As a result of the effects of over-capacity and a reduction in freight rates following the recent recession, to reduce costs ship operators have implemented “slow steaming” operations. When two-stroke engines are operated in slow steaming mode, fuel bills are significantly lower; however, the risk of developing cylinder cold corrosion increases if cylinder lubrication is ineffectively managed.
Preventing cold corrosion on cylinder liners is a key challenge for ship owners who practice slow steaming. Technical bulletins describing the challenges faced have been issued by all marine engine suppliers (Figure 1).
Figure 1. The effects of cold corrosion can be seen on this cylinder liner. The surface has uneven pock marks; these enable severe sliding wear to occur.
Furthermore, the expansion of low Sulfur Emission Control Areas (SECA) on global trade routes and future limits on sulfur in fuel are altering normal practice. There is an increased use of dual fuel handling, which requires switching between intermediate and high BN lubricants based on the fuel used. As there are different sulfur contents in different fuels, the BN offered by the lubrication oil needs to be changed based on the fuel used.
This results in the modification of cylinder lubrication oil feed rates becoming a daily routine. For this optimization, routine monitoring of residual drain oil BN is required. Current approaches are not suitable for this monitoring frequency as it can take a considerable amount of time to measure each cylinder.
A new handheld BN analyzer was designed in cooperation with Winterthur Gas & Diesel (WinGD) and Wärtsilä Services, offering ship owners a quick and precise method for determining BN for the optimization of cylinder lubrication.
This article details the challenges of conventional laboratory and field techniques for the measurement of BN and describes a new IR approach using a handheld infrared analyzer.
Background
Wärtsilä Services, Winterthur Gas & Diesel, MAN Turbo and major lubricant suppliers, emphasize that there is the need for more frequent piston underside oil analysis on their two-stroke engines and offer specific recommendations for slow steaming operating conditions (Figure 2).
Figure 2. Wartsila’s recommended limits for BN in residual cylinder oil for two- stroke engines
Further to monitoring residual BN by determining the piston underside drain oil, designers also suggest that the inlet (fresh) cylinder oil BN is measured, especially if the ship operator is performing blending on board.
Based on the technical bulletin recommendations (Fig 2), the piston underside oil BN value should be above 25mg KOH/g. Frequent direct monitoring onboard is needed to maintain the residual oil BN within these guidelines.
Understanding Cylinder Lubrication
It is important for ship owners that premature cylinder liner failure is prevented. When all of the key variables are balanced and in control as shown in Figure 3, reliable operation can be achieved.
The key variables are load rate, cylinder temperature, cylinder oil BN, lubricator feed rate and fuel sulfur content. These variables are all interdependent. For instance, if there is an increase in the sulphur content then there needs to be a change in the BN and feed rates accordingly.
Figure 3. Many variables affect cylinder health. Direct measurement of all these variables is needed to ensure proper engine health.
Using engine control system data or BN measurements (Figure 4), the ship engineer can vary the cylinder lube oil feed rate (LOFR) to compensate for changes in these variables when steaming. Most cylinder lubricating systems work on a feed forward basis because the engine control system does not receive any feedback based on piston running conditions.
It is, therefore, easy to lag behind in terms of altering lubricator spray settings when there is a change in fuel sulfur concentrations (transition from HFO to MGO). This can be a gradual change taking up to 72h.
This lag can lead to under or over lubrication. Under lubrication may result in sliding wear and cold corrosion, whereas over lubrication may cause wastage of good cylinder oil and bore polishing/hydraulic lock.
Figure 4. Wartsila Pulse Lubrication System allows for tighter control of cylinder lubrication.
Traditionally, cylinder feed rate charts that engineers depend on have fuel sulfur content as the variable across a presumed BN range based on the new oil values (Figure 5). These guides were initially developed for long journeys where the high sulfur fuel oil (HFO) composition was known and there were minimal changes. It becomes a delicate balance preventing cold corrosion as hourly variations in sulfur rates occur when ships enter into SECA zones and when BN oil is switched.
