Analytical Ferrography - Make It Work For You

Analytical ferrography is among the most powerful diagnostic tools in oil analysis today. When implemented correctly it provides a tremendous return on your oil analysis dollars. Yet, it is frequently excluded from oil analysis programs because of its comparatively high price and a general misunderstanding of its value.

The test procedure is lengthy and requires the skill of a trained analyst. As such, there are significant costs in performing analytical ferrography not present in other oil analysis tests. But, 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.

Figure 1. Ferrogram Slide Maker
Separates Particles from the Oil

Principle

To perform analytical ferrography the solid debris suspended in a lubricant is separated and systematically deposited onto a glass slide. The slide is examined under a microscope to distinguish particle size, concentration, composition, morphology and surface condition of the ferrous and non-ferrous wear particles.

This detailed examination, in effect, uncovers the mystery behind an abnormal wear condition by pinpointing component wear, how it was generated and often, the root cause.

Ferrogram

Analytical ferrography begins with the magnetic separation of machine wear debris from the lubricating oil in which it is suspended using a ferrogram slide maker (Figure 1). The lubricating oil sample is diluted for improved particle precipitation and adhesion. The diluted sample flows down a specially designed glass slide called a ferrogram. The ferrogram rests on a magnetic cylinder, which attracts ferrous particles out of the oil (Figure 2).

Due to the magnetic fluid, the ferrous particles align themselves in chains along the length of the slide with the largest particles being deposited at the entry point. Nonferrous particles and contaminants, unaffected by the magnetic field, travel downstream and are randomly deposited across the length of the slide. The deposited ferrous particles serve as a dyke in the removal of nonferrous particles. The absence of ferrous particles substantially reduces the effectiveness with which nonferrous particles are removed.

After the particles are deposited on the ferrogram, a wash is used to remove any remaining lubricant. The wash quickly evaporates and the particles are permanently attached to the slide. The ferrogram is now ready for optical examination using a bichromatic microscope.

Figure 2. Ferrogram Slide Maker
Separates Particles from the Oil

Particle Identification

The ferrogram is examined under a polarized bichromatic microscope equipped with a digital camera. The microscope uses both reflected (top) and transmitted (bottom) light to distinguish the size, shape, composition and surface condition of ferrous and nonferrous particles (Figure 4). The particles are classified to determine the type of wear and its source.

Particle composition is first broken down to six categories: white nonferrous, copper, babbitt, contaminants, fibers and ferrous wear. In order to aid the identification of composition, the analyst will heat treat the slide for two minutes at 600ºF.

• White nonferrous particles, often aluminum or chromium, appear as bright white particles both before and after heat treatment of the slide. They are deposited randomly across the slide surface with larger particles getting collected against the chains of ferrous particles. The chains of ferrous particles typically act as a filter, collecting contaminants, copper particles and babbitt.

• Copper particles usually appear as bright yellow particles both before and after heat treatment but the surface may change to verdigris after heat treatment. These also will be randomly deposited across the slide surface with larger particles resting at the entry point of the slide and gradually getting smaller towards the exit point of the slide.

• Babbitt particles consisting of tin and lead, babbitt particles appear gray, sometimes with speckling before the heat treatment. After heat treatment of the slide, these particles still appear mostly gray, but with spots of blue and red on the mottled surface of the object. Also, after heat treatment these particles tend to decrease in size. Again, these nonferrous particles appear randomly on the slide, not in chains with ferrous particles.

• Contaminants are usually dirt (silica), and other particulates which do not change in appearance after heat treatment. They can appear as white crystals and are easily identified by the transmitted light source, that is, they are somewhat transparent. Contaminants appear randomly on the slide and are commonly dyked by the chains of ferrous particles.

• Fibers, typically from filters or outside contamination, are long strings that allow the transmitted light to shine through. They can appear in a variety of colors and usually do not change in appearance after heat treatment. Sometimes these particles can act as a filter, collecting other particles. They can appear anywhere on the ferrogram, however they tend to be washed towards the exit end.

Figure 3. The Metal Alloy of the Particles
Determines Whether They Line up
On or Adjacent to the Magnetic Field

Ferrous particles can be broken down to five different categories, high alloy, low alloy, dark metallic oxides, cast iron and red oxides. Large ferrous particles will be deposited on the entry end of the slide and often clump on top of the other. Ferrous particles are identified using the reflected light source on the microscope. Transmitted light will be totally blocked by the particle.

• High Alloy Steel - particles are found in chains on the slide and appear gray-white before and after heat treatment. The distinguishing factor in the identification between high alloy and white nonferrous is position on the slide. If it is white and appears in a chain, it’s deemed to be high alloy. Otherwise, it’s considered white nonferrous The frequency of high alloy on ferrograms is rare.

• Low Alloy Steel - particles are also found in chains and appear gray-white before heat treatment but then change color after heat treatment. After heat treatment they usually appear as blue particles but can also be pink or red.

• Dark Metallic Oxides - deposit in chains and appear dark gray to black both before and after heat treatment. The degree of darkness is indicative of the amount of oxidation.

