Wellhead Annulus Pressure Holds: When the Tree Looks Tight and the Annulus Does Not

Sustained casing or annulus pressure with a tree that looks tight is still a well-control record.

What you saw

Annulus pressure rebuilds after a bleed.

Hold rules

Hold further production increases. Notify the well-control owner. Hold adjacent operations that share the pad if the procedure says so.

Reconstruction

Last packer test, thermal cycle, and bleed volumes.

What not to do

Do not keep bleeding as a routine without a documented decision.

What the record has to show

The hold is only as good as the timestamps. Write the first observation, the last passing check, the equipment ID, the product or lot still in play, and who has authority to release or scrap. If those five facts are missing, the investigation starts late even when the science is obvious.

Supervisors often want a verbal all-clear. That is how a second shift inherits an undocumented hold. Put the hold in the system the same hour it is called, even if the root cause is still open. Name the owner. Name the next check. Name the lots.

People, not just instruments

The operator who saw the first signal should stay in the room long enough to show the next person where to look. Handoffs that only say see the log lose the sequence of what was tried. A short witness note beats a perfect template filled in two days later.

Training records matter when the same failure repeats on weekends. Check who last ran the check, who last calibrated, and whether the procedure in the drawer matches the controlled copy. If the procedure was revised last month and the binder was not, that gap belongs in the file.

When to escalate

Escalate when product has already left the cell, when a safety device is in the story, or when two independent checks disagree. Do not wait for a full investigation tree. A same-day technical hold with a named owner is better than a polished record opened after the lot ships.

If a customer, notified body, or FDA inspector could reasonably ask for the chart, keep the chart. Screenshot the HMI. Photograph the indicator. Those files age better than memory. Store them with the lot, not in a personal folder.

Closing the hold

Closing needs a documented check that would have caught the original failure. If you only replace a part and restart, you have a repair, not a close. State the new limit, the next due date, and which lots remain restricted.

Disposition language should name lots, not affected material. QA, production, and the laboratory should sign the same list. If a lot is split, write the split. If a lot is reworked, write the rework order before anyone touches it.

Related checks worth running the same day

Walk the utilities that feed the asset: air, steam, power, water, vacuum, or nitrogen. Many holds start as a utility excursion that never made it into the batch record. Pull the utility chart for the same window as the process chart.

Look one step upstream and one step downstream. The hold you see is often the first visible symptom of a drift that started on the previous unit. Write that walk-down even if you find nothing. Finding nothing is still a result.

How this shows up on the next audit

Auditors do not start with your conclusion. They start with the time the signal appeared and ask what product moved after that minute. If your hold clock starts at the meeting, not at the observation, the timeline already has a hole.

Keep the raw files. Keep the failed printouts. Keep the names of people who were actually there. A tidy summary written the next week will not replace those.

What the record has to show

The hold is only as good as the timestamps. Write the first observation, the last passing check, the equipment ID, the product or lot still in play, and who has authority to release or scrap. If those five facts are missing, the investigation starts late even when the science is obvious.

Supervisors often want a verbal all-clear. That is how a second shift inherits an undocumented hold. Put the hold in the system the same hour it is called, even if the root cause is still open. Name the owner. Name the next check. Name the lots.

People, not just instruments

The operator who saw the first signal should stay in the room long enough to show the next person where to look. Handoffs that only say see the log lose the sequence of what was tried. A short witness note beats a perfect template filled in two days later.

Training records matter when the same failure repeats on weekends. Check who last ran the check, who last calibrated, and whether the procedure in the drawer matches the controlled copy. If the procedure was revised last month and the binder was not, that gap belongs in the file.

When to escalate

Escalate when product has already left the cell, when a safety device is in the story, or when two independent checks disagree. Do not wait for a full investigation tree. A same-day technical hold with a named owner is better than a polished record opened after the lot ships.

If a customer, notified body, or FDA inspector could reasonably ask for the chart, keep the chart. Screenshot the HMI. Photograph the indicator. Those files age better than memory. Store them with the lot, not in a personal folder.

Closing the hold

Closing needs a documented check that would have caught the original failure. If you only replace a part and restart, you have a repair, not a close. State the new limit, the next due date, and which lots remain restricted.

Disposition language should name lots, not affected material. QA, production, and the laboratory should sign the same list. If a lot is split, write the split. If a lot is reworked, write the rework order before anyone touches it.

Related checks worth running the same day

Walk the utilities that feed the asset: air, steam, power, water, vacuum, or nitrogen. Many holds start as a utility excursion that never made it into the batch record. Pull the utility chart for the same window as the process chart.

Look one step upstream and one step downstream. The hold you see is often the first visible symptom of a drift that started on the previous unit. Write that walk-down even if you find nothing. Finding nothing is still a result.

How this shows up on the next audit

Auditors do not start with your conclusion. They start with the time the signal appeared and ask what product moved after that minute. If your hold clock starts at the meeting, not at the observation, the timeline already has a hole.

Keep the raw files. Keep the failed printouts. Keep the names of people who were actually there. A tidy summary written the next week will not replace those.

What the record has to show

The hold is only as good as the timestamps. Write the first observation, the last passing check, the equipment ID, the product or lot still in play, and who has authority to release or scrap. If those five facts are missing, the investigation starts late even when the science is obvious.

Supervisors often want a verbal all-clear. That is how a second shift inherits an undocumented hold. Put the hold in the system the same hour it is called, even if the root cause is still open. Name the owner. Name the next check. Name the lots.

People, not just instruments

The operator who saw the first signal should stay in the room long enough to show the next person where to look. Handoffs that only say see the log lose the sequence of what was tried. A short witness note beats a perfect template filled in two days later.

Training records matter when the same failure repeats on weekends. Check who last ran the check, who last calibrated, and whether the procedure in the drawer matches the controlled copy. If the procedure was revised last month and the binder was not, that gap belongs in the file.

When to escalate

Escalate when product has already left the cell, when a safety device is in the story, or when two independent checks disagree. Do not wait for a full investigation tree. A same-day technical hold with a named owner is better than a polished record opened after the lot ships.

If a customer, notified body, or FDA inspector could reasonably ask for the chart, keep the chart. Screenshot the HMI. Photograph the indicator. Those files age better than memory. Store them with the lot, not in a personal folder.

Closing the hold

Closing needs a documented check that would have caught the original failure. If you only replace a part and restart, you have a repair, not a close. State the new limit, the next due date, and which lots remain restricted.

Disposition language should name lots, not affected material. QA, production, and the laboratory should sign the same list. If a lot is split, write the split. If a lot is reworked, write the rework order before anyone touches it.

Related checks worth running the same day

Walk the utilities that feed the asset: air, steam, power, water, vacuum, or nitrogen. Many holds start as a utility excursion that never made it into the batch record. Pull the utility chart for the same window as the process chart.

Look one step upstream and one step downstream. The hold you see is often the first visible symptom of a drift that started on the previous unit. Write that walk-down even if you find nothing. Finding nothing is still a result.

How this shows up on the next audit

Auditors do not start with your conclusion. They start with the time the signal appeared and ask what product moved after that minute. If your hold clock starts at the meeting, not at the observation, the timeline already has a hole.

Keep the raw files. Keep the failed printouts. Keep the names of people who were actually there. A tidy summary written the next week will not replace those.