Case Study- Shipboard Inert Gas Generator Problems
In the first part of the last century, vessels carrying fuel and other hydrocarbon cargo utilized open holds with hatch covers, which afforded some level of natural ventilation of the tankage. This practice was not particularly safe, as volatile vapors could migrate into the boiler/engine room, possibly causing an explosion and fire, and also exposed crew to potentially toxic hydrocarbon vapors. After WWII, hold design migrated to enclosed tanks, which contained hydrocarbon vapors much better, but also forced closer attention to the control of potentially explosive atmospheres.
The tank environment has to be maintained to avoid the lower and upper explosive limits where any ignition source could cause an explosion and fire.
Steam-powered vessels might divert combustion flue gases from the boiler into the holds in order to keep the O2 content of the holds well below the LEL (around 10-12% O2). Dedicated Inert Gas Generators (boilers) were developed in Norway by Wärtsilä Moss, and Foster-wheeler licensed the technology in the US. These units provided slightly pressurized, cooled, and clean flue gases solely for the purpose of inerting, ie. no heat recovery or steam generation. More recently, nitrogen generating systems have entered the market, which utilize a hollow fibre molecular sieve principal, grading air by molecular size. This is similar in principal to medical oxygen concentrators used by patients with breathing problems, except that shipboard inerting applications are much larger in scale, and the nitrogen is the desired molecule, with the resulting oxygen typically vented overboard.
Whether the inerting system utilizes flue gases or nitrogen, it’s important to measure and control the resulting oxygen content of the inerting gas to 2-5% before directing it into the holds. Additional in-hold measurement with portable analyzers is also recommended.
A fuel barge plying the great lakes was having trouble with a zirconium oxide oxygen analyzer probe used to confirm the final O2 of an inert gas generation system (IGS). They could not offload their cargo until they could guarantee that the analyzer was reading correctly, and that the oxygen content of the inert flue gases being generated were safe to replace the volume of the holds as the liquid product was pumped out. The analyzer was working, but reading 14% with the generator off, vs. normal ambient air of 20.95% O2. It was decided to remove the analyzer for inspection. Upon removing the conduit strain relief for the electrical cable about a pint of oily water drained out of the probe. The most likely source of this contamination was the plant air being utilized as reference air. The oil-lubricated compressor had some leaks, and the drier had been bypassed in the distant past. It should be noted that while contaminated plant air can have many negative consequences on valves and other actuating devices, the effect on a ZrO2 analyzer is particularly bad, since the analyzer depends on the reference side of the sensing cell (the entire inside of the probe) to be exactly ambient- 20.95% O2.
The probe was completely disassembled and cleaned with solvent. The metal parts were then heated with a blow torch to burn off any solvent/oil residue left, and the probe was reassembled with a new heater assembly, and powered up. Reference air was supplied from a simple aquarium pump in order to expedite the burn-out of hydrocarbons. While clean and dry ambient reference air is preferred for normal operation, the reference side of the probe can also just be left open to atmosphere, which will diffuse into the probe.
Smoke from the oil residue poured out of the probe as it heated up for about an hour, after which it was calibrated and successfully cal-checked by flowing calibration gas to verify that the readings were correct.
Gas Generator Problems
IGS boiler would start up, but it kept shutting down after 10-15 minutes. It was postulated that the controls were tuned to operate with the O2 analyzer in poor condition, and the controls would not run with the repaired and calibrated probe. A portable analyzer was reading higher O2 than the installed O2 probe, and also reading high CO at the maximum scale of 2000 PPM.
It was suggested that the burner nozzle and seals may not be atomizing the diesel fuel properly, causing poor combustion at the burner. The Teflon seals of the burner nozzle were replaced, and after several tries, the IG unit came up well in manual, with the O2 readings holding steady at around 2%. Further seal work was conducted, and the Inert gas system was deemed satisfactory for operation.
Summary and Recommendations
- Inert Gas Generators maintain a safe condition inside tankage containing any hydrocarbon liquid. Proper operation of the gas generator is required for most operations, esp. off loading.
- The measurement of Oxygen confirms that the inert flue gas has a safe low level of oxygen, and also ensures efficient and clean operation of the inert gas generator.
- Proper calibration and maintenance of the O2 analyzer is important. A spare O2 probe should be kept in inventory.
- Reference air must be maintained at exactly ambient levels (20.95% O2), either by flowing clean and dry instrument air to the inside of the probe, or by leaving the reference air port open to atmosphere.
CO and other hydrocarbon gases cause lower readings of the ZrO2 probe- Be aware that sensing cells of zirconium oxide analyzers are heated to 700 +C, and uses platinum electrodes. Any CO or other combustibles that might result from burner problems will burn on the sensing cell, which will consume some O2, causing the probe to read lower. This phenomenon is by design to ensure a lower O2 reading when a given burner has problems.
The heated ZrO2 sensor should also never be utilized directly in the holds being protected. The hydrocarbon vapors present will cause lower O2 readings, causing a false sense of security. The heated cell can also become a source of ignition if the gas mixtures in the hold reach explosive limits.
- Inert gas generators operate at a positive pressure to ensure that tankage is pressurized with the inert gas. O2 probes should ideally be calibrated while the generator is operating and developing pressurized flue gas. A upward shift of app. .3% O2 readings can be expected if calibrations are conducted with the generator off-line.
- While heated zirconium oxide O2 probes do a good job of measuring the final oxygen at or near the point of inert gas generation, they should not be utilized in the actual cargo holds being protected. Since the sensor is heated to around 1450 F (700C), the se
- Some control strategies for Inert Gas Generators are designed to be fail-safe, and do not permit operation in manual control. Safe operation in manual control may be possible with a good O2 reading, and trained and attentive operator.