北京工业职业技术学院学报

在瞬变电磁法一维反演领域，阻尼最小二乘法(DLS)因算法简单、计算效率高而被广泛应用，但初始模型构建难题限制其发展。因此，提出一种改进流程：先借助二分搜索算法拟合全程视电阻率，结合烟圈法反演估算初始模型；再运用G-S逆拉普拉斯一维正演试算，确定初始模型；最终利用DLS对实测数据进行反演。DLS反演结果能有效区分地层，与实际地质吻合度较高，显著提升了DLS在多地层反演中的稳定性和准确性。

In the field of one-dimensional inversion for Transient Electromagnetic Method(TEM), the Damped Least Squares(DLS) algorithm has been widely adopted due to its simplicity and computational efficiency. However, its development has been constrained by challenges in initial model construction. This study proposes an improved workflow. First, the binary search algorithm is employed to fit full-process apparent resistivity combined with Smoke Ring Method inversion to estimate the initial model. Subsequently, G-S inverse Laplace one-dimensional forward modeling is applied to verify and refine the initial model. Finally, DLS is implemented for measured data inversion. The DLS-derived outcomes can effectively differentiate stratigraphic layers and exhibit a high degree of consistency with actual geological columnar sections. This significantly enhances the stability and accuracy of DLS in multi-layer inversion.