In-depth oil analysis is essential to the health and reliability of production machinery.
Oil analysis may be one of the last frontiers of industrial maintenance, where large amounts of money can be saved for a relatively small investment. By reducing failure-related costs, it is not beyond the realm of possibility to expect a return on investment in excess of 500 percent. Savings of more than $1 million are possible, depending on the size of the plant.
Not all lube oil analysis programs are equal. Just because a plant engages in some form of oil analysis does not mean that the machinery is well protected or that the program is effective.
It is important to note that oil analysis is only one part of a comprehensive tribology program that includes vibration monitoring and analysis, ultrasonics, and thermography. Oil analysis supplements vibration analysis by revealing two key root causes of machinery failure: changes in oil chemistry and oil contamination.
Evaluate the current situation
Here are three questions to ask:
Sampling. Collecting as few as 10 oil samples per quarter is considered adequate in some plants. This is probably not enough, however. On the other hand, some intensive programs involve gathering more than 1000 samples per month. In general, collecting fewer than fifty samples per month is an indication of an incomplete oil analysis program in most plants.
Samples should be collected and tested often enough to detect contamination and chemistry problems and establish trends. If a seal failure could allow contamination leading to damage in three months, then monthly samples will be necessary to identify a problem early enough so that steps can be taken to repair the seal.
Because every plant is unique, there is no single answer to questions regarding the number of oil samples to be gathered or collection frequency. On average, most industrial mills or plants can expect excellent cost savings based on information gained by collecting and analyzing between 50 and 200 samples each month. Here is one rule of thumb: if there are 3000 vibration points in the oil-lubricated pumps, motors, compressors, turbines, gearboxes, air handlers, and hydraulic systems in a plant, at least 100 oil points should be sampled monthly.
Testing. The going price for industrial oil analysis is about $32 per sample, but may be as low as $8 each. Is a single sample worth $32? What will it cost to repair a damaged machine? A greater investment is probably economical protection for millions of dollars worth of production equipment.
Industrial machinery is subject to contamination and chemistry-related faults leading to abnormal wear mechanisms typically involving abrasion, fatigue, adhesion, and corrosion. A thorough oil analysis for industrial applications seeks to identify lubricant components that support wear due to abrasion, adhesion, and corrosion. Typical tests include spectrometric oil analysis, total acid number, water by Karl Fischer, particle count with size distribution, and automatic and analytical wear debris analysis (WDA).
Too many industrial oil samples are subjected to low-cost analysis when they really need the particle counting, particle size distribution, and wear debris analysis that come only with a more expensive program. Some maintenance departments choose low-cost tests because they do not understand the value of looking for particles larger than 10 microns. Purchasing agents may insist on lower cost analysis. Or, the oil supplier may give away oil analysis as a value-added service.
Savings. If substantial cost savings cannot be attributed to oil analysis, serious changes to the current program should be considered because it is not producing results that are attainable. Successful oil analysis programs do pay off. A saving of $250,000 in the first year is not unusual, as an effective oil analysis program will identify potential problems that can be corrected by appropriate maintenance actions.
If no follow-through is called for, the program is a waste of time and money. It would be better to do no oil analysis than to have a program with no maintenance follow-up. At least management will not be lulled into a false sense of security by thinking that plant assets are being fully protected. Establish a strong program
The operational life of most industrial equipment is directly related to the contamination and chemistry of the lubricants, which are root causes of abrasion, adhesion, and corrosion. When a program is established to recognize presence of contaminants and to identify the types and sizes of particles present in a lubrication system, a giant step has been taken toward predicting if and when a ma-chine will fail in order to initiate corrective measures.
Is it better to do testing and analysis on site or to rely on a well-equipped off-site laboratory to test samples collected in the plant? There are a number of good arguments for doing the work on site, including better control, immediate results and immediate retest if needed, analysis by technicians that are familiar with the equipment, and the ability to test more lubricants more often.
In general, on-site oil analysis makes sense for large industrial plants with more than 100 oil systems. An effective on-site program monitors machine wear, system contamination, and oil chemistry. Emphasis must be placed on the identification of the primary root causes of abrasive wear, fatigue wear, adhesive wear, and corrosive wear.
Considering the wide range of equipment in an industrial plant and the number of faults to be monitored, an on-site program must have a range of capabilities, including both quantitative and qualitative wear debris analysis, particle counting, water contamination monitoring and oil chemistry testing.
The key to the success of an on-site program is a well-trained, in-house champion with a vision for improvement.
No matter who does the testing and analysis, successful oil analysis programs generally encompass:
Wear debris analysis (WDA). WDA measures the nature and severity of wear mechanisms quantitatively and qualitatively. An automatic wear debris analyzer or ferrous density monitor not only measures particle size, it screens out the relatively few samples requiring in-depth visual analysis. Qualitative analysis is performed by a trained technician who uses a microscope to view both ferrous and nonferrous wear debris on a glass slide or filter patch.
