Rate of Penetration Optimization and Prediction Based on a Geological Label-Parameter Spectrum Framework

In response to the challenges of complex parameter coupling, rough formation description, and poor model interpretability in predicting and optimizing the rate of penetration (ROP) during drilling, this study aims to achieve geology-driven intelligent optimization of drilling parameters. A "Geological Label-Parameter Atlas" system integrating geological feature quantification and multi-parameter visualization analysis is proposed. Based on actual data from a block in the South China Sea, factor analysis is introduced to reduce multiple geological parameters, such as depth, pressure, seismic wave velocity, and lithology, into three common factors. Accordingly, 12 types of geological labels with clear physical meanings are constructed. The drilling parameter atlas, developed based on these geological labels and target ROP intervals, visually displays the reasonable value ranges and optimal values for over ten engineering parameters under different geological conditions. Based on this atlas, quantitative parameter optimization strategies can be formulated during actual drilling by comparing the deviation between current parameter curves and the atlas, while in the design stage, quantitative prediction of ROP intervals can be achieved by calculating the matching probability between design parameters and multiple sets of atlas.