Turbine oils are subjected to a wide range of conditions – extreme heat, entrained air, moisture, contamination by dirt and debris, inadvertent mixing with different oil, etc. – that degrade the integrity of the hydrocarbon base stock and deplete the additive chemistries, causing irreversible molecular changes. There are two primary degradation mechanisms in turbine applications – oxidation and thermal degradation.
Oxidation is a chemical process where the oxygen reacts with the oil molecules to form a number of different chemical products, such as carboxylic acids. The rate at which this occurs depends on a number of factors. Temperature is perhaps the most critical one since the rate of oxidation doubles for every rise of 10 degrees C. The temperature above which this occurs is influenced by the oxidation stability of the oil and the presence of catalysts and pro-oxidant conditions such as water, air, certain metals, fluid agitation, and pressure.
Thermal degradation is the breakdown of the oil molecules by heat (high temperature), forming insoluble compounds that frequently are referred to as soft contaminants. Typically, thermal degradation occurs as a result of micro-dieseling, electrostatic spark discharge, and hot spots. Micro-dieseling is the combustion of imploding air bubbles creating adiabatic compressive heat (often exceeding temperatures above 1,000 degrees C). Electrostatic spark discharge results from the internal molecular friction that generates high-voltage electric charges such as where oil passes through very tight clearances at high flow rates, producing temperatures over 10,000 degrees C.
Over time, it has become clear that the oxidation performances of the different base stock classes are quite different. The high natural oxidative resistance of Group II turbine oils combined with specific antioxidants employed (usually based in phenol and amine compounds) provide a non-linear behavior in terms of their molecular degradation over time. As a result, the majority of standard oil analysis tests offer little to no warning as the lubricant starts to degrade and generate system deposits. Instead of degradation occurring in a linear and predictable fashion, many of the modern turbine oils fail rapidly.
Changes in the oil’s molecular structure due to additive depletion and the development of insoluble particulates are among the first oil degradation conditions that affect equipment performance. The sequential process will be the formation of sludge and varnish, which are common occurrences in turbo-generators. Besides these oxidation and thermal degradation byproducts being the main contributors for the development of varnish and deposit problems in turbines, they interfere with other important properties in steam turbine lubricants, such as demulsibility and the detrainment of air. Therefore, it is vital that appropriate diagnostic analysis be performed to detect these conditions in critical and sensitive lubrication systems.
Degradation trend of different base stock oils
Ferrography is a technique that provides valuable information about wear evolution in machinery through analysis of a representative lubricant sample. Developed by Vernon Westcott at the U.S. Navy in the 1970s as a condition-monitoring technique, it has been applied by hundreds of worldwide users to all kinds of lubricated systems.
Analytical ferrography deposited patterns
The potential of ferrography is not only limited to predictive maintenance strategies. Its important contribution to tribology studies, by assisting in a better understanding of the wear mechanisms and of the lubricant effects on the contact surfaces, turns this versatile technology into one of the most powerful diagnostic tools to assess machine health, providing valuable information about the past, present and future condition of the machine’s lubricated components.
The test procedure is lengthy and requires the skill of a well-trained analyst. As such, there are significant costs in performing analytical ferrography not present in other oil analysis tests. However, if time is taken to fully understand what analytical ferrography uncovers, most agree that the benefits significantly outweigh the costs and elect to automatically incorporate it when abnormal wear is encountered.
In analytical ferrography, the solid debris suspended in a lubricant sample is separated and thoroughly deposited onto a glass slide while passing across a bipolar magnetic field. When the sample flow is completed, a solvent “wash” cycle removes any lubricant remaining on the substrate, resulting in a “ferrogram,” where the particles are all arranged by size and permanently attached to the slide for optical analysis using a biochromatic microscope. The particles are then examined and classified by size, shape, concentration and metallurgy. This information carried by the wear particles is valuable for the identification of the wear mode and mechanism.
Steam Turbine Monitoring
This case study is about the condition monitoring of the lubricant in a steam turbo-generator at a local cellulose industry plant. The turbine is a 26 MW Siemens G 800-2. It has been in service for 22 years, operating continuously, with a lubricating oil reservoir holding 8,500 liters of ISO VG 46 oil to lubricate and cool bearings, gears and oil shaft seals and to act as a hydraulic medium for operation of the governor and steam control valves.
Since its initial operation in 1988, this turbine worked with solvent-refined base stock oil (Group I). However, due to a manufacturer upgrade, this oil was replaced by a hydrocracked base stock (Group II) in 2002. In the meantime, about 6,000 liters of makeup fluid was added, along with a few periodic oil reservoir fill-ups, making the circulating fluid a blend of these two base stocks.
The turbo-generator was operating and performing normally, and no occurrences of anomalous functions of the lubricated components had been recorded. Nevertheless, a close monitoring of the oil condition was ensured by analyzing the turbine oil periodically.
