Coal-rock Identification While Drilling Method and Application Based on Multi-source Data Fusion

The technology of intelligent and accurate detection while drilling is an important support in the geological support system of intelligent and green coal mining. In view of the difficulty of coal and rock identification and information fusion during underground drilling in coal mines, this study deeply analyzes the change law of natural gamma while drilling and drilling rig operation parameters under different coal and rock conditions. And the multi-source drilling information fusion mode is explored to realize the effective fusion of multi-source monitoring data. A new coal-rock identification method based on support vector neural dynamics (SVND) model is proposed, and a recognition model with soft interval and kernel function is constructed. The model recognition simulation experiment is compared and analyzed. The results show that compared with random forest and decision tree methods, SVND model constructed by multi-source data sets has higher accuracy in coal and rock identification, and can realize accurate identification of typical strata with relatively stable lithology changes. The research results effectively improve the accuracy and reliability of coal and rock identification while drilling, and provide a new technical way for geological transparency.