Predictive and reliability-centered maintenance programs are far more apparent today than even five years ago. At the heart of these progressive trends are technologies such as ferrography.

Ferrography, or wear-particle analysis, is the identification of all particles suspended in the lubricating fluids of any oil-wetted machinery. This technology was developed by the U.S. Navy in the 1970s. Today, it is available worldwide through commercial laboratories.

Ferrography provides a non-invasive look at historic, current and future conditions of a machine's lubricated components. This is all accomplished without the time and expense of physical examination.

Analytical methods Identifying the size, shape, composition and concentration of particles is the core of ferrography. Once a trained analyst determines these factors, an association between the wear particles and the specific component of origin can be determined. This is done through direct examination of the particles.

Glass substrate, or ferrogram analysis, is one common method of particle identification. Predict/DLI of Cleveland developed a method of particle distribution that uses a magnetic gradient field. A combination of incline, sample preparation and a magnetic field ensure all particles present in the lubricant sample are deposited on the substrate for examination. Particles ranging in size from less than 1 micron to greater than 2,000 microns are released on the substrate.

To further aid particle categorization, this method establishes consistent ferrogram patterns or maps (Figure 1, page 26). Ferrous wear particles are deposited in strings between the poles of the magnetic assembly positioned below the substrate, perpendicular to the flow of the sample. They are released in a general order of size, with the largest ferrous particles being collected at the entry end of the substrate.

Non-ferrous wear particles are released in a random manner throughout the length of the substrate, often appearing between the strings of ferrous particles. Contaminants, such as sand and dirt, fibers and friction polymers also are distributed in an irregular fashion throughout the length of the substrate.

Chemicals fix the particles to the slide and aid dispersal of the lubricant. Reduction of sample surface tension through the use of diluents, increased sample temperatures and mechanical means further aids in the release of particles from the sample.

This method of ferrographic examination provides a complete picture of the internal components of a piece of machinery. An analyst can identify all particles-from ferrous wear particles to contaminants such as insect parts-and evaluate the effect of their presence.

Cumulatively, the particles present in a sample carry with them the story, or fingerprint of the internal workings of an individual piece of equipment. Identifying these particles and the wear mechanisms that generated them can effectively demonstrate the equipment's operating history and current state of performance, as well as generate alarms to future wear conditions.

Normal rubbing wear produces platelet particles typically ranging in size up to a major dimension of 15 microns (Figure 2). They are generated by two sliding surfaces and are usually of a benign nature, unless concentrations are substantial enough to affect lubricant quality.

Bearing platelet wear (Figure 3) has a flaked appearance and is easily misidentified as normal rubbing wear. Bearing wear typically occurs in larger formations than normal rubbing wear. The morphology of these particles may be of greater significance than their size. The difference between a case-hardened bearing platelet (Figure 4) and a low-alloy steel bearing platelet (Figure 5) is indicative of the severity of bearing wear.

The concentration of bearing platelets from abnormal wear is significantly lower than the number of particles generated by other wear mechanisms, making the verification of their composition a crucial alarm in noting abnormal wear patterns.

Gear wear is a combination of rolling and sliding wear. The rolling action produces an irregular shaped particle with a generally smooth surface (Figure 6). The sliding motion produces striations very similar to those produced by severe sliding wear.

Gear wear particles are typically very large compared with other particles. Their composition may often be of greater significance than their size. The progression from high-carbon alloy steels to low-carbon alloy steels indicates wear severity.

Cutting wear is created by one surface penetrating another (Figure 7). It is perhaps the easiest particle to identify and indicates the most devastating types of wear. There are two classifications of this wear: two- and three-body.

Two-body cutting wear results when the softer of two surfaces is gouged by the harder surface, leaving relatively long, wire-like cutting wear. Three-body cutting wear occurs when the softer surface becomes imbedded by very hard particles, such as sand or dirt, and cuts the adjacent hard surface, creating short curly particles.

Severe sliding wear is identified by parallel striations on the particle surface and sharp fractured edges. These become more prominent with wear severity (Figure 8). Excessive load or speed are common root causes of these particles.

Spheres are associated with rolling bearing fatigue (Figure 9). The magnified surface of these particles appears dimpled like a golf ball. Their presence, depending on quantity and size, can indicate impending abnormal bearing wear long before any actual spalling occurs. However, it is possible that in higher-than-normal loads and in clean lubricating systems, these spheres may not be produced in significant quantities to alarm on their own.

Sand and dirt are some of the most common contaminants found in lubricating fluids and can also be the most damaging. The very fluid designed to protect component surfaces can carry with it the most devastating contaminants. These particles (Figure 10) often cause cutting wear.

Other commonly noted particles include red oxides, which are associated with water contamination. They are a form of iron oxide and can be used to identify a current moisture problem and historic difficulties.

Black oxides indicate periods of marginal lubrication. These particles are heavily oxidized in appearance, which proves an obstacle in identifying their original composition.

Corrosive debris is most often found in heavily concentrated amounts at the exit end of the ferrogram. It is generally smaller than 1 micron. These particles may be used to monitor a change in overall lubricant quality. Corrosive debris in large enough quantities can be associated with a rise in lubricant acidity.

