering, Academy of Militory Science of PLA, Luoyang 471023, Henan, China
2.
Academy of Military Sciences, People's Liberation Army, Beijing 100071, China
3.
College of Computer and Information, Hehai University, Nanjing 211110, Jiangsu, China

摘要

摘要: 针对毁伤试验数据少、不均匀、不连续、范围窄等带来的计算精度不高的问题。研究通过数据挖掘技术进行毁伤效应计算。利用数据库管理毁伤数据，通过数据清洗技术识别并清除数据异常点，以保证数据库中数据的质量。建立了算法评价方法以选择最优经验算法。通过特征选择对高维毁伤数据进行降维，确定毁伤效应的主要控制参数进行神经网络学习和k-近邻检索。在此基础上建立基于数据融合的“三阶段”毁伤效应计算模型，可依据试验数据、经验算法和神经网络模型进行毁伤效应计算。实际应用表明，所提出的计算方法，能够满足实际应用需求。
- 数据挖掘 /
- 毁伤效应 /
- 数据质量分析 /
- 特征选择 /
- k-近邻检索 /
- 神经网络
Abstract: Aiming at low calculation accuracy of damage effect caused by less data, uneven, discontinuity and narrow distribution of damage experimental data, data mining technology is introduced to calculate damage effect. The database manages damage metadata and the data cleaning technology is used to identify and eliminate dead points’ data in order to control the data quality in database. An algorithm evaluation method is established to select the optimal empirical algorithm. The dimensionality reduction of high-dimensional damage data is achieved through feature selection and the main control parameters are chosen to train neural network model and k-nearest neighbor search. The “three-stage” damage effects calculation model based on data fusion has been established. The model can be used to calculate weapon damage effect based on experimental data, the empirical algorithm and the BP neural network model. The software has been developed to complete the damage calculation, and the results shows that the proposed method can meet the needs of practical application.
- data mining /
- damage effects /
- data quality analysis /
- feature selection /
- k-nearest neighbor search /
- neural network

HTML全文

图 1 地下工程目标毁伤评估源数据结构

Figure 1. Structure of source data for underground engineering target damage assessment

下载: 全尺寸图片幻灯片

图 2 数据样本频率分布图

Figure 2. Frequency distribution of data sample

下载: 全尺寸图片幻灯片

图 3 经验算法评价流程

Figure 3. Evaluation process of empirical algorithms

下载: 全尺寸图片幻灯片

图 4 “三阶段”法毁伤效应数据挖掘流程

Figure 4. Flow of damage effect data mining based on “three steps”

下载: 全尺寸图片幻灯片

图 5 基于KNN的相似检索流程图

Figure 5. Flow chart of similar searching based on KNN

下载: 全尺寸图片幻灯片

图 6 地下工程毁伤效应计算流程图

Figure 6. Calculation flow chart of damage effect to underground engineering

下载: 全尺寸图片幻灯片

表 1 经验算法和BP神经网络模型MAPE值

Table 1. MAPE values of empirical algorithms and BP neural network model

模型	δ_map/%	模型	δ_map/%
BRL公式^[15]	71.34	Forrestal公式^[18]	56.90
Young公式^[16]	52.76	BP预测模型	40.11
NDRC公式^[17]	65.92

下载: 导出CSV

表 2 地下工程毁伤计算所需主要控制参数

Table 2. Main control parameters required for damage calculation of underground engineering

参数名称	输入值	备注	参数名称	输入值	备注
战斗部型号	***	下拉菜单选择	主坑道轴线长	100 m
弹体质量	874 kg		坑道等效直径	5 m
弹体直径	0.37 m		防护门位置	50 m
弹头长度	1.111 m		防护门抗力	0.2 MPa
弹头形状	卵形		围岩材料种类	岩石	下拉菜单选择
装药TNT当量	242.7 kg		围岩材料波速	3000 m/s
弹体长度	2.511 m		围岩材料强度	46 MPa
弹着点坐标	（10 m, 0 m）		围岩材料密度	2000 kg/m³
着靶速度	200 m/s		衬砌材料种类	钢筋混凝土	下拉菜单选择
弹着角	5°		衬砌材料强度	60 MPa
攻角	0°		衬砌层厚度	0.5 m
坑道类型	直通式出入口	下拉菜单选择	口部防护层厚度	5 m

下载: 导出CSV

参考文献(20)

