Stuck pipe: the early signs and the three mechanisms

Stuck pipe rarely arrives without warning. The string usually tells you hours ahead, in the drag at a connection, the pressure while drilling, or the time it stands still against permeable rock. This article sets out the three mechanisms, the signs that belong to each and how to score them, then replays a tight-hole event in a horizontal well where the first alarm came seven hours before the worst pull.

Three ways a string gets stuck

The freeing worksheet used on most rigs sorts stuck pipe into three mechanisms, because each needs a different response.

Pack-off and bridging. Solids fill the annulus around the string: cuttings that were never lifted out of a horizontal hole, cavings from a shale that is failing, junk or cement. Circulation becomes restricted, then impossible. It often happens while pulling out or just after the pumps start.

Differential sticking. The string lies still against a permeable formation with the mud pressure above the pore pressure. The filter cake seals around the contact, and the overbalance acting on that sealed area holds the pipe against the wall. It needs three things together: overbalance, permeable rock and a string that is not moving. Circulation stays free, which is how you tell it apart.

Geometry. The hole is smaller or more crooked than the string can pass: a dogleg, a ledge, a key seat, an undergauge section, or a stiff BHA meeting a change in angle. Reactive or overstressed shale belongs here on the worksheet, because it shows up as tight hole. DrillIQ watches it as a fourth mechanism, wellbore instability, because its cure is mud weight and inhibition rather than reaming.

The early signs, and when to read them

Each sign means something only in the right rig state. Pressure while the pumps are on, drag when the pipe moves, still time when it does not. These are the checks DrillIQ runs every 15 seconds, grouped by mechanism.

Most signs need an expected value, such as the pick-up a soft-string model predicts or the pressure the last connections gave at this flow. So a torque and drag model calibrated on the well's own connections sits under every stuck-pipe watch; the method is in our torque and drag worked example.

Scoring the signs

Each check compares one measured value with two limits, Watch and Warning, and turns it into a severity between 0 and 1. The risk score on the screen is the worst severity times 100.

v

the measured value: pressure over baseline, friction over the section's, minutes still as a share of the allowed time

W, A

the Watch and Warning limits for that check

s

severity: 0.30 at Watch, 0.55 at Warning; 0.85 is kept for a string that no longer moves as asked

Three rules turn severities into a verdict. A mechanism takes its worst check. Two checks at Watch make a Warning. A pack-off needs two kinds of sign to reach Warning, pressure with torque, returns or drag, because one gauge alone drifts for ordinary reasons. Alarms are suppressed in the slips and on surface, need three 15-second cycles to open, and close only after five minutes back at low, so a pump break does not split one problem into three alarms.

Differential sticking has its own arithmetic. The overbalance at the bit is

with densities in SG, TVD in metres and ΔP in bar. DrillIQ's default allowed static time against permeable rock is 60 minutes up to 10 bar, 20 minutes up to 20 bar, 10 minutes up to 35 bar and 5 minutes beyond. Shale and salt get four times as long. Watch comes at half the allowed time, Warning when it runs out.

The example well

The example is the same horizontal well as our torque and drag article, set in conditions typical of North Oman: a carbonate reservoir at 1,500 m TVD under a sequence of reactive shales, a water-based drill-in fluid and a 6⅛″ lateral from the 7″ liner shoe at 1,760 m MD. Inside the reservoir a thin shale streak at 2,180–2,183 m has started to cave since the bit crossed it. We replay the last eleven hours before a tool failure forced a trip.

A hand check of the numbers

Four of the engine's numbers, worked out on paper:

1. Pore pressure. 2,175 psi at 1,500 m TVD is 2,175 / (1.4223 × 1,500) = 1.0194 SG.
2. Overbalance. (1.12 − 1.0194) × 0.0981 × 1,500 = 14.8 bar. That falls in the 10–20 bar band, so the string may stand still for 20 minutes against the carbonate. A 10-minute flow check uses half of it, which is Watch.
3. The first alarm. At the 2,195 m connection the friction that matches the pick-up is 0.405 against the section's 0.335: a rise of 0.070, with limits of 0.06 and 0.10. The severity is 0.30 + 0.25 × (0.070 − 0.06) / 0.04 = 0.36, a risk of 36 and Watch.
4. The Warning. At 2,224 m the friction is 0.500, a rise of 0.165, past the 0.10 limit: 0.55 + 0.20 × 0.065 / 0.10 = 0.68, a risk of 68.

The engine's own string model is simpler than a full torque and drag run, so its baseline reads 0.34 where the full model reads 0.25 on the same stands. The alarm reads the rise from the well's own baseline, and both models agree that the 2,224 m stand picked up about four times the drag of the 2,195 m one.

Replaying the event

We fed the eleven hours, 5-second surface data, through DrillIQ's Stuck Pipe engine exactly as it runs live, and counted the alarms the crew would have seen.

Read it from the left. While drilling from 2,160 m every mechanism sits low; instability holds near 15 because the mud is 0.10 SG over the collapse gradient. Then the connections start to drag: the 2,195 m stand picks up 2.0 klbf over the model and the 2,224 m stand 8.4 klbf, and geometry, which carries the friction checks, climbs to Watch and then Warning. The pack-off signs follow inside the open Warning, and the trip adds a tight spot and a flow check.

The worst pull came at 06:04, 26 klbf over pick-up at 2,203 m. The first Watch had come 7 h 4 min before it and the first Warning 4 h 21 min before. Nothing alarmed outside the event. Four hours is long enough to change the plan: circulate the hole clean before the trip, pump out or back-ream through 2,180–2,240 m instead of pulling on elevators, and keep the overpull well inside the limit while working the string.