A steam turbo-generator at a cellulose industry plant
Turbine Oil Analysis
A lubricant analysis program was applied quarterly, taking two samples from the oil reservoir and sending it to independent laboratories. The standard methods used at one of the laboratories to assess the condition of the turbine oil were:
• Kinematic viscosity at 40 degrees C (ASTM D445)
• Water by Karl Fisher (ASTM D6304)
• Insoluble particulates (ASTM D4898)
• Acid number (ASTM D664)
• Neutralization number (ASTM D974)
• Elemental spectroscopy (ASTM D5185)
• Rust (ASTM D665-A)
• Demulsibility (IP 19)
• Foam (ASTM D892)
• Flash point (ASTM D92)
• Air release (DIN 51636)
• Cleanliness code (ISO 4406)
• Linear sweep voltammetry (LSV), (ASTM D6971)
Simultaneously, at another laboratory, ferrography and Fourier transform infrared (FTIR) analysis were performed along with other techniques. These analyses allowed a complemented diagnosis not only of the condition of the oil but also of the turbine wear rate conditions.
In this case study, among all the standard test results obtained, those that showed some indications of fluid degradation were the demulsibility, air release, particle count and LSV. As can be seen in the table above, the oil viscosity and acid number are within the range over the time period. Water contamination and foam tendency are maintained low. However, the particle contamination is high for all the evaluated period, the phenolic content falls below critical in some samples and the demulsibility is also affected significantly.
The sequential events in the oil degradation produces an eventual depletion of the antioxidant additives. The aminic/phenolic antioxidant mixtures actuate as a complex system. The aminic inhibitor works to neutralize the free radicals that cause oil oxidation, but it is then regenerated by phenolic, which is a good free-radical trap. When phenolic levels fall below a critical level, the oil is in danger of rapid degradation, resulting in the formation of soft contaminants and varnish. Soft contaminants are typically less than 2 microns in size and cannot be removed through standard mechanical filtration. They are insoluble and polar in nature, and are unstable in a non-polar oil environment, such as hydrocracked base oil (Group II).
Analytical results from standard oil tests show the oil viscosity and acid number are within the range over the time period.
The high ISO Codes obtained, mainly in terms of small particles (less than 4 microns), can be related with this turbine oil degradation process. Demulsibility is also compromised by the presence of polar contaminants.
For the lubrication of turbo-generator bearings, the cleanliness level with respect to particles in the oil is of the utmost importance. Consequently, a proactive action is taken through periodical on-line oil purification (filtration during 24 hours) to achieve the system cleanliness in accordance with OEM recommendations (ISO 18/16/12). However, a swift increase of the ISO Codes is consistently verified during the operation of this turbine.
The ferrography analysis completed for the same period revealed valuable information on the oil’s solid contamination. In all ferrograms, the presence of soft contaminants that resulted from oil thermal degradation and additive depletion was observed. This information is essential to identify the reason for the persistent high ISO Codes obtained in particle counting. Although soft particulates are not harmful in terms of wear, they contribute to the generation of surface deposits, as detected through ferrography.
Figure 1 shows two photomicrographs of these particles deposited on a ferrogram as observed under white/green light and polarized illumination. The polarized light allows the identification of non-metallic particles (crystalline and amorphous materials, for instance) by the brightness of light reflected. Note the brown pattern evidenced by some of these particles.
Figure 1. These two photomicrographs show turbine oil crystalline contaminants (1,000x magnification).
The particles in the ferrogram of Figure 2 are very small in size, and due to polarity, they easily aligned along the magnetic field of the ferrograph. These particles have the tendency to form agglomerates, which when overstressed with the oil, form a large coherent structure by a molecular polymerization.
Figure 2. Particles aligned on the ferrogram to the magnetic field
The varnish build-up seems to be a consequence of this physicochemical process, as can be realized by the photomicrographs in Figure 3, obtained in different oil samples. All these kinds of particulates have polar affinities and high molecular weight and tend to be adsorbed onto dipolar metallic surfaces as a sticking matter, which in turn captures hard contaminants as they flow within the system. They are capable of shutting down a turbine or causing serious damage, which is frequently related to bearings and servo applications.
Another technique employed to monitor the oil condition was FTIR, which is used to measure organic molecular components, monitor additive depletion (antioxidants) and identify organic degradation byproducts (oxidation). The monitoring of specific antioxidant depletion in used lubricants is still considered a relatively new research area. However, some studies show that the rate of antioxidant depletion is related to lubricant degradation or affected by the antioxidant mix or base stock type used to produce the lubricant.
