This article investigates an air breather installed on the oil reservoir for lubricating a gearbox and discusses its capacity to capture the oil fog generating from the reservoir. The author analyzed the root causes of oil fog generation, confirmed its cause to be a repair error from drilling a new hole in a labyrinth seal chamber, and then recommended how to prevent it.

A customer complained of a problem with the air breather installed (Figure 1). It could not completely capture the oil fog, although it could greatly reduce the fog escaping from the air breather. Under the air breather, the oil from the fog escaped and collected outside, pooling on the outer top surface of the reservoir within a week (Figure 2).

The field investigation confirmed this condition, and the escaping fog felt oily on the fingertips. While we saw the heavy fog billow out, we asked ourselves what could cause this heavy fog to generate in the reservoir, and which countermeasures could prevent the fog from forming?


Figure 1. Air Breather on Oil Reservoir of MVR Compressor

We studied the lubrication system, discussed the conditions that generate the oil fog, and made a fault tree diagram for the oil fog generation and its escape to the atmosphere (Figure 3). Following this diagram in our search for the causes, we investigated the field conditions of highly possible causes for the continuous outflow of heavy fog, such as reservoir design, oil lubricant condition, air breather condition, and invasion of air into the gearbox.


Figure 2. Oil Leak on Top of Reservoir

Field Investigation of Air Breather
Capacity of Air Breather for Oil Fog Capture
The filtration capacity of the air breather was 99.7 percent for capturing oil fog particles greater than two microns (that is, ß2=300) and 99.8 percent (ß3=400) for removing solid contaminant particles greater than three microns, according to the product catalog. Comparing use with this air breather and without it, we found that the fog escaping was greatly reduced as follows: The fog still escaping was considered two microns or less and can be shown not as individual particles, but as a slight smoke billowing out of the air breather.

The design flow rate of the air breather was 1,000 liters per minute (lpm) under the differential pressure of 350 mmAq. Because the delivery flow rate of the oil pump was 139 lpm and the suitable flow rate of the air breather was two times the pump delivery, the present air breather of 1,000 lpm flow rate was a sufficient capacity compared to the maximum 278 lpm required.


Figure 3. Fault Tree Analysis on Fog in Reservoir and Its Escape to Atmosphere

Improper Installation and Operation of Air Breather
The installation state was good as shown in Figure 1, except the clogging indicator was discovered popped out during the investigation. It was said that the indicator had not been pushed in when installed. The result of the experiment for 10 days after pushing it in was the same as the initial state of being pushed in. Therefore, we concluded that the operation of the air breather was still normal. This air breather was found to have no check valve for bypass of the flow.


Figure 4. Planned Oil Flow in Reservoir

The possible size of oil mists depends on the process of its generation. The mists generated by the mechanical process, such as spray of oil jet or pressured air through the Venturi tube, shall be more than three microns. They easily become large drops falling on the surface due to their weight and collision with each other.

Conversely, the mist by distillation or evaporation of oil is so small that it's invisible to the naked eye and ranges from 0.005 micron (of a single molecule) to three microns (several molecules together). Therefore, the mists that were not captured and escaped from this air breather are considered to range between 0.005 to two microns.

Cause of Oil Fog Related to Oil
Our second question is why so much oil fog was generated from the oil reservoir. It too was answered by an intensive survey of field conditions at the plant.

Table 1. Oil Temperatures Measured at Outer Surfaces of Reservoir

Oil Temperature
The survey results of the oil temperatures of the reservoir are shown in Table 1. Temperatures of the outer surface were measured by an infrared thermometer (HIOKI 3444, Japan) because of the lack of a thermometer installed on the reservoir.

Reviewing the oil temperatures shows that the baffle appeared nonexistent and, as shown in Figure 4, the oil returned from the gearbox seems to flow along two divided paths; that is, 1) from return pipe directly to suction pipe in the rear tank and 2) from return pipe through the front tank, connection duct under the motor, and rear tank to the suction pipe.