In a caving shale the drag spoke first. Cavings in the hole show as drag when the BHA is pulled through them at a connection, and only later pack the annulus while drilling. The first pressure and torque signs came at 02:26, 3 h 38 min before the worst pull, about half the warning the drag gave.

The pack-off signs up close

From 02:15 the pressure while drilling climbs above the engine's expected 2,900 psi, the baseline from the run's last connections at the same flow. At 02:26 the drilling torque reaches its Watch line, 1.0 kft·lbf over the last stands'. A minute later the pressure passes its own Watch line, 5 % over the baseline. Two kinds of sign now agree, so the pack-off mechanism goes to Warning, and holds there from 02:28. At 02:33, with the pressure 209 psi over and the torque 1.7 kft·lbf over, the driller picks up and back-reams the stand. Off bottom the pressure stays 100–120 psi over its baseline while the back-ream works the cavings up, and falls away as bottoms-up clears them.

Two details make this work. The limits scale with the pressure, the larger of 100 psi and 5 % for Watch, so a well running at 4,000 psi gets a 200 psi Watch line, clear of its normal scatter. The baseline is taken at the same flow rate: a 10 % change in rate moves the pressure about 19 %, more than a pack-off does.

Differential sticking and static time

At 14.8 bar the flow check after the trip was safe, and the engine said so with a short Watch at half the allowed time. Many carbonate reservoirs in North Oman have produced for decades, though, and some sit well below their original pressure. We ran the same replay with the reservoir depleted to 0.85 SG. The overbalance becomes 39.7 bar and the allowed static time 5 minutes. The same flow check goes to Watch after 2.5 minutes still and to Warning at 5, and the first ordinary connection of the window raises a Watch at 3 minutes in the slips.

In a depleted reservoir, plan the still time before the well needs it: rotate slowly during flow checks and surveys where the procedure allows, keep connections short, and do not leave the BHA against the reservoir while waiting.

When the pipe does stick

The first minutes decide most outcomes, and the first decision is the mechanism. The worksheet asks four questions: how the pipe was moving, and whether it now goes down, turns and circulates. DrillIQ scores the answers for each mechanism.

The responses differ. For a pack-off, stage the pumps down to 200–400 psi, work torque and slack off, jar down, and never pull harder. For differential sticking, circulate at the maximum rate, slack off and jar down with torque inside the limit, then spot a freeing pill if it has not come free within the first hour. For geometry, jar opposite to the direction of travel when it stuck, starting light.

The torque limit while working the pipe is the lower of 80 % of the drill pipe's torsional yield (API RP 7G: 41,918 ft·lbf for 4″ 14 ppf S-135, so 33,534 ft·lbf) and the tool joints' make-up torque, 16,000 ft·lbf for these NC40 joints. The joints govern.

The free point tells you where the string is held. With the free-point constant A_sE / 12,000, which is 2,500 × 3.805 in² = 9,512 for this pipe, a stretch of 26 in under a 40 klbf pull puts the free point at 9,512 × 26 / 40 = 6,183 ft, or 1,885 m. That is the answer for a straight hole. Here 27.6 % of the pull is lost to friction in the builds above the lateral, so the engine's drag-corrected answer is 2,093 m: the drill pipe is free to just above the HWDP, and the BHA is held. A back-off planned on the plain figure would be 208 m too high.

A checklist for the rig team

• Read each sign in its rig state: pressure with the pumps on, drag when moving, still time when static.
• Plot every connection's pick-up against the model and watch the friction per stand as well as the hook load.
• Treat drag rising for two stands in a shale streak as the first sign of cavings, before any pressure change.
• Compare standpipe pressure only with a baseline at the same flow rate.
• Know the allowed static time for today's overbalance, and recompute it in depleted zones.
• Before tripping through a hole that alarmed, circulate it clean and plan to pump out or back-ream the interval.
• If it sticks, answer the four worksheet questions first, then follow that mechanism's sequence and torque limit.
• Use the drag-corrected free point in deviated hole.

How DrillIQ handles it

DrillIQ's Stuck-Pipe Risk module runs the four mechanism checks every 15 seconds on live rig data, with the rig state from KPI & Rig State and Torque & Drag's published loads when the well has them, or its own soft-string model calibrated on the connections. Each alarm carries its evidence and actions, escalates from the driller to the toolpusher and company man at 10 minutes and the office at 20, and is acknowledged and labelled by the crew. The Hole map shows every tight spot by depth, and Trip plan carries them into the next trip.

When the pipe sticks, the worksheet, the mechanism's sequence, the limits and the freeing calculations sit on one page, built from the well's own string, mud and survey. Replay runs the live engine over stored history; the event above was worked out with it.

Book a demo → See all capabilities

References

1. Helmick, W.E. and Longley, A.J. Pressure-Differential Sticking of Drill Pipe and How It Can Be Avoided or Relieved. API Drilling and Production Practice, 1957.
2. Outmans, H.D. Mechanics of Differential Pressure Sticking of Drill Collars. Transactions of AIME, Vol. 213, 1958.
3. Johancsik, C.A., Friesen, D.B. and Dawson, R. Torque and Drag in Directional Wells: Prediction and Measurement. SPE 11380, 1984.
4. API RP 7G. Recommended Practice for Drill Stem Design and Operating Limits. American Petroleum Institute.
5. Bourgoyne, A.T., Millheim, K.K., Chenevert, M.E. and Young, F.S. Applied Drilling Engineering. SPE Textbook Series, Vol. 2, 1986.

About this example. The well in this article is a representative scenario set in typical North Oman conditions and worked through with DrillIQ's engines. It is not a client's well, and its figures are not field results.