基于优化常规双参时间函数的煤矿开采动态沉陷预测研究  

Dynamic Subsidence Prediction of Coal Mining Based on Optimizing Conventional Dual Parameter Time Functions

作　　者：周华安 戴广礼 李怀展 刘超 元亚菲 ZHOU Hua′an;DAI Guangli;LI Huaizhan;LIU Chao;YUAN Yafei(School of Environment and Spatial Informatics,China University of Mining and Technology,Xuzhou 221116,China;The Fifth Geological Brigade of Jiangsu Provincial Bureau of Geology and Mineral Resources,Xuzhou 221004,China;Collaborative Innovation Center for Resource Utilization and Ecological Restoration of the Old Industrial Base in Jiangsu Province,Xuzhou 221116,China)

机构地区：[1]中国矿业大学环境与测绘学院,江苏徐州221116 [2]江苏省地质矿产局第五地质大队,江苏徐州221004 [3]江苏省老工业基地资源利用与生态修复协同创新中心,江苏徐州221116

出　　处：《金属矿山》2025年第3期148-157,共10页Metal Mine

基　　金：国家自然科学基金区域重点项目(编号:U21A20109);徐州市青年科技人才项目(编号:KC23020);江苏省地质局重点科研项目(编号:2021KY08)。

摘　　要：煤矿开采动态沉陷预测对于保证开采过程中地面建(构)筑物安全及实施边采边复具有重要作用。针对基于常规双参时间函数(Weibull、MMF、Logistic及Bertalanffy)进行开采动态沉陷预测时存在的预测精度偏低及模型拟合程度不高的问题,提出了一种自适应混合寻优人工鱼群算法(Adaptive Hybrid Optimization Artificial Fish Swarm Algorithm,AHO-AFSA)实现双参时间函数参数最优值求解。该算法采用自适应视野和步长对人工鱼群算法(Artificial Fish Swarm Algorithm,AFSA)的固定视野和步长进行改进,并将相对成熟但易陷入局部极值的粒子群算法与尚未广泛应用于地表动态沉陷预测的人工鱼群算法(AFSA)相结合,实现了利用寻优算法求取双参时间函数参数精度的提升。同时以静态概率积分预测模型为基础,通过复化辛普森公式优化地表点静态沉降值的求解过程,并基于此构建了煤矿开采地表动态沉陷预测理论模型。通过实测数据验证得出:利用优化求解双参的自适应混合寻优人工鱼群算法,基于Weibull、MMF、Logistic和Bertalanffy 4种时间函数模型的总体相对误差精度分别提升了2.44%、0.35%、1.48%和3.11%,总体拟合误差在10.3%以内,算法用于反演双参时间函数参数进行动态沉陷预测具有较高精度。研究成果对于基于寻优算法反演双参时间函数参数的煤矿开采动态沉陷精准预测具有参考价值。Predicting dynamic subsidence in coal mining is important for ensuring the safety of surface buildings(structures)during the mining process and simultaneously implementing mining and reconstruction.In response to the problems of low prediction accuracy and low model fitting in predicting dynamic subsidence in coal mining based on conventional dual parameter Weibull,MMF,Logistic,and Bertalanffy time functions,an Adaptive Hybrid Optimization Artificial Fish Swarm Algorithm(AHO-AFSA)to solve the optimal values of dual parameter time function parameters is proposed.The fixed field of view and step size of the Artificial Fish Swarm Algorithm(AFSA)is replaced by the adaptive one.And the relatively mature particle swarm algorithm that is prone to getting stuck in local extremes is combined with the Artificial Fish Swarm Algorithm(AFSA)that has not yet been widely used in predicting surface dynamic subsidence to achieve an improvement in the accuracy of using optimization algorithms to obtain parameters of the dual parameter time function.Based on the static probability integral prediction model,the solution process of static settlement values of surface points was optimized with the compound Simpson formula,and the theoretical model for predicting dynamic surface settlement in coal mining was constructed based on this.Through experimental data validation,it was found that using the Adaptive Hybrid Optimization Artificial Fish Swarm Algorithm with optimized dual parameters,the overall relative error accuracy based on four time function models,Weibull,MMF,Logistic,and Bertalanffy,was improved by 2.44%,0.35%,1.48%,and 3.11%,respectively.The overall fitting error was within 10.3%.The algorithm has high accuracy in inverting dual parameter time function parameters for dynamic subsidence prediction.The research results have reference values for the precise prediction of mining dynamic subsidence based on the inversion of dual parameter time function parameters using optimization algorithms.

关 键 词：开采沉陷 双参时间函数 概率积分法 动态预测 人工鱼群算法 

分 类 号：TD325[矿业工程—矿井建设]