In many cases, this step produces the most useful information of all, including the concentration, shape, size, texture, color, and optical properties of the particles. A trained technician can determine types and causes of wear and contamination (abrasion, adhesion, fatigue or corrosion) quite accurately using this technique.
Abrasive wear particles normally are an indication of excessive dirt or other hard par-ticles that are cutting away at load-bearing surfaces. Adhesive wear particles will reveal problems with lu-bricant starvation that results from either low or high load, high temperature, slow speed, or inadequate lubricant delivery. Fatigue wear par-ticles may be associated with mechanical problems, such as improper fit, misalignment, imbalance or some other condition. Corrosive wear particles indicate the presence of corrosive fluids, such as water or process materials contacting metal surfaces.
This knowledge, which reveals the condition of a piece of equipment when the sample was taken, is useful in predicting when corrective action will be needed and what must be done.
Particle counting with size distribution. Water and dust, the most common contaminants in oil, are primary causes of abrasion, corrosion and fatigue wear. Effective oil analysis programs quantify both water and dust.
Particle counting is the accepted method for measuring total concentration of particulate debris, as well as size distribution. Both are important for monitoring the condition of the lubricant and effectiveness of the filtration system. Particle counters for on-site oil analysis should actually measure multiple size ranges leading to a determination of size distribution. Both bench-top and portable unitss are available, but bench-top use of a portable particle counter can be cumbersome.
A new ppm distribution method combines particle counting and WDA for maximum impact. Fig. 2 shows parts per million (ppm v/v) of solid particles vs size distribution for those particles. Each peak in the ppm distribution plot represents a different source of contamination or wear debris in the oil. If there are multiple peaks in the distribution, there should be a separate group of debris on a filter patch or glass slide corresponding to each peak in the plot. Each particle group can be attributed to a root cause event associated with contamination or wear events.
Water contamination. There are many ways to measure water in oil. Visual appearance, crackle test, and time-resolved dielectric are three common methods of identifying water contamination problems. The exact measure of water concentration is best left to a laboratory using Karl Fischer titration. Corrective actions depend on whether the water is in solution, emulsion, or free state. In general, emulsified and free water are most damaging.
Oil chemistry. Chemical instability in lubricants is often caused either by ingress of process materials into the fluid or by breakdown of the fluid. Breakdown occurs due to high temperature exposure and/or aeration, possibly due to foaming.
Another serious form of chemical instability is the result of water or coolant contamination. These corrosive fluids not only attack metal surfaces, they also consume vital additives that are needed for anti-oxidation, anti-wear and other functions in the fluid.
Chemistry monitoring normally involves comparison of a used oil sample with new oil. Visual examination can reveal color changes from amber to reddish-brown, indicating chemical deterioration. Quantitative on-site methods for measuring oil chemistry include dielectric, voltammetric, and TAN test kit. Dielectric increases 0.1 to 0.02 and TAN increases 1.0 to 2.0 each represent significant chemical deterioration of lubricating oil.
Staffing and training
A basic understanding of the importance of oil analysis is imperative for all maintenance personnel. The single most important ingredient in a successful oil analysis program, however, is the champion behind it. One individual must be assigned to take the lead, and that person must be passionate about the opportunity he or she has been given to save the company money.
Skill training and certification are essential. Many equipment vendors offer training and certification for their instruments and methods. In addition, general tribology training is available from various sources. The Society for Tribologists and Lubrication Engineers (STLE) provides training standards such as Certified Lubrication Specialist and Oil Monitoring Analyst. The standards are high, and the exams are not easy. Formal training is crucial. Which department should perform on-site oil analysis? The reliability team in the maintenance department is the first choice. A good alternative is the technical services department, where laboratory analysis for environment and process monitoring already take place. A third possibility is an outside contractor to collect the samples and perform oil analysis on site.
In any of these scenarios, the findings must impact equipment maintenance. Oil analysis without corresponding corrective actions will not be effective.
Periodic auditing is suggested for best practice. Each audit should compare this plant with plants identified as industry benchmark (e.g., those setting the highest standards for oil analysis practices). The audit report should include an assessment of performance and cite ways to improve.
After the initial audit, a continuous improvement plan should be drafted. The plan should set objectives for the next 24 months. Then, quarterly reviews should measure progress. Each quarterly review should be summarized in a status report to the maintenance and plant managers. All reports must include available financial evidence of savings.
Benefits outweigh costs
Many industrial plants are faced with downsizing and out-sourcing for maintenance activities. Some predictive maintenance teams that formerly comprised four to six people have been cut in half. How can these plants possibly increase the number of oil samples collected from 10 per month to more than 100 per month? Moreover, how can they begin doing the oil analysis?
The plant collecting 10 or 20 samples per month is missing problems costing far more in labor and other expenses than the cost of collecting more samples. It takes only about one week per month to collect and test 100 samples. The payoff in both labor and cost savings is far greater than the time spent doing this work.
Obviously, new programs must be justified, but history provides dozens of documented case histories with anecdotal evidence backed by the knowledge of tribology experts.