Switching Prediction of Sliding and Rotating Modes in Compound Directional Drilling Based on Sliding Window

Compound directional drilling consists of sliding and rotating modes. The two modes alternate to achieve trajectory extension, where the sliding mode achieves trajectory deviation and the rotating mode maintains the trajectory. When to switch between the two modes plays an important role in obtaining a satisfactory drilling trajectory. In this study, the switching of the two modes is considered as a prediction problem and is achieved by using the random forest algorithm. First, the collected raw data were preprocessed, where the linear interpolation was used to fill the missing values and the feature importance assessment was applied to select seven decision variables. Secondly, the pre-processed dataset consisting of the seven decision variables was divided into multiple data blocks through the sliding window mechanism. Lastly, the random forest algorithm was adopted to predict the results for the data blocks. A case study shows that our method provides an effective solution to predicting when to switch the two modes in compound directional drilling.