You’ve got an A-annulus building pressure, or production rates are dropping, and the data points to a tubing leak. The instruction from the asset team is clear: “Find the leak.” So you call out a logging unit, run a temperature survey, and come up inconclusive. Then you try a noise log, still no definitive answer. Two runs, significant NPT, and you’re no closer to a repair.
This scenario is frustratingly common. The problem often isn’t the tools themselves, but the approach. “Find the tubing leak” isn’t a single question; it’s three distinct questions that require different diagnostic methods. Does a leak exist? Roughly where is it? Exactly which joint or connection is it? Starting with a high-resolution survey to answer the first question is a wasted run. Stopping after answering the second and attempting a repair often means you’re fixing the wrong depth.
The Engineering Reality: Why Tools Seem Definitive But Aren’t
Each leak detection technique has a clear physical basis: temperature responds to fluid movement, noise to turbulence, and pressure to communication. All these principles are true, yet all three are routinely defeated by well conditions that a superficial interpretation doesn’t account for. The key is understanding each method’s strengths, weaknesses, and the specific conditions under which it provides reliable data.
The goal is to build a diagnostic sequence that minimizes NPT and maximizes the chance of a successful, single-run identification. This starts long before any logging tool is mobilized.
Operational Approach: Matching the Tool to the Problem
Pressure Methods: The First, Free Step
The starting point for any leak investigation is always pressure behavior, because it costs nothing but time. Monitoring the relationship between tubing pressure and annulus pressure over time helps you discriminate between communication, thermal effects, and external sources. If your annulus pressure tracks tubing pressure with a short lag, you have communication with the tubing. If annulus pressure rebuilds after bleed-down to the same value, at a rate independent of tubing pressure, you’re likely dealing with an external source or trapped annular pressure, not a tubing leak. If the annulus bleeds to zero and stays there, you don’t have a tubing leak.
Before you even think about mobilizing a logging unit, run a diagnostic bleed and rebuild test. Bleed the annulus to zero, carefully recording the volume removed. Then, monitor the pressure rebuild. The volume removed helps distinguish a gas cap under pressure from a liquid-filled annulus with a small leak. The rebuild rate differentiates an active feed from thermal re-equilibration. The rebuild ceiling – whether it returns to the same value or a lower one – tells you if you have a fixed-pressure source or a depleting one. This data is invaluable and free.
For more specific depth information from pressure alone, you can use a packoff or straddle test. This involves isolating a section with a retrievable plug or an inflatable packer, then pressuring each side and observing. While genuinely diagnostic, it’s slow. Each step tests only one interval, making it efficient only when the leak has already been narrowed down to a few hundred meters.
Temperature Logging: Sensitive, But Tricky
Fluid moving through a small orifice changes temperature. Gas expanding through a leak cools via Joule–Thomson expansion, producing a cold anomaly. Liquid crossing a leak typically causes a warm anomaly if fluid from a hotter depth arrives at a cooler one, or a cold anomaly if flow is downward from a cooler depth.
However, practical difficulties abound. The thermal anomaly is often displaced from the actual leak in the direction of flow behind the tubing, sometimes by tens of meters. A leak small enough to impact integrity might be too small to produce a detectable thermal signature at low flowing rates. Furthermore, a static well’s temperature profile approaches geothermal, suppressing the very contrast the method relies on.
Temperature logging works best when the well can be deliberately manipulated. Inject into the annulus or tubing at a controlled rate, log during and after, and interpret the difference between flowing and shut-in passes rather than either alone. This is the crucial operational detail: never interpret a single temperature pass. The information is in the delta between passes taken under different flow conditions. A single pass in a well with unknown thermal history yields an ambiguous curve. Two passes with a controlled change resolve most of that ambiguity, and the second pass costs a fraction of the mobilization.
Noise Logging: Pinpointing the Source
Turbulent flow through a restriction radiates acoustic energy across a broad frequency range. Spectral analysis helps separate the signature of fluid crossing a small hole from background flow noise. Because the source is local, the depth resolution is generally better than temperature logging, often down to meters.
