Bearing applications are ideal places to look for process and cost improvements around your plant. As illustrated by the following examples, cost-justifying these enhancements often is very easy—in many cases, much easier than you might have imagined.
To cool or not to cool pump bearing housings Many refinery applications deal with flammable fluids. Safety and reliability are crucial in these services—that is why the API Standard 610 was developed. Although not legally binding, this well-known industry standard covers the basic requirements that impart reliability to process pumps. It is of the utmost importance in severe pump applications.
"Hot oil" (300 F/149 C to 850 F/399 C) is one of the most severe pump services there is, in that it creates high thermal expansion of pump components. The pump supports, in turn, must maintain shaft alignment at these elevated temperatures. Casing centerline support, without a separate support for the bearing housing, is normally used to address this problem. On the overall subject of safeguarding acceptable pump component temperatures, however, there is the (often incorrect) notion that all types of pump bearings require cooling. They dont.
Cooling water IS NOT desirable for rolling element bearings…
Decades of solid experience with literally thousands of pumps have shown that cooling water is not needed for the majority of centrifugal pumps in process plants worldwide. Regardless of convention and tradition, there is no net advantage to the use of cooling water in pump bearing housings equipped with rolling element bearings. Five statements explain these long-term findings:
In any case, for rolling element bearings that are properly installed and loaded per the manufacturer's allowable guidelines, cooling the oil is neither necessary nor is it advantageous. It may only be necessary to choose a higher viscosity oil, if indeed warranted.
Cooling water IS desirable for sleeve bearings…
The correct oil viscosity is needed for reliable oil application and long bearing life. In other words, an oil ring will not function the same way in lubricants with substantially different viscosities. Furthermore, the oil film thickness developed in the bearing will differ for different oil viscosities.
It is intuitively evident that some heat is being conducted to pump bearings. Moreover, oil shear in bearings produces heat. In many applications involving sleeve bearings, though, the oil may indeed have to be kept cool through the use of cooling water jackets that surround the bearings or by cooling water coils immersed in the oil. Cooling thus tends to ensure that the correct lube oil viscosity is being maintained. Nevertheless, there are many sleeve bearings that will simply not require cooling water. As explained in Ref. 1, they can be identified by temporarily shutting off cooling water in a controlled test, during which the steadystate oil temperature of a premium ISO Grade 32 synthetic is observed to remain below approximately 170 F (77 C).
To restate, while cooling may still be needed for sleeve bearings lubricated by oil rings, using cooling water jackets can prove disastrous to bearing life. Cooling water jackets that only partially surround bearing outer rings have often restricted the uniform thermal expansion of operating bearings and have been known to force bearings into an oval shape. There have been many instances where, as a result, bearing operating temperatures were higher with "cooling," and lower after the cooling water supply was disconnected and the jackets left open to the surrounding atmosphere. [Ref.1]
Similarly, cooling water flowing through a coil immersed in the lube oil also has been known to cause problems. Adherence to this "traditional" cooling method very often invites vapor condensation in bearing housings. Unless the bearing housings are hermetically sealed [Ref. 2], moist air fills much of the bearing housing volume. When it is cooled, the air sheds much of its water vapor in the form of liquid droplets. Since this condensate causes the lube oil to degrade, cooling the bearing environment can be indirectly responsible for reduced bearing life in pumps.
Regrettably, some pump manufacturers and installation contractors have been painfully slow in endorsing the deletion of cooling water. Others have been equally slow advocating superior synthetic lubricants and certain highly advantageous application methods. Where cost-justified, advantageous application could refer to pure oil mist ("dry sump") for both effective lubrication of operating pumps and, especially, the protection of non-running pumps against harmful environments [Ref. 3].
For rolling element bearings that are properly installed and loaded per manufacturer's allowable guidelines, cooling the oil is neither necessary nor helpful. Instead of following misguided tradition, and for bearing housings from which cooling liquid has been removed, reliability-focused users choose the right synthetic base stock (generally diester or diester/PAO blend formulations, [Refs. 2 and 3]). Reliability- focused users also select the correct viscosity and, as discussed previously, avoid the use of oil rings because of their serious limitations at today's higher operating speeds.
