Early Symptom Detection on the Basis of Real-Time Evaluation of Downhole Conditions: Principles and Results From Several North Sea Drilling Operations

摘要

During drilling operations, downhole conditions may deteriorate and lead to unexpected situations that can result in significant delays. In most cases, warning signs of the deterioration can be observed in advance, and by taking proactive actions, drillers can avoid serious incidents such as packoffs or stuck pipes. A new analysis methodology, relying on an automatic real-time computer system, has been developed to detect those early indicator conditions.The methodology involves constantly computing the various physical forces acting inside the well (mechanical, hydraulic, and thermodynamic). These physical forces are coupled by an automatic model calibration, which then gives a reliable picture of the expected well behavior. Through analysis of the deviations between modeled and measured values, an estimation of the current state of the well is derived in real time. Changes in the well condition are an early warning of deteriorating well conditions. This paper precisely describes the real-time analysis and the results during some drilling operations.The software has been used for monitoring 15 unique wells located in five different North Sea fields. All major situations were signaled in advance at different event time scales: Rapidly changing downhole conditions (such as pulling a drillstring into a cuttings bed) were typically detected 30 minutes ahead of the actual event, medium-duration deteriorations were detected up to 6 hours before the incident, and slow-changing downhole conditions were signaled up to 1 day in advance. Several examples that illustrate the detected incidents over distinct time periods are described.The availability of good-quality real-time data streams makes it possible to implement such analysis tools in an integrated operation setup. Early symptom detection can be used to make decisions in a timely fashion, on the basis of quantitative performance indicators rather than subjective feelings and personal experience.