我国许多矿区都有小窑采空区的存在，小窑采空区边界不清且富水性较强时，煤矿生产必须对由此引起的安全隐患重点排查。目前常用探测小窑采空区的物探方法分为地面和矿井两类，但地面探测受到探测深度和精度的限制，很难指导实际的掘进生产工作。矿井随掘超前探测技术是探查迎头前方是否存在小窑采空巷道的有效方法之一，也是目前随掘探测应用最为广泛的手段。但由于探测资料处理及解释的不充分，使得资料的反演精度受到一定限制。 基于上述背景，本文从点源人工场和超前探测的基本理论出发，在查阅相关文献，结合实习期间现场经验的基础上，应用COMSOL有限元数值模拟软件建立了合适的地电模型，讨论了改变激励电流和改变低阻巷道空间位置时模型的电位分布规律。得出观测电位大小与激励电流大小成正比；以及低阻巷道的异常响应必须经过处理才能较清晰识别的结论。而后对计算出的数据进行处理，形成探测方法成果，并在此基础上提出了改进“归一化”校正的思路，对两种处理方法进行了对比讨论。得出异常响应大小与低阻巷道距离迎头远近成反比；巷道位置距迎头大于30m时会带入难以避免的“假异常”；经过改进后“归一化”的校正能够使得异常特征更易分辨的结论。 最后，通过乡宁县某煤矿的几例实验，对比讨论了两种处理方法的优缺点及效果。本文提出的校正办法是建立在理想条件下的，而实际的探查环境往往非常复杂。因此，该技术在实际应用中的效果仍需要进一步的研究和总结。尽管如此，通过对以往实例的总结发现该校正办法仍有较为明显的效果。

Coal mines goaf are being in mining areas in China, which the boundary of coal mine goaf is not clear. Meanwhile, when water yield property is stronger, the safety production of coal mine goaf must be checked. At present, the methods of geophysical exploration in coal mines goaf are divided into two categories, the surface ground and mine. But the surface detection is limited by the detection range, the actual excavation production is difficult to guide. Underground digging in advanced detection technology is one of the effective methods, which is to probe ahead if there is a coal mine goaf roadway, and is also the most widely used method. But because short of the detection methods and data processing, the data of inversion interpretation is limited. This report mainly focuses on the background the researcher mentioned before. This study from the point of artificial field and the basic theory of advanced detection, establishes the geoelectric model that using the finite element numerical simulation software COMSOL, which is combining the experience of field during the internship. This model comes to the conclusion that observation voltage becomes to the direct ratio with excitation current, and the exceptional response of low resistance of roadway must be treated to a clear identification. This study discusses the changed excitation current and low resistance body potential distribution regularity of the spatial location model, and then disposing the data which is calculated by theory. Finally, the detection results are formed. This study put forward the idea which is to improve the "normalization" correction, comparing two kinds of methods. This can conclusion that exceptional response becomes to the inverse ratio with the head-on distance of low resistance roadway. When the roadway position more than 30m from head-on, the "false anomaly" is hard to avoid. After improvement of correction in "normalization", the anomaly characteristics can make easier to distinguish on before. Introducing a few experiments by Xiang Ning coal mine, this study discusses the comparative advantages and disadvantages of two methods, and the problems should be paid attention in practice, finally a conclusion can be formed. The actual detection environment tend to be very complex, this paper that puts forward the correction method is based on ideal conditions. Therefore, this technology effect still need study and summary in the practical application, even so, the summary of previous instance found that this method is still have obvious effect.