The cost benefits can be considerable. For example, by no longer using cooling water on hundreds of pumps with rolling element bearings, one petrochemical plant estimated that it could save $120,000 in water consumption, piping maintenance and water treatment cost annually. Deleting cooling water also was expected to prevent at least three pump failures per year (at $6700 each). The annual savings were estimated to exceed $140,000 at this facility. This, again, is just one instance where the implementation of a cost-saving measure has shown positive reliability impact and instantaneous payback. All it takes is for reliability professionals to familiarize themselves with 30-years-worth of well-documented prior experience and the scientific principles behind it—which are very easy to explain to plant personnel. If necessary, a simple wellinstrumented test will convince even the most traditional doubters of the laws of physics.
Double-row angular contact bearings with two inner rings
Pump bearings can fail for a number of different reasons. Two of these are bearing overload and bearings being too lightly loaded. What at first seems like a paradox— failure due to loads being too light—can be explained by the analogy of an aircraft landing. The tires skid until their peripheral velocity matches the forward velocity of the aircraft. Skidmarks on the runway and smoke and noise coming from the tires attest to this fact.
Until the late 1990s, failure avoidance through upgrading was pursued by some pump users, while others were content with frequent bearing replacements. Thus, upgrading the double-row angular contact ball bearings (DRACBBs) in pre-sixth edition API-610 standards, or in ANSI and ISO-style pumps to meet "ANSI-Plus" standards generally meant one of two things:
When it became evident that conventional, 30° contact angle, double-row angular contact ball bearings often represented the limiting factor in achieving extended pump life, the need to devise a retrofit bearing with characteristics approaching those of the API 610-recommended 40° contact angle back-to-back orientation arose. This prompted a multinational bearing manufacturer to develop DRACBBs with two inner rings and 40° contact angles. [Refs.1 and 4]
Using a specially designed DRACBB with two inner rings (Fig. 1) requires no labor-intensive retrofit work and represents an ideal upgrade for many API and ANSI/ISO pumps (Fig. 2). The DRACBB bearing employs a closely controlled axial clearance. This optimized clearance promotes load sharing between the two rows of balls— a design that reduces the possibility of skidding in the inactive ball set without the use of a preload. Skidding produces heat and often causes the oil film to be wiped off. Either way, metal-to-metal contact will result and bearing failure risk increases exponentially.
Preloading a bearing can help prevent skidding, but may also have the undesired effects of generating excessive heat or contributing to poor bearing performance. While one ball set is supporting the axial load, the backup set becomes inactive, supporting only a portion of the radial load. Without sufficient loading, the motion of the balls in the inactive set leads to skidding and heat generation.
The new separable inner ring DRACBBs' shaft and housing fits are identical to those for standard SRACBB and DRACBB with comparable sizes. ISO k5 is the recommended shaft tolerance for this bearing in most pump applications. This tolerance produces an interference fit between the bearing inner ring and the shaft. Interference fits are necessary for bearings supporting any radial load. A lighter fit using modified tolerances may be necessary for bearings mounted on shafts made of stainless steel, or for the occasional bearings that have a large temperature differential between the inner and outer rings.
The mounting recommendations for DRACBBs with two inner rings differ slightly from those applicable to the conventional double-row configuration. But, the benefits of selective upgrading to these retrofit bearings will more than make up for the inconvenience of looking up a different instruction sheet.
Assume that an incremental outlay of $40 per bearing plus $120 in conversion cost were to lead to an $8,000 repair avoidance on a large ANSI/ISO pump every four years (or $2000 annually). In that case, $160/4 years = $40/year has returned $2000/year. That's a rather attractive benefit-to-cost ratio of $2000/40, or 50:1. Or, project an ultra-conservative scenario of spending $200 on conversion and extending a previous 1.5-year MTBR to a post-conversion MTBR of three years. In that case, avoiding even a $6000 repair will still yield a solid 10:1 payback.
As this four-part series has shown time and again, many reliability improvements are readily available and easy to justify from an economic standpoint. As an example and dealing only with compressors and pumps, advances in high-performance polymer materials and synthetic lubricant technology can lead to significant extensions in equipment run times, or mean times between repairs. If a reliability professional must wrestle with a population of centrifugal pumps, he or she might do well to consider several of the easily cost-justified enhancements described in this series of articles. For instance, it would seem appropriate to look into:
Removal of cooling water from bearing housings equipped with rolling element bearings Other upgrade opportunities were described years ago in this magazine (formerly known as Lubrication & Fluid Power). They include the merits of:
There surely are other "things" that can be done to decrease pump and compressor failures—this series just highlighted several of the simplest and most cost-effective. Needless to say, reliability-focused plants and users will follow up with the speedy implementation of these and other cost-justified upgrade and enhancement measures.