The following article was an effort to combine theory with actual practical statistics. Many things change, but many also stay the same. What would one see as a “big splash of change” in pumps in the past 25 years?
The biggest change is the ever-evolving trend toward instrumentation, computer control, data acquisition systems and similar automation-related technologies. However, what did not change is “how pumps feel about it.”
The number one and two forerunners of failures are mechanical seals and bearings—or perhaps some would say bearings and seals. Despite all the developments in technologies, bearings and seals keep failing the same way they did 25 years ago. There have not been any really significant, ground-shaking new bearing designs or fundamentally different seal designs that have come on the market.
Bearing and seal failures have never been an exact science—when bearing housing vibrations begin to increase from 0.05 inches per second (in/sec) to 0.1 in/sec, no guru of reliability in the world can tell you that their life will be shortened in half. If vibration went up from 0.2 in/sec to 0.8 in/sec, all specialists will tell you that the failure is “imminent.” However, I have seen cases where the pump kept running for another seven years, still at 0.8 in/sec vibrations, and was eventually changed to another design for completely unrelated reasons.
So, these are the trends. Ironically, we can repeat the old phrase adapting it to pumps: “The more pumps change, the more they stay the same!”
In the future, what we need to do as a community is to begin quantifying reliability with real field data and link that to theory—combining both. The statistical approach to reliability has been for years in a rudimentary form, and we need to address it. If we do, we will begin to learn what to tell a maintenance manager when he sees vibrations go up from 0.05 in/sec to 0.1 in/sec—to predict if the bearing, based on this analysis, will fail next Tuesday or in 10 years.
Proper lubrication is a key to long, trouble-free life of centrifugal pump bearings. In recent years, the issue of lubrication has received renewed attention from pump users in chemical plants, pulp and paper mills, refineries and other industries.
Budgetary pressures have forced many plants to reduce maintenance capital. Many knowledgeable maintenance workers have been laid off.
Not surprisingly, the ability to maintain pumping equipment properly is reduced, resulting in increased outages, lost production and rising maintenance costs.
Users have started to look to pump manufacturers to pick up the slack and help solve pump reliability problems, extend component life, and increase mean time between failure (MTBF) and mean time between scheduled maintenance (MTBSM).
Statistics show (Ref. 1) that most pump failures are related to bearings and seals. In this article, we will look at bearings, analyzing how design changes affect bearing life in a quantifiable way.
The need for improved pump reliability and increased MTBF led to a new design. When comparing the cross sections of two single-stage, end-suction ANSI pumps, both have identical wet ends (impeller and casing), but the power end and the seal chambers are different.
Improvements in the seal chamber are significant. The new design has a larger chamber to ensure better heat transfer and cooler operation of the mechanical seals. The previous design incorporated a tight stuffing box.
The new power end design features approximately three times the volume of the oil sump, an oil level sight glass to assure the proper oil level versus the constant level oiler, improved cooling via a finned cooler insert versus bottom cooling pockets, labyrinth oilframe seals versus lip seals, and stiffer footing for reduced vibrations.
A testing program has been conducted to compare the two designs under extremely adverse operating conditions, such as running endurance testing at overspeed and below minimum flow. This program resulted in quantifiable correlations between changes in pump design and their effect on life extension. Feedback from users comparing two designs was also obtained, specifically in relation to the operating temperature of the bearing frame surfaces.
Analysis of the Power End
With regard to the power end, the belief that “the bigger the better” is not uncommon in the pumping community. This idea has some merit, but manufacturers often overlook the importance of quantifying the benefits of a particular design or modification. Frequently, little information is given as to how much life extension can be obtained by, say, having a deeper sump, or how much added value and savings can be realized from the increased bearing frame heat transfer surface.
It is clear that a systematic approach to identify, measure and improve pump component design is impossible without a proper balance of theory, experimentation, user feedback and data from real world installations. Theory and experimentation should be balanced by clear communication between manufacturers and users.
Increased Frame Outside Heat Transfer Surface
Heat is transferred from the pump bearings to the oil and through the housing frame walls to the outside air. Some of the heat is also conducted through the casing to and from the pumpage, depending on the temperature of each. Typically, the difference in temperatures is small for the pumping conditions of chemical plants, and the effects are omitted for simplicity.
Our investigation has shown (Ref. 3) that the larger surface area can result in a nearly 40 F reduction in bearing operating temperature. The cooler bearings, in this case, result in approximately 13 percent longer life.
