提高煤矿微震定位精度的台网优化布置算法  

OPTIMIZATION ALGORITHM OF NETWORK CONFIGURATION FOR IMPROVING LOCATION ACCURACY OF MICROSEISM IN COAL MINE

作　　者：巩思园[1,2] 窦林名[2] 马小平[1] 牟宗龙[3] 陆菜平[3] 

机构地区：[1]中国矿业大学信息与电气工程学院,江苏徐州221116 [2]中国矿业大学煤炭资源与安全开采国家重点实验室,江苏徐州221116 [3]中国矿业大学矿业工程学院,江苏徐州221116

出　　处：《岩石力学与工程学报》2012年第1期8-17,共10页Chinese Journal of Rock Mechanics and Engineering

基　　金：国家重点基础研究发展计划(973)项目(2010CB226805);江苏高校优势学科建设工程资助项目(PAPD);江苏省博士后科研资助计划(1002004B)

摘　　要：针对大规模台网布置组合优化问题,建立台网优化布置目标函数,提出包括模型数据准备模块、遗传算法求解模块和台网布置方案定位能力评价模块的微震台网布置方案求解模型。模型数据准备模块中,首先根据综合指数法和台站候选点确定原则为遗传算法求解模型提供初始参数,然后由评价模块对求解的最优方案进行定位能力评价。遗传算法求解模块中采用入选候选点在前,落选候选点在后的台网布置方案自然数编码个体表达形式,结合创建的用于保证监测小能量震动的惩罚函数,构建台网布置方案个体的适应度函数。为防止算法过早收敛,使用混合交叉和变异操作算子,规定交叉和变异范围必须包括入选候选点,以提高算法效率。试验和现场应用结果表明,该算法能够快速找到最优解,且计算时间不随组合方案规模的增加而显著增加。数值评价技术可验证采用该算法求解的台网布置方案较优,显著降低重点监测区域内的震源定位误差,其最大值降幅达230 m。For optimizing the large-scale network configuration of combinatorial optimization problem, a model, including the blocks of input data preparation, genetic algorithm and location capability assessment of network, is constructed to solve the establishment of optimal objective function of microseismic network layout. First, initial parameters are provided to genetic algorithm by the data preparation block based on the comprehensive index method and the general principle of choosing candidate points; and then the locating ability of the obtained solution is evaluated by the assessment block. In the block of genetic algorithm, natural number coding method is utilized to express the individual of microseismic network layout, in which winning candidates are in front of the losing to keep more information. Combining with the created penalty function, which is used to guarantee the capability of monitoring and recording weak tremor, the individual fitness function is built to evaluate the performance of network. In order to prevent the premature convergence of algorithm, crossover and mutation operators are mixed and operated in the range of containing winning candidates to increase computing efficiency. The application results from on-site and experiment show that, the genetic algorithm can quickly find the best solution and its computing time does not increase significantly with the expansion of combination. The network configuration solved by the genetic algorithm is better than the layout used currently, which is demonstrated by the numerical emulation technique, and significantly decrease the hypocenter location errors in the key monitoring areas; and the maximal reduction of location errors reach to 230 m.

关 键 词：采矿工程 微震 台网布置 遗传算法 自然数编码 惩罚函数 适应度函数 

分 类 号：TD32[矿业工程—矿井建设]