Adaptive monitoring for geological drilling process using neighborhood preserving embedding and Jensen–Shannon divergence

Since the geological drilling process involves numerous variables and the relationships between them are also complex, it is not easy to implement an accurate description of operating conditions by conventional methods. Hence, it has a wide range of application possibilities for comprehensively monitoring the processes. An adaptive monitoring method for the geological drilling process using Neighborhood preserving embedding and Jensen–Shannon divergence is proposed in this paper. Firstly, the times-series clustering method identifies the operating modes under normal drilling conditions to construct operating zones for multiple operating modes. Then, in each sub-mode, manifold learning-based monitoring models are established. In addition, an adaptive updating criterion is developed for model updating to accommodate additional information from the accumulation of new running data. Finally, several comparisons are conducted utilizing historical data from a geothermal well worksite. The experimental result reveals that the proposed method can effectively monitor the operating performance in the drilling process with the highest accuracy of 91.11 % and the minimum monitoring delay of 12 s. The proposed method achieves much better effectiveness and flexibility through real-world process scenarios due to the distributed structural division and adaptive control limits in this paper.