A Novel Data Science Approach to Borehole Dysfunction Analysis

Scott Hess, Kristopher A. Innanen, Roman Shor, Alex Vetsak

During drilling activities, monitoring the quality of the borehole is an essential task. Borehole dysfunctions such as constrictions, ledges, and differential sticking can cause significant amounts of non-productive time that decreases overall operational performance. The dysfunctions are most troublesome during tripping operations when the drill string is moved in and out of the borehole. Identification of dysfunctions before significant operational issues occur, such as stuck pipe, can allow proactive mitigating actions to reduce the impact on the tripping operation. Modelling methods for expected sliding friction for good borehole conditions provide a baseline for hook load operating parameters during tripping out of the borehole. Assuming the baseline hook load estimation is somewhat accurate, the anomalous high hook loads above baseline, referred to as overpull, provide measurements that should capture the resistance in the borehole due to dysfunctions. The focus of this study is to utilize overpull signatures and the drill string configuration to produce a resistance depth profile that can provide better depth resolution to place dysfunctions along the wellbore and also characterize the dysfunction mechanism. This paper represents the initial steps of developing a forward modelling strategy using a source signal (i.e. the drill string) convolved with a resistance signal (i.e. the dysfunctions) to produce overpull signals. Initial tests show promising similarities compared to overpull real data.