Core-Log Depth Adaptive Matching Using RDDTW

Core-log depth matching is crucial in petrophysical interpretation and reservoir parameter prediction. The accuracy and reliability of core-log matching determine the quality of all subsequent petrophysical and reservoir parameter models. At present, this task is mainly performed manually by geological experts, and the process depends on expert experience and is time consuming and laborious. The existing methods ignore the matching of core-log data similarity and overall trend changes. Given this situation, this paper proposes regularized derivative dynamic time warping (RDDTW), which is used to extract and fuse the shape and trend information between the core and the log porosity curve to realize adaptive depth matching. In addition, the method includes a new weight function, called excessive warping regularized function (EWRF), which is proposed to eliminate the overfitting effect of excessive warping. The study applies the proposed methodology to a data set comprising 10 exploration wells in the Ordos Basin, with an average porosity of 10.31%. Utilizing particle swarm optimization (PSO), the approach iteratively refines optimization. A comparative analysis assesses the method’s effectiveness in core-log matching tasks against traditional Pearson correlation coefficient (PCC) techniques and manual interventions. The results show that RDDTW can realize core adaptive relocation more efficiently and accurately in various strata. Compared with the PCC, the proposed method increases the porosity R2 by 9.62% and decreases the porosity mean square error by 1.57%.