Confined Space Entry: Falling Oxygen Inside a Steam Drum
Two boilermakers are replacing generating tubes through a 12 in. × 16 in. manway into a 42 in. ID steam drum about 20 ft long. The drum was drained, steam-blown and water-washed, and nitrogen was used to blanket it during cooldown. A 4-gas meter at the manway reads O2 20.9%, LEL 0%, CO 0 ppm, H2S 0 ppm. The entry permit is written for 8 hours. The hole watch is also acting as fire watch for the welder tacking nozzles outside the drum. After 75 minutes the worker inside reports a headache and shortness of breath; a sample taken near the far end reads 19.3% O2. A 500 cfm blower with 25 ft of 8 in. duct sits outside the manway. Ambient temperature is 92°F at high humidity.
1. Interpret the two oxygen readings. What mechanism best explains a falling O2 with zero LEL, zero CO and zero H2S, and why is the manway reading misleading?
Read the worked answer
What the numbers say
A drop from 20.9% to 19.3% oxygen with LEL, CO and H2S all at zero is not combustion, not a solvent, and not a process leak of hydrocarbon. Oxygen has been displaced by something the meter cannot see. The known candidate in this drum is the nitrogen used for blanketing.
Why the manway sample is not the drum
- The meter was sampled at the opening, where room air is entrained by the worker's own movement and by stack effect.
- Nitrogen is only slightly lighter than air, so it does not stratify neatly, but in a 42 in. ID drum with one opening it is pushed to and sits at the far end until it is purged.
- 19.5% is the entry floor. 19.3% at the far end means the breathing zone is already out of limits even though the entry point reads 20.9%.
The correct conclusion is that a single-point sample at the portal proves nothing about the far end of a 20 ft vessel.
2. The worker is symptomatic. State the immediate actions in order, and explain why the headache cannot be used as evidence that the atmosphere is safe or unsafe.
Read the worked answer
Order of actions
- Have the entrant exit immediately using his own legs if able; if not, non-entry retrieval with the harness and retrieval line. Do not send a second person in without supplied air.
- Keep the blower running and move the duct discharge to the far end so it pushes the pocket out the manway.
- Re-test at the far end and at the breathing zone with a bump-tested meter before anyone re-enters; a reading taken while ventilation is running is the reading that matters.
- Get the entrant to fresh air, evaluate, and re-issue the permit only after the atmosphere and the isolation are corrected.
On the headache
Headache is a nonspecific symptom. At 19.3% O2 the impairment is real but a headache is more likely from heat (92°F, high humidity, PPE) or from a low-level contaminant than from the oxygen number itself. It is not proof of a toxic exposure and it is not proof of safety. The meter and the isolation control the decision, not the symptom. If the meter cannot be trusted - no bump test that shift, wrong sensor - that alone stops the entry.
3. Evaluate the permit and the manning. What controls are missing, and how would you verify the ventilation is actually working?
Read the worked answer
Permit and manning failures
- The nitrogen line must be positively isolated - a closed valve is not isolation. Blind it or disconnect and tag it, and lock out the valve.
- The hole watch and fire watch are two jobs that both require continuous, undivided attention. One person cannot do both; a dedicated attendant with radio, retrieval line and a rescue plan is required.
- The permit was written for 8 hours before the atmosphere changed. After a measured oxygen drop the permit must be re-evaluated and re-issued.
- A 12 × 16 in. manway limits entry and rescue, so the rescue plan must be non-entry retrieval with a tripod or davit, rehearsed before entry.
Verifying ventilation
Drum volume is about π × (1.75 ft)² × 20 ft ≈ 192 ft³, so a 500 cfm blower gives roughly 2.5 air changes per minute on paper. On paper is not enough. Verify by:
- Positioning the duct to discharge at the far end and confirming flow at the discharge with a smoke tube or anemometer.
- Accessing a second opening, if available, to give the air a path out rather than back out past the entrant.
- Testing at the far end while the blower runs and confirming the reading holds above 19.5% at the breathing zone for a sustained period.
- Checking that the blower intake is not sitting next to a running engine or welder, which would pull CO into the drum.