• Cast Iron - particles appear gray before heat treatment and a straw yellow after the heat treatment. They are incorporated in chains amongst the other ferrous particles.

• Red Oxides (Rust) - polarized light readily identifies red oxides. Sometimes they can be found in chains with the other ferrous particles and sometimes they are randomly deposited on the slide surface. A large amount of small red oxides on the exit end of the slide is generally considered to be a sign of corrosive wear. It usually appears to the analyst as a “beach” of red sand.

After classifying the composition of particles the analyst then rates the size of the particles using a micrometer scale on the microscope. Particles with a size of 30 microns or greater are given the rating of “severe” or “abnormal.” Severe wear is a definite sign of abnormal running conditions with the equipment being studied.

Figure 4. Red and Green Filters Help
the Analyst Differentiate Wear Particles
from Organic and Translucent Materials

Often, the shape of a particle is another important clue to the origin of the wear particles. Is the particle laminar or rough? Laminar particles are signs of smashing or rolling found in bearings or areas with high pressure or lateral contact. Does the particle have striations on the surface? Striations are a sign of sliding wear. Perhaps generated in an area where scraping of metal surfaces occurs.

Does the particle have a curved shape, similar to drill shavings? This would be categorized as cutting wear. Cutting wear can be caused by abrasive contaminants found in the machine. Is the particle spherical in shape? To the analyst, these appear as dark balls with a white center. Spheres are generated in bearing fatigue cracks. An increase in quantity is indicative of spalling.

Conclusion

Analyzing the size, shape, color, magnetism light effects and surface detail of wear particles, a skilled analyst can paint a picture above the nature, severity and root cause of abnormal wear. This information enables maintenance to implement effective corrective action.

Case Study I

Equipment: Ingersoll Rand single-stage centrifugal pump.

Application: A steel mill uses this pump to boost the water pressure it supplies to the hot mill for descaling. Thin scale forms rapidly on the hot slab and must be removed before entering the mill stands. The high pressure descaling water breaks it loose from the slab and removes it. The pump has babbitt sleeve bearings with a Kingsbury thrust bearing on the outboard bearings. This is a 12,000 gpm, single inlet pump driven by a 2000 hp motor at 1780 rpm.

History: Oil samples from this pump are routinely sent for analysis and historically have exhibited normal machine and lubricant conditions. Spectroscopic analysis and direct reading ferrography are routinely performed in order to trend the wear of the lubricated bearings. Analytical ferrography is automatically performed when there is a machine abnormality.

Problem Sample: The oil sample drawn on 7/3/00 exhibited excess iron wear, abnormal silicon and excessive total ferrous wear. The machine was given a critical condition and analytical ferrography was automatically initiated to pinpoint the source of the problem.

100X	500X
Analytical Ferrography

Analytical ferrography showed high amounts of dark metallic oxides and low alloy. This correlated with the data from the spectrometer and the direct read ferrography.

Conclusions: The maintenance department issued a work ticket to change the oil and inspect the pump bearings. Upon inspection it was found that the thrust bearing was severely damaged. The thrust shoes were deeply scored and the thrust assembly was badly damaged. The pump was pulled from operation.

The routine oil analysis uncovered a problem that was easy to repair. The inspection occurred on a planned basis, thus preventing possible downtime. If the pump were left in service, it would have catastrophically failed. This pump costs more than $50,000!

Case Study II

Equipment: Westinghouse 5,000 HP Motor.

Application: A steel mill uses this motor to drive a double reduction gearbox on a roughing mill at an input speed of 450 rpm. The motor has babbitt sleeve bearings and is coupled to the gearbox with a #11 Kop-Flex gear coupling.

History: Oil samples from this motor are routinely sent for analysis and historically have exhibited normal machine and lubricant conditions. Spectroscopic analysis and direct reading ferrography are routinely performed in order to trend the wear of the lubricated bearings. Analytical ferrography is automatically performed when there is a machine abnormality.

Problem Sample: The sample drawn on 5/18/00 exhibited excess lead wear, excess tin wear and a high total ferrous wear count. The customer was called and submitted another sample. The sample drawn on 5/24/00 exhibited the same abnormalities and the machine was given a marginal condition and analytical ferrography was automatically initiated to pinpoint the source of the problem.

Click Here to View Spectrometer and Direct Read Ferrography Data

500X	500X
Analytical Ferrography

Analytical ferrography showed high amounts of low alloy steel and dark metallic oxides. The colored particles on the top slide are low alloy steel that changed color after heat treatment.

Conclusions: The results of the oil analysis initiated a visual inspection, which revealed that the coupling appeared to be locked. The maintenance department issued a work order to inspect the motor bearing and coupling. Upon removal of the bearing cap, it was found that the faces of the bearing were wiped. The bearing surface that the shaft rode on was also starting to wipe. The bearing was changed.

Replaced Motor Bearing

Maintenance completed the repairs on a planned basis with no delay to the mill. If left unrepaired, the motor bearing would have failed catastrophically. Such a failure while the mill was running would have caused a four-hour delay to the mill with a cost in excess of $60,000!