Figure 5. Lube oil feed rate chart based on the sulfur contentof the fuel being used
The ideal metric for feed rate control is the BN of the piston underside oil (Figure 6). Conventionally, it has not been used in this manner mainly due to the challenge in quickly obtaining a reliable and precise BN result. LOFR settings can be accurately adjusted up to the hour by using this information.
Figure 6. BN results answer many questions for both engine health and lubrication considerations.
Understanding BN Measurement
Base Number (BN or neutralization number) is a measure of the alkalinity reserve of a lubricant. The latest cylinder oils that are qualified for marine engines use detergents like barium, calcium, magnesium sulfonates or carbonates and calcium salicylates and phenates, which are used in formulations developed for addressing the highly corrosive environment.
The referee methods in use by global testing facilities are potentiometric titration techniques, where an acidic titrant is added to the lubricant in a tri-solvent developed to facilitate dissolution.
Conventionally, in cylinder oil BN analysis, perchloric acid (HClO4) titrant has been used in a glacial acetic acid solvent mix (ASTM D 2896) due to its capability of neutralizing both strong and weak basic additives in cylinder oils.
For four-stroke or system engine consuming low sulfur oils, another extensively used method ASTM D4739 is commonly specified. In this method, the titrant is HCl, and it is the preferred method for trending BN on engines over a specific time period, especially in the presence of wear debris.
Currently available referee methods are such that they can be carried out in a laboratory with trained technicians who understand the test and the appropriate infrastructure for handling dangerous chemicals, ventilation and glassware (Figure 7).
ASTM D2896 in particular needs further infrastructure specifications with duct work as perchloric acid fumes are highly corrosive. The appropriate titration of a high BN sample can take more than half an hour determining an end point, and hence, most labs today use a certain amount of automation in titrators. These procedures are official, but unsuitable for onboard monitoring.
Figure 7. Titration apparatus used in laboratories.
Infrared Spectrum Analysis
Infrared (IR) spectroscopy, as applied to cylinder oil analysis, is a method whereby infrared energy is transmitted via an oil sample of a known pathlength and the resultant spectrum is analyzed. Infrared energy absorption is based on the molecular vibrations present in a sample.
Interaction of these vibrational modes is done only with specific and repeatable frequencies (or wavelengths) of IR radiation and result in a clear pattern in the absorption spectrum that can be studied for quantitative chemical data about the oil sample.
The mid-IR region is mostly used (4000 to 400 wavenumbers). The alkaline reserve additives in cylinder oils and the acid build up with depletion of those reserves are seen as changes in the infrared spectrum of a sample.
For determining TBN or TAN of lubricant samples in a laboratory setup, infrared spectroscopy has been used for bringing down the time for arriving at a value and to avoid high chemical costs on routine samples. Direct probing of the lubricant "as-is" allows for trending of the change in base or acid content of the sample even if the actual value is not known.
The use of multivariate methods for correlating a change in the infrared spectrum to a titrated TAN or TBN value been possible with unique calibrations based on fluid type; however, these methods show sensitivity to type mix-ups, formulation changes and fluid contamination.
Since there are challenges while transferring the calibrations to unknown samples, these techniques were limited to the laboratory until the FluidScan Infrared Analyzer was developed in 2009.
This analyzer was the fruit of research performed for several years for satisfying an extended oil drain and fluid condition assessment requirement for a US military cost reduction and environmental improvement initiative.
The analyzer is self-contained and portable, offering instantaneous fluid condition assessment to the user. A patented flip top sampling cell is used for convenient and rapid field analysis, eliminating sample preparation and tedious cleanup.
The core of the FluidScan is a mid-infrared patented spectrometer without any moving components. The infrared light transmitted via the fluid in the flip top cell into a waveguide is collected by the spectrometer as shown in Figure 8.