Figure 5. Oil Suction Pipe (Third from Left) on Rear Tank, Compressor Side

The flow paths are as follows: 1) return pipe (48.3°C) from the gearbox ? compressor-side rear tank (39.5°C) and suction pipe and 2) return pipe (48.3°C) from the gearbox ? compressor-side front tank (41.3°C) ? motor-side front tank (33.6°C) ? connection duct ? motor-side rear tank (29.0°C) ? compressor-side rear tank (39.5°C) and suction pipe. The flow path directly from the return pipe to the suction pipe is too short to be cooled.


Figure 6. Oil Level Gauge Showing Foam Inside

Therefore, the sucked oil is hot and laden with wear particles and air bubbles not separated enough to be suitable for recirculation. The air bubbles may cause cavitation to occur on the delivery side of the pump and burn the oil at the surface of air bubbles compressed adiabatically if the pressure is high.

Because the oil temperature was about 40±10°C in the reservoir, however, it can hardly be regarded as the cause of the heavy fog generation.

Table 2. Physical Properties of Base Oil of R&O 46 Turbine Oil

Oil Jet Sprayed Within Gearbox
Another factor to discuss is the oil jet sprayed to the ring gear mesh of the planetary gear train within the gearbox. This transmits power from the electric motor to the compressor impeller, increasing the speed from 1,800 to 11,600 rpm. The dissolved gas in the oil jet may expand adiabatically into the free gas phase within the gearbox and generate heavy fog when it collides with the gears meshed.

This mist fog may return to the tank, but fortunately, the return pipe was submerged deeply under the oil level to prevent the mists from entering into the headspace of the tank. Therefore, the mist generated within the gearbox cannot be regarded as the direct cause of the heavy fog coming out of the reservoir.


Figure 7. FTIR Measurements of New and Used Oil

Insufficient Oil Volume Kept in Reservoir
The oil volume kept within the reservoir was approximately 1,600 L, which is more than 10 times the delivery of the pump (139 lpm). It is sufficient only when oil flows around the long path, which utilizes the whole tank capacity by the baffle plate to divide the suction-pipe side from the return-pipe side. Such arrangement can enable the settlement of wear particles and other solid contaminants and the separation of air bubbles to be performed effectively, while flowing along the long path and cooling down as low as possible.

However, the foam covering the surface as viewed through the oil sight glass in Figure 6 shows the insufficient effect of air separation and defoaming performance, and doubts the existence of a baffle plate and its sound structure enough to divide both sides of the suction and return parts. This entrained air may cause cavitation in the pump which can burn the oil to produce carbon soot particles and nitrates. It also helps generate sludge floating on the oil surface and the varnish stuck on the cold machine surface.

Fortunately, the pressure was approximately 1.4 bar, too low for severe cavitation to occur (this usually requires more than 70 bar). Reviewing the oil analysis in terms of Fourier transform infrared (FTIR) shows no trace of nitrates and soot produced as the cavitation by-products, except the reduction of antioxidant additive at the IR wave number of 2,350/cm due to the heat generation (Figure 7).


Figure 8. Compressor and Gearbox

Degradation of Base Oil
Such a condition may imply the degradation of base oil and additives. We reviewed the oil analysis report for other properties such as viscosity and acid number. The used oil was R&O 46 turbine oil. As shown in Table 2, the kinematic viscosity was 42.6 centistokes (cSt) at 40°C, acid number 0.03 mgKOH/g, and water content 50 ppm.

The used oil was in good condition compared to new oil and its control limits. The slight reduction of acid number may be caused by the dissipation of antioxidant additives and may increase again after oxidation of base oil has progressed. But in its present state, it was still normal. Therefore, we investigated other causes.

Root Causes of Oil Fog Related to Equipment
We have investigated the lubricant state and its operation conditions without suspicious causes. Now we need to change the focus of investigation to the equipment- maintenance side.

Reviewing the history of faults for this equipment showed that the compressor's impellor was broken about eight years ago, on January 28, 2,000, when the gearbox's shaft, shaft seal and bearing were damaged. After replacing the shaft seal modified by the outsourced repair company, the generation of heavy fog was reported.

Therefore, to investigate the modified part, we reviewed the internal design such as seal structure and the pressure balance, on drawings provided by the repair company.