Quantifying ferrous wear In any machine, one of the two surfaces in contact must be ferrous. Plain bearings are always in contact with a ferrous shaft, and brass bevel gears have steel worm gears as meshed contact. Although the non-ferrous surface may wear first, corresponding ferrous wear is always seen.

There are statistical databases for virtually every type of equipment in use today that can be used to cross reference the amount of ferrous wear in a particular component against that of similar, if not identical components. Through these information systems, industry averages (normal wear rates) have been established.

Various methods of quantifying ferrous wear are used by different laboratories. One method uses optical density. A powerful magnetic field causes particle deposition into a glass precipitator tube. The tube is then subjected to two channels of the optical density emission source (Figure 11).

Particles larger than 5 microns, up to a maximum of approximately 2,000 microns, are deposited at the beginning of the magnetic field, directly under the first optical emission source. Particles smaller than 5 microns are released several millimeters down the tube under the second light source.

As the particles are distributed throughout the length of the tube, the change in light density is measured and reported in two categories:

* DL-density of particles larger than 5 microns; and

* DS-density of particles smaller than 5 microns.

>From these totals, a wear particle concentration (WPC) is calculated: DL + >DS = WPC. The WPC, indicative of the rate of wear, is compared to industry >averages for the specific type of equipment being tested.

Abnormal wear usually manifests in the appearance of particles larger than 10 microns. The percentage of large particles (PLP) can be calculated:

PLP = ((DL-DS) / WPC) Yen 100

The PLP is often used to indicate the onset of abnormal wear conditions.

The whole picture Assembling all the elements of the analysis allows an evaluation of equipment condition to apply a rating of critical, marginal or normal. Often, laboratories provide graphic images of significant particles.

Physical recognition of wear patterns is the primary key to this technology, however, in many instances, the particle morphology is as, if not more, important. Various composition verification methods are used.

Subjecting a ferrogram to a controlled heat source at different temperatures for a specified length of time is a common procedure. Oxidation of the particles determines composition. Applying the heat source in stages makes the effects of oxidation readily apparent. Introduction of acids and bases can be particularly useful identifying non-ferrous particles.

Support tools such as viscosity measurements are routinely taken and can further aid the analyst in rating existing conditions. This measurement, when compared to the lubricant manufacturer's specifications, can also be an indicator of lubricant condition.

Contamination testing for the presence of water, glycol or fuel (dependent upon the equipment being tested) is also commonly performed.

Trending the rate (WPC) and severity (PLP) of wear indicates significant changes. In conjunction with the subjective analytical overview of all the particles present in a sample, an analyst is able to assess equipment condition, pinpoint current and potential problems, and make recommendations for corrective measures.

Having the ability to identify and monitor component deterioration provides the time and opportunity to cost effectively schedule necessary maintenance. Expensive and unexpected failures can be virtually eliminated.

Used-oil analysis or ferrography? Used-oil analysis (spectrographic analysis) has been around for decades and is a common maintenance tool. Spectrographic analysis generally identifies the presence of predetermined elements in a lubricant and reports this information in parts per million. Through other routine testing, it can also demonstrate the current physical properties of a lubricating oil.

A typical used-oil analysis includes:

* viscosity measurement by vibrating cylinder or oil bath;

* elemental analysis;

* FTIR spectrometer;

* total acid number;

* total base number;

* presence of water (crackle);

* quantity of water in parts per million; and

* particle count.

Combining these tests provides an invaluable tool for determining lubricant condition. They afford the opportunity to extend drain intervals and monitor the remaining serviceable life of a lubricant. In the case of a lube system containing several thousand gallons, this can be a substantial saving.

Routine testing for contaminants such as glycol, water and fuel may alarm to, and prevent a catastrophic event.Monitoring lubricant condition also can ensure that component surfaces are not subjected to insufficient lubrication protection for extended periods of time. In effect, deterring the overuse of a lubricating fluid can reduce the rate and severity of wear.

Unfortunately, oil analysis cannot detect particles indicative of impending component failure. Abnormal wear is generally indicated by particles larger than 10 microns. Knowing the composition of the normal wear particles (smaller than 10 microns) present in a lubricant is of no value when equipment condition is a concern.

Ferrography was designed to monitor equipment condition. It can be used to monitor component deterioration in order to maximize service life without the risk and associated costs of secondary damage. It may be implemented to monitor a critical piece of manufacturing or processing equipment to prevent downtime and subsequent loss of production.

Wear-particle analysis can warn of potential failures well in advance of physical manifestation, allowing timely, cost-effective scheduling of needed repairs. In addition, ferrography may eliminate the need for routine overhauls or component replacement, reducing parts inventory and maximizing repair personnel's productive time.

Ferrography's greatest liability is also its greatest attribute. It relies on a person, not a machine, to examine wear particles and interpret information. Industry averages have been established to suggest acceptable wear rates, but the crucial determination of the severity and implications of the wear is left largely to the subjective interpretation of the analyst.

Sample preparation is often time consuming, resulting in a comparably higher cost than used-oil analysis. Add to this the cost of a thoroughly trained analyst, and the total price may be three to five times the per-sample price of spectrographic analysis.

Both technologies possess unique attributes and capabilities. The best maintenance tool depends on the type of information required.

There are five major wear mechanisms:

* abrasion;

* fatigue;

* corrosion;

* adhesion; and

* lubricant breakdown.