基于神经网络的离散元细观参数标定方法研究——以甘肃黑方台黄土滑坡为例

doi:10.13722/j.cnki.jrme.2019.0531

Abstract
Figure/Table
References (0)
Related Citation (15)

Download: PDF (3853 KB) HTML (1 KB)
Export: BibTeX | EndNote (RIS)

Abstract There are high nonlinear characteristics between the micro-mechanical parameters and the macro- mechanical parameters of the discrete element. At present，the manual calibration of the parameter calibration method has great randomness and blindness，the calibration process is inefficient，and the calibration result is reproducible. Therefore，it is urgent to find a new accurate and efficient parameter calibration method. This paper takes 7 typical loess landslides occurred in Heifangtai from 2015 to 2017 as the research object. Firstly，according to the previous mechanical test and related literature，the values of six eigen parameters are determined. And then，the range of values of the remaining six parameters is calibrated by the indoor angle of repose test and the GEMM material database. Through the remaining six parameters sensitivity analysis，the energy density of the JKR model is determined to be 100 J/m2，and the series of orthogonal tests with 6 remaining landslides as variables are designed. Obtaining the parameters of each group and their corresponding simulation forms(horizontal width of the sliding source area，longitudinal width of the sliding source area，wall height of the landslide，overall landslide drop，slip distance，lateral width of the accumulation area). Then use this as a sample to carry out neural network training，and determine the mapping relationship between the simulated shape and the parameters. The optimal simulation parameters of the six landslides is inverted and calibrated by this mapping relationship，and use it as the value range of numerical simulation parameters of loess landslide in this area. Finally，the reliability and accuracy of the parameter range are verified by the forward performance of the 7th landslide，proving this method can avoiding the complexity and randomness of artificial calibration parameters under the same geological conditions.

Key words： slope engineering discrete element mesoscopic parameters neural networks parameter inversion

	Service

	E-mail this article
	Add to my bookshelf
	Add to citation manager
	E-mail Alert
	RSS
	Articles by authors
	ZHOU Xiaopeng1
	2
	XU Qiang1
	ZHAO Kuanyao1
	CHEN Wanlin1
	JU Yuanzhen1
	ZHOU Qi1

Cite this article:

ZHOU Xiaopeng1,2,XU Qiang1, et al. Research on calibration method of discrete element mesoscopic parameters based on neural network landslide in Heifangtai，Gansu as an example[J]. , 2020, 39(S1): 2837-2847.

URL:

http://rockmech.whrsm.ac.cn/EN/10.13722/j.cnki.jrme.2019.0531 OR http://rockmech.whrsm.ac.cn/EN/Y2020/V39/IS1/2837