[1]	任辉启, 穆朝民, 刘瑞朝, 等. 精确制导武器侵彻效应与工程防护[M]. 北京: 科学出版社, 2016: 54−58.
[2]	KANTARDZIC M. Data mining: concepts, models, methods, and algorithms [M]. 2nd ed. Hoboken: Wiley Publishing, Inc, 2011: 135−137.
[3]	HE Z, WU Q, WEN L J, et al. A process mining approach to improve emergency rescue processes of fatal gas explosion accidents in Chinese coal mines [J]. Safety Science, 2019, 111: 154–166. DOI: 10.1016/j.ssci.2018.07.006.
[4]	RYAN S, THALER S. Artificial neural networks for characterizing Whipple shield performance [J]. Procedia Engineering, 2013, 58: 31–38. DOI: 10.1016/j.proeng.2013.05.006.
[5]	RYAN S, THALER S, KANDANAARACHCHI S. Machine learning methods for predicting the outcome of hypervelocity impact events [J]. Expert Systems with Applications, 2016, 45: 23–39. DOI: 10.1016/j.eswa.2015.09.038.
[6]	李建光, 李永池, 王玉岚. 人工神经网络在弹体侵彻混凝土深度中的应用 [J]. 中国工程科学, 2007, 9(8): 77–81. DOI: 10.3969/j.issn.1009-1742.2007.08.016. LI J G, LI Y C, WANG Y L. Penetration depth of projectiles into concrete using artificial neural network [J]. Engineering Sciences, 2007, 9(8): 77–81. DOI: 10.3969/j.issn.1009-1742.2007.08.016.
[7]	金胜兵, 刘军, 张磊, 等. 数据挖掘技术在混凝土侵彻深度分析中的应用[J]. 解放军理工大学学报(自然科学版) [2020-09-21]. http://kns.cnki.net/kcms/detail/32.1430.N.20170602.1014.002.html. DOI: 10.12018/j.issn.1009-3443.20170223003/2017.06.02. JIN S B, LIU J, ZHANG L, et al. The application of data mining technology in the analysis of the projectile penetration depth in concrete[J]. Journal of PLA University of Science and Technology (Natural Science Edition) [2020-09-21]. http://kns.cnki.net/kcms/detail/32.1430.N.20170602.1014.002.html. DOI: 10.12018/j.issn.1009-3443.20170223003/2017.06.02.
[8]	杨秀敏, 邓国强. 常规钻地武器破坏效应的研究现状和发展 [J]. 后勤工程学院学报, 2016, 32(5): 1–9. DOI: 10.3969/j.issn.1672-7843.2016.05.001. YANG X M, DENG G Q. The research status and development of damage effect of conventional earth penetration weapon [J]. Journal of Logistical Engineering University, 2016, 32(5): 1–9. DOI: 10.3969/j.issn.1672-7843.2016.05.001.
[9]	张国星, 强洪夫, 陈福振, 等. 钻地弹侵彻地下工事问题的研究与发展 [J]. 飞航导弹, 2018(6): 34–38. DOI: 10.16338/j.issn.1009-1319.20170372. ZHAGN G X, QIANG H F, CHEN F Z, et al. Research and development of the penetration of ground-penetrating projectiles into underground engineering [J]. Aerodynamic Missile Journal, 2018(6): 34–38. DOI: 10.16338/j.issn.1009-1319.20170372.
[10]	梁国栋. 钻地弹攻击地下目标的效能评估[D]. 南京: 南京理工大学, 2007: 18−24.
[11]	吴越. 钻地弹对典型地堡侵爆复合毁伤效能影响要素研究[D]. 南京: 南京理工大学, 2014: 45−67.
[12]	潘丽娜. 神经网络及其组合模型在时间序列预测中的研究与应用[D]. 兰州: 兰州大学, 2018: 32−56.
[13]	黄丽. 基于云平台的预测分析算法的研究与实现[D]. 北京: 北京邮电大学, 2016: 1−32, 24−35.
[14]	FLETCHER L, KATKOVNIK V, STEFFENS F E, et al. Optimizing the number of hidden nodes of a feedforward artificial neural network [C]//Proceedings of 1998 IEEE International Joint Conference on Neural Networks Proceedings. IEEE World Congress on Computational Intelligence. Anchorage: IEEE, 1998. DOI: 10.1109/IJCNN.1998.686018.
[15]	GWALTNEY R C. Missile generation and protection in light-water-cooled power reactor plants: ORNL NSIC-22 [R]. Oak Ridge: Oak Ridge National Laboratory, 1968: 1−23.
[16]	YOUNG C W. The development of empirical equation for predicting depth of an earth penetrating projectile: SC-DR-67-60 [R]. Albuquerque: Sandia National Laboratories, 1967: 1−11.
[17]	NDRC. Effects of impact and explosion: division 2 [R]. Washington: National Defence Research Committee, 1946: 1−7.
[18]	FORRESTAL M J, ALTMAN B S, CARGILE J D, et al. An empirical equation for penetration depth of ogive-nose projectiles into concrete targets [J]. International Journal of Impact Engineering, 1994, 15(4): 395–405. DOI: 10.1016/0734-743X(94)80024-4.
[19]	HAN J W, KAMBER M, PEI J. Data mining: concepts and techniques [M]. 3rd ed. Amsterdam: Elsevier Inc, 2012.
[20]	BECKMANN N, KRIEGEL H P, SCHNEIDER R, et al. The R*-tree: an efficient and robust access method for points and rectangles [J]. Acm Sigmod Record, 1990, 19(2): 322–331. DOI: 10.1145/93605.98741.