露天矿排土场滑坡预警方法的研究及应用

发布时间：2018-03-11 19:02

本文选题：排土场滑坡　切入点：稳定性分析　出处：《北京科技大学》2015年博士论文　论文类型：学位论文

【摘要】：随着我国采矿业的迅速发展,排土场在数量上还是规模上都逐渐增加。近年来,排土场滑坡事故呈多发态势,造成了严重的生命财产损失。排土场滑坡预警方法的研究对实现企业的可持续发展、社会的和谐稳定等方面具有重要的理论意义和现实的深远影响。本文通过对近40年来国内外排土场的50个典型滑坡事故案例分析,得出排土场滑坡的主要影响因素;依据极限平衡原理,采用GeoStudio2007软件对高村排土场边坡稳定性进行建模计算,得出最小安全系数值1.346,说明高村排土场的稳定性较好。考虑滑坡发展、发生的时间因素,从影响排土场滑坡的内因和外因两方面,提出了露天矿排土场滑坡短期预警指标体系和中长期预警指标体系,划分了五个预警等级,并给出等级划分标准及相应的预警准则。针对预警指标的特点,应用未确知有理数的概念对短期预警指标赋权重,采用基于粗糙集的专家打分法对中长期顸警指标赋权重。根据短期预警指标的预警准则和权重值,建立了基于可拓理论的排土场滑坡短期预警模型,确定了排土场滑坡短期预警的经典域、节域和待评价物元,给出了关联函数的计算方法,由最大关联度和级别变量特征值得出排土场滑坡的预警等级。工程实例应用效果较好。整理了100组排土场滑坡的案例库,给出了基于案例推理(CBR)的排土场滑坡中长期预警方法,采用框架法表示滑坡案例,RBF神经网络和欧式距离相结合的算法来实现案例的检索,由用户和专家来完成案例的修正与调整,给出了案例学习和存储的阈值。工程实例应用效果较好。最后,采用Visual Basic6.0程序语言和Access数据库设计并开发了C/S模式的露天矿排土场滑坡预警管理系统,包括用户管理、数据管理、案例管理、预警管理和应急管理五大模块；主要实现了排土场滑坡的短期和中长期预警功能,软件界面友好、操作简便,在高村排土场进行试点应用,为企业的安全生产提供信息化、智能化和科学化的管理平台。
[Abstract]:With the rapid development of mining industry in our country, the number or scale of dump is increasing gradually. In recent years, landslide accidents in dump are more and more frequent. The research on the early warning method of dump landslide has important theoretical significance and profound practical influence on realizing the sustainable development of enterprises and social harmony and stability. Based on the analysis of 50 typical landslide accidents in recent 40 years at home and abroad, the main influencing factors of dump landslide are obtained, and the slope stability of Gaocun dump is modeled and calculated by using GeoStudio2007 software according to the principle of limit equilibrium. The minimum safety factor is 1.346, which indicates that the stability of Gaocun dump is good. Considering the development of landslide and the time factor of occurrence, from two aspects of internal cause and external cause, this paper puts forward the short-term early warning index system and medium and long term early warning index system of landslide in open pit dump, and divides five early warning grades. According to the characteristics of early warning index, the concept of unascertained rational number is applied to assign weight to short-term early warning index, and the expert scoring method based on rough set is used to assign weight to medium and long term warning index. According to the early warning criterion and weight value of short-term early warning index, a short-term early warning model of dump landslide based on extension theory is established, and the classical domain, segment domain and matter-element to be evaluated for short-term early warning of dump landslide are determined. The calculation method of correlation function is given, and the early warning grade of dump landslide is obtained from the eigenvalue of maximum correlation degree and grade variable. In this paper, the case base of 100 groups of landslide in dump is sorted out, and the method of mid-long term early warning of landslide in dump based on CBR (case based reasoning) is given. The algorithm of combining RBF neural network with European distance is used to realize the retrieval of the case, and the frame method is used to express the case of landslide and combine RBF neural network with Euclidean distance. The case correction and adjustment are completed by users and experts, and the threshold of case learning and storage is given. Finally, using Visual Basic6.0 program language and Access database, the paper designs and develops the landslide early warning management system of open pit dump based on C / S model, which includes five modules: user management, data management, case management, early warning management and emergency management. The short and long term early warning functions of dump landslide are mainly realized, the software interface is friendly, the operation is simple, the pilot application is carried out in Gaocun dump, and the information, intelligent and scientific management platform is provided for the safety production of enterprises.
【学位授予单位】：北京科技大学
【学位级别】：博士
【学位授予年份】：2015
【分类号】：TD76

【参考文献】