Research on the Correlation Analysis Between Logging Parameters and Rock Drillability Based on Random Forest

The rock drillability grade value is a comprehensive parameter that measures the ease or difficulty of rock fragmentation during drilling. Determining this parameter is crucial for guiding bit selection and predicting drilling speed. There exists a strong correlation between certain logging parameters and the rock drillability grade value. To explore this correlation, a common practice is to use the Pearson correlation coefficient method, which mainly investigates the linear relationship between variables. However, relationships among complex formation characterization parameters are often nonlinear. To evaluate the pros and cons of the two analytical methods when dealing with nonlinear and complex data, a comparison was made, and the random forest algorithm was preferentially selected as the analytical tool to enhance the precision of the prediction model. Thus, this paper analyzes the correlation between logging parameters and the rock drillability grade value based on Pearson correlation analysis and the random forest algorithm. Using the Matlab programming language, logging parameter correlation analysis software named Correlation 1.0 was developed. The random forest algorithm was preferred to screen out logging parameters with high correlation to the drillability grade value (drilling depth and formation density), providing a new approach for establishing a high-precision formation rock drillability model.