Noise logging excels at identifying leaking connections and detecting flow behind pipe in cemented annuli where temperature contrast is weak. Its primary weakness is quantification – the relationship between acoustic amplitude and leak rate is complex, depending on geometry, fluid, and differential pressure in ways that resist calibration in a real well. Treat a noise log as a locator, not a flow meter.
A critical operational requirement for noise logging is maintaining a differential pressure across the leak during the survey. A well shut in with tubing and annulus equalized will produce no flow through the leak, and therefore no signal. Establishing and holding a differential during the run is part of the job, and this differential should ideally simulate the direction of flow that would exist during normal operation.
Worked Example: Is the Leak Big Enough to Log?
Consider an A-annulus that rebuilds from 0 to 340 psi in 72 hours. The annulus is liquid-filled with a gas cap of approximately 1.8 m³ at the top. Accounting for gas compressibility, the volume of gas entering at annulus conditions is estimated at roughly 0.7 m³ over that period.
- Rate ≈ 0.7 m³ / 72 h ≈ 0.0097 m³/h ≈ 162 cm³/min at annulus conditions.
This is a small leak. At that rate, the velocity through a defect of even 1 mm² is around 2.7 m/s – sufficient to produce a detectable acoustic signature but marginal for a thermal anomaly against a geothermal background. In this case, running spectral noise first with a maintained differential is the logical choice; a temperature survey would likely be inconclusive and consume a run establishing that.
Decision Checklist: A Sequence to Avoid Wasted Runs
Follow this sequence to maximize your chances of a successful, single-run leak location:
- Establish Existence and Source: Use annulus pressure behavior, bleed volume, and rebuild profile. Do not mobilize any logging tool until this data is documented.
- Estimate Leak Rate: Use the bleed data. The estimated rate dictates which logging methods have any chance of detection.
- Check Completion Record: Review for plausible candidates – connection type, previous interventions, known corrosion exposure, gas lift valve ports, sliding sleeves. A leak at a side-pocket mandrel is far more likely than in the middle of a joint, and this knowledge focuses interpretation.
- Run Combined Survey: Deploy a combined temperature and spectral noise survey in one descent. Ensure a controlled differential is established and held. Take at least two passes with a deliberate change in condition between them.
- Interpret Together: Always interpret the two measurements in conjunction. Agreement between an acoustic source and a thermal anomaly displaced downstream of it is a strong result. Either alone is just a hypothesis.
- Confirm with Packoff Test: If the repair is expensive, confirm the identified depth with a packoff test before committing to the intervention.
Failure Modes and Lessons Learned
Most failed leak-location campaigns fail at the planning stage, not at the interpretation stage. A common mistake is running an expensive survey on a leak too small to produce a detectable signature. The result isn’t a negative finding; it’s simply no information, which then often justifies another costly, inconclusive run.
Remember that interpretation pairing provides the most reliable signatures. An acoustic source with a cold anomaly immediately above it, in a gas well, strongly indicates gas leaving the tubing and traveling up the annulus. An acoustic source with a warm anomaly below it suggests downward flow. An acoustic source with no thermal response at all could mean a very small leak rate, or a source behind the outer string rather than in the tubing itself.
Another frequent pitfall is failing to establish and maintain the necessary differential pressure during the logging run. Without flow across the leak, neither temperature nor noise tools will register a signal, regardless of the leak’s size. This requires active well control and coordination with the logging crew.
Bottom Line
Effective tubing leak location relies on a disciplined, sequential approach, starting with basic pressure diagnostics before escalating to logging tools. The bleed-down data, available for the cost of a pressure gauge and some patience, is the most critical input to determine if any logging tool can even see the leak. Define your question, understand your well conditions, and choose your tools strategically to avoid wasted runs and NPT. Have a question about your well? Reach out via the contact page.