December 2016

Columns

Reliability: Troubleshoot hot bearings in hot pumps

The reliability engineering unit supervisor (REUS) working at a 400-Mbpd refinery wrote that he was greatly concerned about a bearing temperature excursion. He realized that certain rules of thumb had been handed down to his maintenance employees, and that not all of these conveyed the same numbers and guidance.

Bloch, Heinz P., Hydrocarbon Processing Staff

The reliability engineering unit supervisor (REUS) working at a 400-Mbpd refinery wrote that he was greatly concerned about a bearing temperature excursion. He realized that certain rules of thumb had been handed down to his maintenance employees, and that not all of these conveyed the same numbers and guidance.

It is fair to say that bearing temperature concerns cannot be dismissed without explanation. Allowable temperatures are related to oil type and application method. Old guidelines, some of which originated over 100 yr ago for mineral oils, are rarely relevant now, and are much too conservative for today’s premium-grade synthetic lubricants. The reasons that problems reappear at some plants and disappear at others may have to do with vendor support and failure analysis expertise. If bearings fail, bearings must be at fault, or so the (erroneous) reasoning goes.

Since both vendor support and owner expertise are in short supply, trial and error solutions are often pursued. After trial and error solutions lead to an impasse, the plant’s owner-operators often (finally) turn to whatever source will give them free advice, and this is what the REUS did in this instance. Fortunately, he transmitted enough data to diagnose the problem with reasonable certainty.

The problem report

We can summarize the query by stating that the hot service pumps at issue—with a fluid temperature of ≥ 200°C—were experiencing high temperatures on the radial bearing non-drive end (NDE) and the drive-end (DE) thrust bearings. The REUS and his colleagues in various operating job functions noted that bearing temperatures became progressively higher during the hot summer season in their region of the world. These hot service pumps had been produced by a well-known manufacturer which, as is so often the case, did not support its products in the manner the owner expected.

The process pumps were between-bearing models with rolling element, deep-groove ball and angular contact ball bearings lubricated by pure oil mist. The pumps incorporated dual mechanical seals and used low-temperature barrier fluids at 15°C (API Flush Plan 32) and 60°C (API Flush Plan 54), respectively. The REUS’s facility has 10 oil mist generators with ISO VG 100 synthetic oil. The pumps at issue are equipped with low-conductivity (stainless steel) shafts, and with cooling fans on both inboard and outboard bearings. The pump manufacturer had defined temperature pre-alarms and danger limits at 95°C and 105°C, respectively.

However, in this refinery’s hot product pumps, the temperature without external cooling air reached up to 110°C while the pumps were running. Interestingly, the temperature rose to 125°C when the pump was stopped and put on hot standby duty. After the facility had installed larger-size oil mist fittings—reclassifiers for a mist flow of ~0.60 standard cubic ft/min. (scfm)—on the thrust-end bearings, the maintenance workers noted no significant temperature change. The results were quite similar when traditional lubrication with oil rings was tried; virtually no significant difference in outcome was detected.

The REUS provided details on oil mist entry and exit locations, but those details had been properly addressed by the highly experienced oil mist provider in the past. The REUS also asked that we urgently review his request and share our experience on hot service pumps with pure oil mist lubrication in similar applications elsewhere.

The reply

Strong evidence indicated that several deviations from normal were at fault. A single deviation by itself would be of no consequence, but the combination of several deviations were likely causing the problems with the pumps in hot service at this plant.

With regard to troubleshooting, we advised the REUS that hundreds of hot service pumps with pumped-fluid temperatures in the vicinity of 300°C, and some as high as 400°C, have been in successful service on pure oil mist for many decades all over the world. Therefore, for the REUS’s facility to have this concern/issue was an indication that the “problem” had to be found somewhere within that installation.

It was rather telling that this owner-operator experienced high temperatures even on a normally lightly loaded bearing, i.e., the radial bearing. Moreover, as the REUS reported, the problem manifested itself with oil mist and also with traditional oil ring lubrication. Oil mist is not the solution to lubrication errors, nor for installation and/or fabrication problems. The standard oil rings in traditional lubrication are designed for ISO VG 32, and will perhaps still work with ISO VG 68 oils. However, only very special oil rings or clamped-on flinger disks will work with ISO VG 100 oils.1

Therefore, his experience with oil rings in ISO VG 100 service is irrelevant to this issue. This lead us to suspect that someone (“persons unknown”) is unaware of the high thermal expansion/growth of shafts with low-conductivity stainless steel shafts. It is likely that the bearing-to-shaft interference fit in his hot service pumps was much too tight, and that the REUS is unaware as to what the fit should be in this instance. His world-scale refinery probably does not spend money on the right type of training, and may have closed the door on buying its bearings from a vendor with many decades of application experience.

A good text would make the case for measuring and recording the bearing interference fits. They should be close to zero in the non-running, at-assembly condition. These special fits should be retrieved from the user-owner’s computerized maintenance management system. Various literature addresses these basics, and favoring bearing manufacturers that still employ application engineering groups would have prevented this issue.

FIG. 1. The effect of preload on bearing life.

At times, equipment manufacturers get away with the assertion that disclosing dimensions is against their proprietary interest and, therefore, confidential. The REUS’s correspondence contained no data on bearing or housing dimensions, shaft speed and so forth. A proper failure analysis mandates access to cross-sectional views of the bearing housings and requisite dimensional details. He and his employer should insist on vendor support, and while deserving praise for including “clues” as to the probable nature of the problem, they probably did not have access to all the facts that must be discovered. Sending cooling air over a bearing housing restricts the bearing outer ring from thermally expanding and increases preload. With the pump stopped, preload is partially relieved, but heat soak becomes more pronounced. Combined with an unusually high interference fit, the radial and thrust bearings continue to exhibit high bearing temperatures under running, non-running and shutdown conditions.

With everything in working order and proper attention given to dimensional issues, the use of an ISO VG 100 synthetic lubricant could be of benefit in hot climates and hot services. However, under conditions of preload-induced temperature stress, an ISO VG 100 may not conduct away as much heat as an ISO VG 68 synthetic. When certain deviations combine, use of the lighter oil will sometimes be advantageous.

Literature shows how to reduce all machinery problems to just seven cause categories. FIG. 1 shows how excessive preload can greatly and rapidly impair bearing life. The REUS’s field experience corroborates the rules of thumb for temperature increases due to such preloads; the effects of these increases can be drastic. HP

LITERATURE CITED

  1. Bloch, Heinz P., Petrochemical Machinery Insights, Elsevier Publishing, Oxford, UK, and Cambridge, Massachusetts, October 2016.

* Portions of this column were excerpted from the author's latest book.1

 

The Author

Related Articles

From the Archive

Comments

Comments

{{ error }}
{{ comment.comment.Name }} • {{ comment.timeAgo }}
{{ comment.comment.Text }}