[13]	CUNDALL P A. A computer model for simulating progressive large scale movement in block rock system[C]// Symposium ISRM. [S. l.]：[s. n.]，1971：129–136.
[2]	YOON J. Application of experimental design and optimization to PFC model calibration in uniaxial compression simulation[J]. International Journal of Rock Mechanics and Mining Sciences，2007，44(6)：871–889.
[1]	POTYONDY D O，CUNDALL P A. A bonded-particle model for rock[J]. International Journal of Rock Mechanics and Mining Sciences，2004，41(8)：1 329–1 364.
[4]	詹永祥，姚海林，张静波. 黏性粗粒土颗粒宏细观力学参数相关性研究[J]. 铁道工程学报，2016，33(11)：12–16.(ZHAN Yongxiang，YAO Hailin，ZHANG Jingbo. Research on the relationship between mesoscopic and macroscopic mechanical parameters of cohesive coarse grained soil[J]. Journal of Railway Engineering Society，2016，33(11)：12–16.(in Chinese))
[3]	周博，汪华斌，赵文锋，等. 黏性材料细观与宏观力学参数相关性研究[J]. 岩土力学，33(10)：3 171–3 175.(ZHOU Bo，WANG Huabin，ZHAO Wenfeng，et al. Analysis of relationship between particle mesoscopic and macroscopic mechanical parameters of cohesive materials[J]. Rock and Soil Mechanics，2012，33(10)：3 171– 3 178.(in Chinese))
[5]	彭国园，周海清，周淑玲. 红粘土细观力学特性的颗粒流模拟[J]. 重庆理工大学学报：自然科学，2017，31(1)：41–45.(PENG Guoyuan，ZHOU Haiqing，ZHOU Shuling. Simulation of me-soscopic mechanic properties on red clay by using particle flow code[J]. Journal of Chongqing University of Technology：Natural Science，2017，31(1)：41–45.( in Chinese))
[7]	吴华锐. 数字边坡信息技术研究[硕士学位论文][D]. 北京：北方工业大学，2017.(WU Huarui. Digital slope information technology research[M. S. Thesis][D]. Beijing：North China University of Technology，2017.( in Chinese) )
[9]	周飞. 甘肃省黑方台黄土斜坡变形特征与滑坡机理研究[硕士学位论文][D]. 成都：成都理工大学，2015.(ZHOU Fei. Study on the deformation characteristics of loess slope and mechanism of landslide in Heifangtai，Gansu Province[M. S. Thesis][D]. Chengdu：Chengdu University of Technology，2015.( in Chinese))
[10]	赵宽耀，许强，亓星，等. 甘肃黑方台黄土滑坡类型及其发育特征研究[J]. 人民长江，2016，47(14)：46–50.(ZHAO Kuanyao，XU Qiang，QI Xing，et al. Research on types and characteristics of loess landslides in Gansu Heifangtai[J]. Yangtze River，2016，47(14)：46–50.(in Chinese))
[15]	李姝. 甘肃黑方台地下水长期作用对黄土强度影响的试验研究[硕士学位论文][D]. 成都：成都理工大学2016.(LI Shu. Experimental study of influence on strength of loess soaked in water for long–term in Heifangtai，Gansu Province[M. S. Thesis][D]. Chengdu：Chengdu University of Technology，2016.( in Chinese))
[17]	亓星. 突发型黄土滑坡监测预警研究[博士学位论文][D]. 成都：成都理工大学，2017.(QI Xing. Sudden loess landslide monitoring and early warning research—a case study of Gansu landslide in Heifangtai Loess[Ph. D. Thesis][D]. Chengdu：Chengdu Universityof Technology，2017.( in Chinese))
[6]	冯夏庭，张治强. 位移反分析的进化神经网络方法研究[J]. 岩石力学与工程学报，1999，18(5)：529–533.(FENG Xiating，ZHANG Zhiqiang. The theory of calcareous soils mechanics and its application in engineering[J]. Chinese Journal of Rock Mechanics and Engineering，1999，18(5)：529–533.(in Chinese))
[8]	许强，彭大雷，亓星，等. 2015年4.29甘肃黑方台党川2#滑坡基本特征与成因机理研究[J]. 工程地质学报，2016，4(2)：167–180.(XU Qiang，PENG Dalei，QI Xing，et al. Dangchuan 2# landslide of April 29，2015 in Heifangtai area of Gansu Province：characteristics and failure mechanism[J]. Journal of Engineering Geology，2016，24(2)：167–180.( in Chinese))
[12]	巨袁臻. 基于无人机摄影测量技术的黄土滑坡早期识别研究[硕士学位论文 [D]. 成都：成都理工大学2017.(JU Yuanzhen. Early recognition of loess landslide based on UAV photogrammetry— a case study of Heifang Terrace[M. S. Thesis][D]. Chengdu：Chengdu University of Technology，2017.( in Chinese))
[18]	彭大雷. 黄土滑坡潜在隐患早期识别研究—— 以甘肃黑方台为例[博士学位论文][D]. 成都：成都理工大学，2018.(PENG Dalei. Study on early recognition for potentially loess landslide –a case study at Heifangtai terrace，Gansu Province，China[Ph. D. Thesis][D]. Chengdu：Chengdu University of Technology，2018.( in Chinese))
[20]	张聚贤，刘伟. BP神经网络算法考虑激活函数后对强震路基塌陷变形预测研究[J]. 地震工程学报，2019，41(2)：406–411.(ZHANG Juxian，LIU Wei. Use of a BP neural network algorithm with activation function to predict the subgrade subsidence and deformation induced by strong earthquakes[J]. China Earthquake Engineering Journal，2019，41(2)：406–411.( in Chinese))
[11]	马建全. 黑方台灌区台缘黄土滑坡稳定性研究[硕士学位论文][D]. 长春：吉林大学，2012.(MA Jianquan. Stability analysis of loess landslide in loess tablelandedge of Heifangtai irrigation area[M. S. Thesis] [D]. Changchun：Jilin University，2012.(in Chinese))
[14]	王国强，郝万军，王继新. 离散单元法及其在EDEM上的实践[M]. 西安：西北工业大学出版社，2010：16–32.(WANG Guoqiang，HAO Wanjun，WANG Jixin. Discrete element method and its practice on EDEM[M]. Xian：Northwestern Polytechnical University Press，2010：16–32.( in Chinese))
[16]	张一希，许强，刘方洲，等. 不同地区饱和原状黄土静态液化特性试验研究[J] .地质科技情报，2018，37(5)：229–233.(ZHANG Yixi，XU Qiang，LIU Fangzhou，et al. Experimental investigation of static liquefation of undisturbed saturated loess in different regions[J]. Geological Science and Technology Information，2018，37(5)：229–233.( in Chinese))
[19]	李骅锦. 数据挖掘在滑坡泥石流预测评价中的应用[硕士学位论文][D]. 成都：成都理工大学，2017.(LI Huajin. The application of data mining in prediction and evaluation of landslides and debris flow[M. S. Thesis][D]. Chengdu：Chengdu University of Technology，2017.( in Chinese))