Location

  • Offshore, Sureste basin, Gulf of Mexico.

Challenge

An operator was preparing to drill an S-shaped exploration well in an area of the Gulf of Mexico that was unfamiliar to them. Based on drilling programs from the region, the operator felt the well presented a high risk of stuck pipe, however, no pertinent offset well data existed for reference.

Solution

The operator chose to implement Exebenus Spotter Stuck Pipe Agent with the expectation that the real-time predictive machine learning (ML) solution would alert the operations team to stuck pipe risks and allow the rig crew to take timely preventive actions.

Results

Exebenus Spotter flagged several potential stuck pipe incidents, which prompted communication between the office and rig in sufficient time for the team to decide on the best course of action. No costly stuck pipe incidents occurred.

Products Used

Drilling an S-shaped well poses a high risk of stuck pipe. In this case, no offset wells existed for reference, elevating the uncertainty. Exebenus Spotter Stuck Pipe Agent was deployed and delivered reliable real-time outputs with minimal false warnings. The 162-day drilling program was completed with no stuck pipe incidents.

OVERCOME HIGH RISK AND UNCERTAINTY TO AVOID COSTLY STUCK PIPE INCIDENTS

In the Sureste basin, Gulf of Mexico, an operator was planning to drill an exploration well in 51 m of water depth. Based on similar drilling programs conducted elsewhere in the region, the operator felt the well’s placement and S-shaped trajectory presented significant risk of stuck pipe.

The reservoir top was situated in the Upper Miocene, immediately across a growth fault (Figure 1). Growth faults are often associated with unpredictable formation changes and unstable geological conditions, making this area particularly prone to borehole instability, poor hole cleaning and an increased likelihood of stuck pipe incidents during drilling operations.

There were no offset wells in the area to learn from, elevating the level of uncertainty. To manage the risk, the operator decided to test Exebenus Spotter Stuck Pipe Agent during drilling operations.

The operator had no prior experience using an AI/ML predictive warning system like Exebenus Spotter, and the engineering team was concerned about the potential of false warnings that would trigger unnecessary reaming or tripping out operations, thus increasing drilling time and cost. Furthermore, the mid-size company could not dedicate a 24/7 monitoring team to the project, and was concerned about responding effectively to the system’s real-time advice.

Figure 1. Post-well geological column. The Stuck Pipe Agent identified risks during multiple independent passes through the depth zone, ranging from 2443 m to 2520 m. This was interpreted as an indication of unstable hole conditions when approaching the Main Growth Fault between the Pleistocene and Miocene formations.

EXEBENUS SPOTTER AGENT USES REAL-TIME DATA TO IDENTIFY RISKY ANOMALIES

Exebenus Spotter Stuck Pipe Agent was deployed. The agent was plugged into the operator’s real-time WITSML server and monitoring system, using readily available surface data. It required no downhole data, no project specific training and no timeconsuming configuration, despite the complexity of the well.

The agent remained active throughout the project, monitoring the drilling program from 36 in to total depth (TD), including an 8 ½ in sidetrack.

To overcome the operator’s concerns about its capacity to monitor the software, the system was set up to display results directly to the mudlogger on the rig and to drilling engineers in the office. One engineer was responsible for watching the real-time advice at critical points and communicating with the mudlogger and rig supervisor, who would decide on corrective actions.

STUCK PIPE AGENT PREDICTS AND PREVENTS STUCK PIPE INCIDENTS

The Stuck Pipe Agent detected early signs of potential stuck pipe throughout the 162-day drilling program, including anomalies related to pipe and fluid drag. The warnings allowed the rig supervisor to take timely corrective action. All potential stuck pipe situations were averted.

 

Concerns about false warnings eliminated

In the setup, the software system’s sensitivity was adjusted to issue the right balance of warnings, thereby reducing the number of false warnings. The client experienced a very low rate of false warnings, averaging only 0.15 per day (24 false warnings overall).

 

Visual observation at the shakers confirmed real-time predictions

A near stuck pipe situation occurred as the shoe of the 20 in casing was being drilled out with the 17 ½ in bottom hole assembly (BHA).

The time-based screenshot (Figure 2) shows general information in the top track, such as block position and standpipe pressure (SPP), followed by tracks of the Stuck Pipe Agent’s three supporting models—Hole Cleaning (HC), Differential Sticking (DS) and Mechanical Sticking (MS).

While drilling out the shoe, the Hole Cleaning (ECD) track flagged two warnings, and the Differential Sticking track flagged one warning. These warnings indicated restricted movement and fluid friction, most likely caused by an accumulation of large cement cuttings around the BHA.

Figure 2. The Differential Sticking (DS) model provided a warning due to string static. While drilling the shoe track, cement cuttings resulted in constrained pipe movement and fluid friction. Subsequently, the Hole Cleaning (HC) model provided warnings that highlighted the obstructions in pipe movement and fluid circulation caused by the cuttings.

Following the warnings, the drilling team conditioned the fluid and reciprocated and rotated the string for 20–30 minutes to move the cuttings above the BHA. Once drilling resumed, the driller circulated and performed one bottoms up, which revealed large cement cuttings at the shakers and provided visual validation of the system’s warnings (as reported in the daily drilling report).

No further hole cleaning or static friction warnings were raised, which was taken to mean that the risk had been mitigated by the intervention.

 

Drilling program completed successfully with no stuck pipe incidents

Exebenus Spotter Stuck Pipe Agent worked out-of-the-box, without customized ML training. Throughout the drilling and casing operations, the system delivered reliable real-time outputs. It highlighted numerous potential risks and confirmed their absence when mitigating actions were performed.

The project highlighted the critical importance of full-time wellbore situation awareness, and convinced the operator that establishing a 24/7 real-time operations monitoring center (RTOC) would further enhance the proven value of the Exebenus Spotter solution.