Figure 8. Optical schematic for the FluidScan IR grating spectrometer. IR energy is passed through the sample in the flip top cell, concentrated and diffracted towards a tuned detector. The signal is then processed on board to provide usable data quickly.
The FluidScan increases optical throughput and spectral resolution in a portable instrument. The FluidScan, hence, offers the optimum spectral range, signal-to-noise ratio and resolution for quick analysis of in-service lubricants. This innovative technology has been optimized for low power consumption, allowing the manufacture of a robust, highly precise analyzer that works on Li-Ion batteries for up to 8h.
Developing a Cylinder BN Application on the FluidScan
Although there is a variation in the additive chemistry used in cylinder oil, there are clear spectral characteristics present in the infrared spectrum, which can be related to TBN value. For developing the application a number of new and residual cylinder oils was collected.
The testing of each sample was performed with a titrator and an IR spectrometer for developing a database with IR spectra and titration results (both ASTM D4739 and D2896) for each sample.
For studying the different additive formulations present, the spectra of new lubricants were qualitatively examined. A principal component analysis (PCA) was performed on the infrared spectra of a number of new cylinder lubricants of different additive formulations and brands. There is a high correlation seen along the first principal component with the BN value of the new lubricant as determined by ASTM D2896 (Figure 9).
Figure 9. Principal Component Analysis (PCA) of the new oil dataset shows good correlation of PC1 with the new oil BN per D2896.
This strong correlation enables the development of a universal calibration for BN for all of these oils. This can be clearly seen in the spectral region between 1000 and 1700cm- 1. The profile and peak intensity of a specific base number range is unique irrespective of the oil blend or brand. (Figure 10).
Figure 10. IR Spectra of dozens of cylinder oils show the peaks, due to additives, that closely follow each other in the 1000 to 1700 cm-1 range.
The residual oil spectra were studied over an intermediate BN range (Figure 11).
Figure 11. The IR spectra for oils with BN in the range of 12 to 38mgKOH show depletion of the alkaline reserve and accumulation of acidic by products.
There is a noted evolution of the peaks with a change in BN that corresponds to the depletion of the alkaline additive package and the build up of by products from the reaction of acids with the BN package. The transformation is usual for all cylinder oil types examined in this study, which will enable a universal calibration for BN with infrared spectroscopy.
Experimental Procedure
For observing the correlation between infrared spectroscopy and base number, a random selection of eight new oils was performed that had two oil types in every starting base number range. Titration was performed with concentrated sulphuric acid (97%) for synthetic creation of samples spanning the BN value, from new to 10mgKOH/g (Figure 12).
Figure 12. Strong sulfuric acid was added to cylinder base oils to model corrosion effects. The spectral changes were consistent with real world samples.
The samples were subjected to overnight stirring and then analysis was performed with infrared spectroscopy and titration. Using this data, a universal cylinder BN calibration was developed irrespective of the additive package or the specific oil type of the sample.
A three factor partial least squares (PLS) regression model (RMSEC = 6.2mgKOH/g) was successfully created and fine-tuned to a frequency range of 1320-1850 cm-1 over the complete BN range of 2 to 102mgKOH/g.
The calibration shows good correlation (R2 = 0.94) with D2896 base number titration results on validation with both new and used oils (Figure 13). The average absolute error of the IR result compared to D2896 titration value was about 2mgKOH/g (Table 1) on an authentic sample set of residual oil in the critical range of 10 to 25mgKOH/g.
Figure 13. The predicted onshore infrared calibration correlates well with the offshore D2896 laboratory result for BN (mgkOH/g).
