摘要：This paper proposes the concept of projection curves based on the theory of CT image reconstruction to probe the internal structure of the working panel prior to formal mining of the working *** well as reducing costs,this method provides safe and efficient excavation of the working *** to the results of the numerical model and the actual working panel,the new method has been proven to be accurate in detecting the location of the fault that extends into the *** faults of the internal working panel,as well as the start and end points of the fault,can be detected by this *** practice has proven that the method is highly reliable,has a highly decisive impact on faults for coal mining,and can be used to guide the safe mining of the working panel.

摘要：为满足煤矿工作面智能精准开采对地质异常分布精准感知的需求,进一步提高工作面电磁波CT反演的精准性,提出了多源先验信息约束下的电磁波CT智能反演模型。在对工作面电磁波CT反演稀疏矩阵方程进行泛函优化转换的基础上,分析了基本遗传算法(Single Genetic Algorithm,SGA)的寻优进化机理,探讨了不同遗传参数对SGA智能算法搜索优化性能的影响,设计了基于多种群自适应遗传算法(Multi-population Adaptive Genetic Algorithm,maGA)的工作面地质异常电磁波CT反演目标函数求解机制。除此之外,提出了范围约束、就近约束和平均值约束3种约束模型,将生产过程中揭露的巷探、钻探及回采等多源先验地质信息作为约束条件,不断调整地质异常反演目标函数搜索的进化方向,最终形成了基于maGA智能算法和多源先验信息约束的电磁波CT反演模型,并通过数值模拟和工程实例进行了测试和验证。试验结果表明:maGA智能算法能够充分结合多种群协同进化和遗传参数自适应调节策略,相比于单一的遗传算法具有更高的全局及局部搜索准确性和稳定性;随着外部约束数据的增加,基于平均值约束的maGA智能算法能够获得更优的电磁波CT反演目标函数收敛结果。以山西大同矿区8208工作面内电磁波CT地质异常反演和解释为例对所建模型进行检验,预测结果与实际揭露结果基本一致,进一步验证了多源信息约束下的智能反演方法在工程实践应用中的可靠性,为工作面内地质异常的精准反演和预测提供了一种新的思路和方法。