基于LM-LSO算法的磷石膏堆场边坡形态与稳定性智能优化方法  

Intelligent Optimization Method for Slope Morphology and Stability of Phos-phogypsum Stack Based on LM-LSO Algorithm

作　　者：李垚熠 陈国庆[1,2] 石明汉 周玉新 聂闻 郑成 Yaoyi LI;Guoqing CHEN;Minghan SHI;Yuxin ZHOU;Wen NIE;Cheng ZHENG(State Key Laboratory of Geohazard Prevention and Geoenvironment Protection,Chengdu University of Technology,Chengdu 610059,Sichuan,China;College of Environment and Civil Engineering,Chengdu University of Technology,Chengdu 610059,Sichuan,China;Yuan’an County Liaoyuan Mining Co.,Ltd.,Yichang 443000,Hubei,China;Sinosteel Ma’anshan General Institute of Mining Research Co.,Ltd.,Ma’anshan 243000,Anhui,China)

机构地区：[1]成都理工大学地质灾害防治与地质环境保护国家重点实验室,四川成都610059 [2]成都理工大学环境与土木工程学院,四川成都610059 [3]远安县燎原矿业有限责任公司,湖北宜昌443000 [4]中钢集团马鞍山矿山研究总院股份有限公司,安徽马鞍山243000

出　　处：《黄金科学技术》2024年第5期882-893,共12页Gold Science and Technology

基　　金：国家自然科学基金项目“陡倾顺层滑坡加速破坏的微震—变形时序特征”(编号:42372326)资助。

摘　　要：大量欠稳定边坡阻碍了人类工程活动,亟需采用合理的优化方案来确保工程安全。为此,提出了基于列文伯格—马夸尔特(Levenberg-Marquardt,LM)算法和光谱优化算法(Light Spectrum Optimizer,LSO)的边坡形态智能优化方法(LM-LSO)。首先,采用LM算法对样本数据进行非线性拟合;其次,采用拒绝不可行解方法处理稳定性系数的约束问题;最后,利用LSO搜索最优方案,并与另外4种算法进行比较分析。该方法应用于四川省某磷石膏堆场边坡优化,得到了最小挖方量的优化设计方案。该方法综合考虑了边坡优化的经济性和稳定性,保障了堆场的堆填方量和安全。Mining operations and engineering projects frequently produce numerous unstable artificial waste slopes,which pose significant hindrances to human activities.Consequently,the implementation of effective slope morphology optimization is essential for ensuring engineering safety and maximizing landfill capacity.The development of an efficient,accurate,and scientifically robust method for slope morphology optimization holds substantial theoretical and practical importance for the management of unstable artificial slope projects.Consequently,we propose an advanced slope morphology optimization method,termed LM-LSO,which integrates the levenberg-marquardt(LM)algorithm with the light spectrum optimizer(LSO).Initially,stability coefficients for various slope configurations were computed utilizing the limit equilibrium method,with cut and fill volumes estimated based on the differences in two-dimensional profiles,thereby generating the sample data.Subsequently,cross-product terms among variables were incorporated to capture nonlinear relationships,and the LM algorithm was applied for the nonlinear fitting of the sample data.In conclusion,an infeasible solution rejection method was employed to address stability coefficient constraints,optimizing the solution using the LSO algorithm and benchmarking it against four other algorithms.This approach was implemented to optimize the terrace height,width,and slope angle of a gypsum stack slope in Sichuan Province.The objective was to minimize the excavation volume while ensuring adherence to stability regulations.The final optimized design parameters include terrace heights(h)of 6.38 m,step widths(l)of 4 m,step inclinations(α)of 25.11°,and a minimal excavation volume(Vmin)of 298.92 m2.The method comprehensively considers the economic feasibility and stability of slope optimization,ensuring both landfill capacity and safety.

关 键 词：LEVENBERG-MARQUARDT算法 光谱优化算法 非线性拟合 稳定性系数 挖方量 磷石膏堆场 

分 类 号：TD216[矿业工程—矿井建设]