Table 1. Selection of used samples
TEST SAMPLES |
PRED BN (mgKOH/g) |
LAB BN (mgKOH/g) |
w508 |
7.77 |
10.98 |
w507 |
8.99 |
12.82 |
w503 |
14.38 |
14.37 |
w512 |
12.08 |
14.46 |
w531 |
12.67 |
15.00 |
w548 |
12.48 |
15.43 |
w533 |
14.75 |
15.66 |
w547 |
13.43 |
16.82 |
w527 |
20.94 |
19.35 |
w539 |
19.48 |
20.04 |
new |
57.16 |
62.70 |
Discussion
BN measurement of the piston underside oil is currently measured offsite at oil analysis labs or increasingly onboard using a shaker test kit. Onboard shaker kits are time consuming, require bench space, involve dangerous chemicals and need the sample to be brought to the kit.
The effort is so involved that only one TBN test is performed on board every day – on the common drain. Additional cost and effort are required for observing the drain oil from each piston to calibrate the lubricating system per cylinder instead of the wastage and damage resulting from just adjusting the feed rate.
BN measurement is suitable along with new cylinder lubrication systems that have a higher accurate feed rate and cylinder spray patterns that minimise cylinder lubricant usage. Other indirect techniques like scrapedown oil analysis are targeted at the iron content which occurs due to poor control techniques.
Real time BN measurement can significantly change the onboard feed rate instructions (Table 2), and hence, the result is integral to the decision making instead of a historical observation.
Table 2.
FUEL TYPE IS... |
ONBOARD BN RESULT IS... |
FEED RATE SETTING SHOULD... |
WHY? |
HFO |
<15 |
Increase |
Under lubrication-scuffing risk |
HFO |
>25 |
Decrease |
Over lubrication hydraulic lock/chemical bore polish risk |
MGO |
>25 |
Decrease |
Over lubrication mechanical bore polish risk |
MGO |
<15 |
Increase |
Under lubrication risk |
Conclusions
The cost and time involved for analysing the shipboard Base Number cylinder oil has been considerably reduced with the new analyzer that eliminates the need for dangerous chemicals and operator interpretation of the results.
The new instrument is portable and can be carried from cylinder to cylinder. BN test results are obtained in just a minute, thus saving a considerable amount of scarce shipboard manpower. The analysis needs just a few drops of oil that drastically reduces the waste stream.
The process of measuring oil conditions is simplified by the instrument and there is no need for any interpretation by operators for precise and repeatable results. The instruments store results and offer automatic alarming, and hence, there is no need for manual logging.
The FluidScan® 1200 is a portable, robust infrared oil analyzer, which determines marine cylinder oil Base Number using a universal calibration. The sample only needs to be introduced for measurement and a BN is obtained. There is no need for a brand of oil. The calibration of the Fluidscan 1200 is performed for a 5 to 10BN range, and the results are in agreement with lab titration systems.
Further to BN testing, the analyzer has on-board software that allows ship owners to test several key oil condition parameters in synthetic and petroleum-based lubricants present in other shipboard machinery, ranging from system oil, hydraulic and gear oil to lubricants from other auxiliary equipment.
Lubrication degradation, contamination and cross-contamination are determined by the device at the point of use by determining key oil condition parameters. This, in turn, enables operators to make quick maintenance decisions.
In this technology, first the fluid is classified and identified through its infrared spectrum into its general chemical family. From this data, the instrument chooses the right set of chemometric algorithms for fluid analysis and for providing quantitative total base number (TBN), total acid number (TAN), nitration, oxidation, sulfation, additive depletion, glycol, water, and soot.
A key benefit in using an infrared oil analyzer is the ability to determine unknown new oils from a built-in fluid library. This way, onboard operators can identify oil mixups or perform incoming quality control for new lubricants in the storage area, thus alleviating the equipment problem risk because of using the wrong oil.
Base number (BN) has always been very important for monitoring proper cylinder lubrication, however, data delivery was highly time-consuming and inconvenient. The new BN analyzer renders acquisition of BN data very easily and quickly, thereby enabling ship engineers to make rapid decisions for bringing down cost in real time.
This information has been sourced, reviewed and adapted from materials provided by AMETEK Spectro Scientific.
For more information on this source, please visit AMETEK Spectro Scientific.