Used oil samples are complex mixtures of different chemicals, including compounds derived from the formulation of the base oil and its additives, and from oil degradation products and contaminants. As a result, a used oil spectrum is complex and essentially the net sum of the spectra of all the individual compounds making up the sample. In fact, because of this complexity, the used oil spectrum alone is of limited value and must be compared against the spectrum of the unused oil to be of significant analytical value.
Figure 4 shows transmittance spectral snapshots of the new and used turbine oil. The black spectrum is that of the new oil (new base stock – Group II), while the red spectrum is from the blend oil in service, which still contains a small percentage of Group I base stock oil. Nevertheless, the spectra revealed identical functional groups.
In analyzing the spectrum overlays, you can clearly see relative molecular changes in the oxidation peaks, as well as thermal degradation of the oil through the signs of nitration. Another molecular alteration is observed where the phenolic antioxidants are characterized. The type of decomposition detected in the used oil spectrum is commonly observed in the FTIR analysis of fluids where thermal breakdown took place.
Figure 3. Ferrogram photomicrographs of the turbine oil particles in different samples (1,000x magnification)
Figure 4. FTIR spectra in transmittance/wavenumber (cm-1) of new and used turbine oils
Static-generated sparks are very common incidents in the filtering systems of turbo-generators. This is a phenomenon of molecular friction occurring as oil flows through small clearances, such as the filter media. Since oil and filter media are both dielectric, this electrical energy builds until a limit is reached, and then sparks are released in the lubrication system in the direction of the ground. These electrical arcs can have an extremely high, localized temperature (about 20,000 degrees C), instantly cracking the hydrocarbon molecule.
Figure 5. Plugged filter from the turbo-generator and filter mesh with black and brown shiny residue (200x magnification)
Since spark discharges generated on filters and other locations are a key root cause of varnish, and some of the previous oil analysis results confirmed that (through additive depletion and high particle counts), one of the duplex-type filters was dismantled and analyzed through an optical microscope.
Evidence of electric discharge can be easily seen through microscopic inspection of the filter media, filter core, filter meshes and from debris carried away from the filter.
Figure 5 shows one of the plugged filters changed in a periodic maintenance action due to a plugged filter alarm, with a microscopic view of the filter mesh. As can be seen, black and brown shiny deposits (sludge and varnish) are present in high concentrations, clogging the filter mesh.
The solvent used for cleaning the filter mesh was collected and used to prepare a ferrogram where significant amounts of ferrous spherical wear particles were identified (Figures 6 and 7). One source of spherical ferrous debris is the erosion wear activated by electrical discharges. The high temperatures attained by the sparks on the steel surface thermally liquefy the steel debris, which acquires a spherical shape due to rapid cooling under the action of surface tension.
The microscopic analysis of the filter core surface showed several small, circular burned holes left by the high-temperature spark discharges on the metal surfaces.
In conclusion, turbine oils must be well-maintained to extend their service life and simultaneously provide the maximum turbine performance. However, the recent upgrade in the turbine oil formulations has caused some controversy. The older analytical techniques are no longer the predictive tools able to monitor the real condition that they once were.
Figure 6. Photomicrograph showing high concentration of ferrous spheres (1,000x magnification) to the magnetic field s
Figure 7. Photomicrographs of small burned holes on the surface filter core (200x and 1,000x magnification)
The generation and presence of soft contaminants are among the main consequences of the actual turbine oil degradation process. There are four likely reasons for this:
Unlike old-generation base oils (Group I), the type of base oil currently used (Group II) does not hold varnish precursors in suspension. These insoluble particles may form deposits.
Group I and Group II turbine oils possess significantly different oxidation properties and failure mechanisms.
The antioxidants precipitate as they are preferential oxides generating insoluble particulates.
The new generation of anti-foam additives has less effective air-release characteristics, and these small air bubbles are adiabatically compressed, causing varnish to appear.
In this case study, it was recognized that only the following techniques used to monitor the condition of the turbine oil were efficient in predicting eminent problems related to the generation of varnish and sludge:
Particle counting (ISO Code) was effective in monitoring particle contamination. This was in spite of the fact that most particle counters are not sensitive to the small size of the polar particles (less than 2 microns). The reason for their efficiency was that the particles have a tendency to form agglomerates, increasing the size of the particulates and thus allowing particle counting to detect them.
The demulsibility of the oil was a critical characteristic to evaluate since it is affected by the presence of polar particles. The alteration of this property could be a signal of extreme particle contamination.
The LSV technology and FTIR are both already recognized as important techniques to monitor the condition of modern turbine oils. They efficiently monitor the condition of the antioxidant package and the creation of soft contaminants.
Analytical ferrography was effective in the detection of soft contaminants and in the identification of their nature. In the hands of a skilled analyst, analytical ferrography is a powerful technique to identify turbine oil-related problems, providing a root cause based on the morphology and characteristics of the insoluble particles, as well as monitoring the progressive mechanism of varnish formation.