Compressor Design
Because the compressor's impeller lacks a balancing hole, the thrust force was exerted on the rear of the impeller by the differential pressure between suction and delivery pressures of the compressor. It was supported by the thrust bearing to the direction of compressor (Figure 8). This was lubricated by the outflow oil from the journal bearing supporting the sun gear shaft of the planetary gearbox, which drives the impeller.

The compressor's shaft seal was composed of several sheets of labyrinth seal, and the leaked gas was vented to outside of the building through a vent pipe connected to the end plate chamber. The suction pressure was 597 mmAq, delivery pressure was 600 mmAq, and the differential pressure was negligible.

Gearbox Design
Concerning the possibility of sugar gas leaking through the labyrinth seal and entering the gearbox, we investigated by unscrewing two eyebolts on top of the gearbox. We found oil mist emitting heavily from a hole on the motor side, and oil drops ejecting from a hole on the compressor side.

A second investigation of this gearbox shows the following:

  1. During the repair of the broken shaft seal, a hole of 4 mmf?was drilled to return the leaking oil in the last chamber before the last land of labyrinth seal. The original purpose of this chamber was to apply the air pressure to seal the shaft.

  2. This modification was not approved by the original equipment manufacturer (OEM). Steps to restore the equipment to its original condition as required by the OEM were not performed because of the continuous operation of the equipment.

  3. Through this hole, the air applied at the pressure of 1.5 bar to this chamber was leaking into the gearbox directly. It flows through the return pipe and the oil in the reservoir to the air breather above the tank, carrying the heavy fog from the oil.

  4. The leaked oil from the shaft seal was radially flung from the slinger ring assembled on the shaft by the centrifugal force of shaft rotating at the high speed of 11,600 rpm. If the internal pressure accumulated by the seal's air supply is stopped, the oil leaked through this shaft seal will stop due to the retrieving effect of oil leaked through the newly drilled hole.

Remedial Measures and Recommendations
Based on this investigation and reviews, the air breather performed the functions described in the product catalog. The root cause of the heavy oil fog was the air that leaked through the hole in the modification. The remedial measures to prevent the oil fog are as follows:

  1. Vent the gearbox by removing the two eyebolts above the gearbox. Install air breathers to capture oil mist generating and escaping from there.

  2. Stop the sealing air supplied to the labyrinth seal. Or as an alternative, after plugging the hole modified as required by OEM, reduce the air pressure applied to the labyrinth seal to less than 0.05 bar.

The complementary ideas to operate more efficiently are as follows:

  1. Install a baffle plate in the reservoir between the suction and return pipes. This will increase the flow path of the recirculation oil to help settle contaminant particles, cool the hot oil, and separate entrained air and gas for preventing cavitation and pump wear.

  2. The cleanliness of new oil was found to be 12 grade in NAS 1638 and ISO 21/17, not suitable for the journal bearing to require less than NAS 7 grade and ISO 16/13 to prevent wear. The new oil should be filtered through a filter element of three microns absolute, the oil cleanliness confirmed below NAS 7 grade, and the used oil in the system filtered continuously through the off-line filter of three microns absolute. This will help assure fault-free operation.

  3. Regular sampling and oil analysis works to control contamination to limits of less than ISO 16/13 or NAS 7. Through these measures, the bearings and gears will be proactively maintained to achieve fault-free operation and extended service life.

  4. The existing filter in the oil system should be tested for performance and have a filtration efficiency greater than 99.5 percent for the contaminant particles greater than six microns.


1. The operation and maintenance manual of the MVR compressor.

2. Oil analysis report for particle counts, shape classification of wear particles, RDE and RFS, and FTIR of used and new oil.

Author's Note
The root cause was identified to be sealing air supplied to the labyrinth shaft seal. As a result, the fog generation stopped after removing the air supply, and the oil leakage onto the slinger ring also stopped. The survey for root cause of oil fog generation was successfully concluded, and the problem was determined to be the sealing air that leaked into the gearbox through the newly drilled hole.

The author wishes to express special thanks to Solge’s president, Why Kim, for his sincere discussion and support, and to K.J Lee, an engineer in charge of this compressor for his support and help in field investigation.