Ontario Power Generation is a government-owned utility company located in the Province of Ontario, Canada. Like most companies, Ontario Power is constantly seeking ways to improve equipment reliability. To this end, Ontario Power started a program to implement a common predictive maintenance program across its three nuclear generating stations.
The strategy selected to evaluate the effectiveness of various proposed improvements was to have each station run a pilot program in one of the PdM technologies; specifically oil analysis, vibration analysis and thermography. Pickering’s pilot project was oil analysis.
As part of the overall proactive strategy, the Pickering plant looked at ways to screen oil for initial alarm conditions. Historically, water has been a main concern at Pickering.
One of the main focuses was to quickly and simply identify high levels of water to allow corrective action to be taken immediately, without waiting for data to be compiled and returned by an outside lab. The hot-plate crackle test was initially considered, but ruled out due to environmental and fire concerns.
The Pickering Station implemented a relatively new technology to measure the percent saturation of the oil as a quantitative screening method for measuring water in oil. Various lubricant and component types were tested throughout the station using a commercial saturation meter with the results compared to the accepted benchmark for water contamination, specifically Karl Fischer moisture.
In oil, water is present in a free, emulsified or dissolved state. When oil is fully saturated with dissolved water, the formation of free and emulsified water begins, with subsequent damage generally associated with water in oil.
Most laboratories provide a water content measurement in ppm or percentage by volume using the Karl Fischer moisture test, however, this test cannot distinguish between dissolved, free and emulsified water. Most users set alarm limits using Karl Fischer and take action when these values are exceeded.
However, because most harm is caused by free and emulsified water, Pickering’s approach was to monitor the percent saturation to ensure any moisture present was in a relatively harmless dissolved state.
Pickering’s initial strategy was to attempt to establish a correlation between its existing knowledge of water measurements using Karl Fischer and the setting of alarm limits based on the measurements by the water sensor.
However, the results in ppm as measured by the Karl Fischer test, were found to vary considerably between equipment types and between oils. This can be explained by the varying amounts of water being absorbed due to the difference in the oil’s formulation, additive package and age. In general, more highly additized oils, as well as older oils tend to hold more dissolved water.
Because dissolved water levels in oil are related to relative humidity of the ambient air, the degree of saturation for equipment working in the same environment with the same oil should be the same, provided there is no serious source of water ingression.
For this reason, it was considered that measuring the relative humidity (RH) of the oil and setting limits based on the percent saturation was a more effective strategy than simply measuring the total ppm of water present.
When establishing alarm levels, a “machine normal” condition was initially established. Because changes in percent saturation levels due to changes in seasonal and/or relative air humidity were expected, fine-tuning the alarm limits was required. A rapid or sudden change, as well a high percent RH level was used to trigger the appropriate alarm.
The water sensor is used by inserting the probe into the oil; it is necessary to give the probe an initial vigorous stir and allow the reading to settle out. This ensures that the sensor has a chance to absorb any moisture present in the oil. This usually occurs within five minutes.
When an oil has a low percent RH level this may take up to 20 or 30 minutes. When an oil is heavily saturated with water it takes considerably less time. After each test, it is necessary to clean the probe with a solvent to remove any residual traces of the oil to avoid corrupting the next test.
As with viscosity, RH testing of the oil should be done at a constant temperature to establish a consistent trend. This is particularly important because percent saturation is strongly temperature dependent; the higher the temperature, the lower the percent saturation.
In each case, test data was obtained at 25°C ±1°C. With these few basic steps and precautions, the saturation meter was found to be an effective tool to monitor moisture content at the Pickering Station, as illustrated in the following case studies.
Data was collected from three steam turbines using Teresso 46 R&O oil in an attempt to correlate the percent saturation with Karl Fischer moisture readings(Table 1). Although minor differences were found from sample to sample, the data shows a direct correlation between Karl Fischer moisture and the water sensor data, indicating that both methods were successful in determining moisture levels in these systems.
Free water was found at 82 percent saturation (130 ppm), with the presence of small water droplets forming as the oil cooled to room temperature. Once the saturation point was discovered, appropriate alarm levels were established.
These pumps are located within the reactor building (containment structure) known as the boiler room. The relative humidity in the boiler room is controlled using dryers, thus ensuring no extraneous or seasonal impact on the data. It was anticipated that the degree of saturation should be the same for each pump. All the samples were collected on the same day with Teresso 68 being the oil tested.
The saturation levels shown in Table 2 were found to be reasonably constant, in the range 33 to 38 percent RH, exactly as anticipated. However, the Karl Fischer values ranged from 48 ppm to 206 ppm with no visual water present in any form.
It was also noted that the oil with higher Acid Numbers (AN) had higher water concentrations as measured by Karl Fischer moisture. In particular, when plotting the correlation between AN, Karl Fischer water ppm a definite trend was established. Since the saturation readings showed that the relative humidity (RH) was constant, another explanation was required to explain the inconsistencies between the Karl Fischer data and the saturation data.
The key to understanding the data is in the acid numbers. Because acid number measures the degree of oxidation of the oil, it is obvious that the samples with higher water levels by Karl Fischer are also those with more oxidized oil.
It is well-known that the age of an oil affects the amount of dissolved water that can be absorbed; the higher Karl readings simply reflect the higher degree of dissolved water in the more highly oxidized samples. Once this was understood, alarm levels for percent saturation (as well as acid number) could be established.
The data from the heat transport pump motors was the clearest validation for using relative humidity rather than Karl Fischer to detect moisture content. Pickering’s goal was to keep free and emulsified water out of the equipment.
Percent saturation allows the company to monitor and control moisture levels below the saturation point, without the interference of oxidized oil effects introduced through Karl Fischer testing.
Using readings in ppm to ascertain a harmful level of water contamination can be difficult due to differences in fluid properties, including oil age, fluid type and additive levels. Percent RH can establish “machine normal” condition and effectively monitor oil contamination levels in a proactive manner to ensure that a beneficial state is maintained.
Percent saturation is an effective quantitative screening tool and has proven to be useful when testing new and used lubricants. The fact that very little training is required translates to repeatable results, which are easily obtained.
An added benefit is that because fewer chemicals are used for testing than Karl Fischer and the sensor is relatively inexpensive, the cost per test performed is significantly less expensive than the Karl Fischer test.
In addition, the saturation instrument can give reliable warnings of free water contamination allowing early rectification of the problem, while percent RH is easier to use for establishing alarm levels than using an alarm limit based on ppm, which will vary due to the amount of water that the oil can hold in the dissolved state as it ages.