Abnormal torque and drag (T & D), which commonly includes overpull, underpull, and high-torque load, are indications of excess frictional effects between the drillstring and the wellbore walls. Numerous conditions can cause these effects, including tight hole, differential sticking, poor hole cleaning, key seats, etc. Failure to observe these anomalies will cause excessive wear on the drillstring and can eventually lead to severe stuck pipe conditions. A new workflow for monitoring T & D is presented in this paper. This workflow, developed for a real-time monitoring system, allows for monitoring various types of data from multiple sources to be received without delay, aligned, and synchronized.

The workflow requires standard surface measurements and contextual data, which are available on most wells. Three main segments with respect to the computation phase are included in the workflow. These segments include T & D measurement points statistics, T & D modelling and calibration, and abnormal T & D alarms. The measurement points are selected from relevant operations and summarize the statistics at different granularities to meet the different objectives, such as the classical broomstick plot or alarm triggering. A hybrid T & D modelling framework was designed to predict the hook load and surface torque accordingly. This framework combines the mathematical capability of a stiff-string model using a finite element method and the experience acquired from obtaining the drilling data. As a result, the physical model can be automatically calibrated and driven by real-time data to compensate the hook load offset due to uncertain variables or inaccurate inputs. An alarm-triggering logic can be developed to capture anomalies based on a comparison between the measured and predicted values.

The new workflow is fully automatic without a need for manual calibration and fixed thresholds. Furthermore, the workflow adjusts itself according to real-time observations, which makes it adaptive to the changing conditions of the well being drilled. The efficiency and reliability of the anomaly detection heavily rely on the input data quality in the perspective of stream computation. In this paper, two case studies are presented containing the results produced by streaming actual well data in a time series manner. The case studies demonstrate the usability and reasonableness obtained by the user when handling the actual operation scenarios.

The work presented in this paper was developed to meet the increase in digital transformation by the oil and gas industry and demonstrates the best use of data for drilling optimization.