Increased Oil Sump Depth
A deeper sump allows contaminants to settle farther from moving parts, resulting in a cleaner layer of oil near the ball bearings. Contamination of the bearing races and the balls is the cause of microscopic deterioration of load surfaces, leading to failure. Statistics show (Ref. 2, 4) that a cleaner oil operation can increase bearing life by nearly 2.1 times (Ref. 3). Similarly, due to decreased air concentration, the oil oxidation rate by air is reduced for the larger sump. Again, for the type of pump studied in this work, this results in a 2 percent extension in bearing life (Ref. 3).
Labyrinth Oil vs. Lip Seals
The effects of oil contamination are further reduced by improved oil seals. A proprietary labyrinth seal design was tested against the lip seal. Both pumps were sprayed with water from a hose, simulating plant washdown. The spray was directed at various angles to the frame at the oil and seals area. The oil was then analyzed for water content. It was found that the previous design equipped with lip seals contained 3 percent water after 30 minutes of spraying, while the new design, with labyrinth seals, showed no water at all. Also, lip seals may cause wear and leakage after approximately only 2,000 operating hours.
Oil Level Sight Glass vs. Constant Level Oiler
A large sight glass allows direct visual observation to ensure proper oil level. It is standard in the new design, although a constant level oiler option is available. A constant level oiler is preferred by many users. When properly installed and maintained it can result in satisfactory operation.
However, because operation of the oiler is “blind,” depending solely on strict conformance to correct (and nontrivial) oil filling and maintenance procedures, it may lead to an incorrect oil level inside the frame. This can lead to hot operation and premature failure. Another problem is known as the oiler “burping” effect, resulting in a higher actual oil level than perceived (Ref. 5). Obviously, the new pump design can be equipped with both the sight glass and oiler if they are desired by the user. Such improvements in design can be combined because they benefit pump reliability independently. Based on this research, when all are added together, an improvement in pump life of up to 125 percent may be obtained.
To support and validate the theoretical derivations and assumptions as outlined below, a testing program was conducted, including lab testing and field data analysis. A comparison between operating temperatures and bearing life for the old and new designs was made. Tests were conducted with oil covering different levels of the lower ball of the bearings. The proper design level corresponds to oil at the middle of the lower ball of the bearing. At design setting, the new frame ran 40 F cooler, with a corresponding predicted life extension of approximately 6,000 hours.
Having measured the life increases, it is not difficult to assess the economic benefits of the new design.
Assuming the average value for MTBF of two years for the old design, a 125 percent improvement results in a four-and-a-half-year bearing life for the new design. The reciprocals of these numbers (1/2 years = 0.5; 1/4.5 years – 0.22) give an approximate number of failures or scheduled maintenance per year. The difference, 0.28, when multiplied by the average cost of repair ($260 in parts and labor and 3,000 pumps per plant) results in a yearly savings of: 0.28 x $260 x 3,000 = $218,400.
In addition, savings resulting from increased uptime and a reduction of lost production at approximately $500 per offline hour, assuming an average four hours per repair for offline time, would be: 0.28 x ($500 x 4) x 3,000 = $1,680,000.
The total, $1.9 million, is annual plant maintenance savings. Obviously, these numbers are approximate and can best be determined by individual maintenance departments using their operating specifics, but the savings potential due to improved design is clear.
Conclusions & Recommendations
The study demonstrated that substantial savings can be realized through improvements in pump design. To gauge such improvements systematically, it is imperative to quantify the benefits of each pump enhancement.
It is also important to maintain a proper balance between the solid theoretical foundations used for the analysis and the laboratory work, field testing and data supporting such theory. Users should seek quantitative data demonstrating improvements from pump manufacturers, including improvements in MTBF and MTBSM, enabling them to determine added value and other economic benefits.
This approach will improve communication between manufacturers and users, and lay the groundwork for the next step: further improvements in pump reliability.
- H. Bloch. PRIME I and II, Pump Seminar Series, 1992/1993.
- SKF General Catalog 4000 US (bearings), 1991.
- L. Nelik. Value Added and Life Extension with Regard to Reliability of X-Series 3196 ANSI Pump. Goulds Pumps, Inc. Internal Report, 1993.
- CRC Handbook of Lubrication, Vol. 1, CRC Press, R. Booser, 1983.
- L. Nelik. Goulds Technology Video Seminar. Constant level oilers versus sight glass, Series 0693-01.