中文摘要：

在山区进行MT勘探时,用规则网格有限元方法模拟起伏地形会受到限制.本文采用非结构三角网格可以有效地模拟任意二维地质结构,如起伏地形、倾斜岩层和多尺度构造等.正演引入自适应有限元方法,其在网格剖分过程中能根据单元误差自动细化网格,保证了正演结果的精度.将自适应有限元与Occam算法结合,且引用并行处理技术提高正反演计算速度.通过对比两个理论模型,讨论了地形对MT正演响应的影响;其次进行了不同地电模型带地形反演展示了本文算法的正确性和适用性;最后将该方法应用于实测MT数据处理,证明了自适应非结构有限元方法是复杂地形下处理MT数据的有力工具.

英文摘要：

The structured grid has its limit when simulating complicated topography,especially applying the MT method in mountainous areas where the effect of terrain cannot be ignored.Unstructured triangular grids permit efficient discretization of complex modelling domains such as those containing topography,dipping strata and multiple scale structures.To ensure the accurate result of forward modeling,we refine grids according element errors during gridding the model.We combine the adaptive finite element and the Occam method,and develop aparallel processing scheme that can efficiently improve the computational speed of forward modeling and inversion.By comparing two synthetic models,first we discuss the terrain effect by building two different models.Then forward modeling and inversion are conducted on a series of models with steep topography.The application of this approach to processing of real data from Southwest China shows that the adaptive unstructured finite element method is a powerful tool to analyze MT data of complex features such as steep surface topography.