From a recent sequence of correspondence, we discern that certain rulings were being made years ago, although the rationale used to make these rulings has not been explained. This example involves a reliability engineer in Australia. He works for a polyethylene manufacturer that formerly belonged to MNXC, a leading multinational corporation. The plant inherited operating and maintenance manuals from MNXC.

The reliability engineer was aware that MNXC in its maintenance practices manual had documented “proven-best practices” for spare-pump operation. MNXC was said to have claimed that the more frequently a pump is started, the lower its mean-time-between-repairs (MTBR). This, MNXC explained, was attributable to pump startup being the most severe or stressful transient operation to which pump seal(s) and bearings are subjected.

Maintenance schedule.

The Australian reader informed us that MNXC’s maintenance practices manual advocated a four-category criticality ranking of all pumps and to then swap the pumps as follows:
Emergency: Every 4 weeks
Vital: Every 12 weeks
Normal: Every 52 weeks
Low: Never.

The reader also noted—quite correctly—that the MNXC maintenance practices manual recommendation does not seem to follow logic. It was apparent to this reliability professional that, for its most critical equipment, MNXC applied a frequent start strategy which, as they correctly inferred, would yield the shortest MTBR. The Australian engineer summarized his correspondence by recalling that we at HP were rather familiar with MNXC’s practices and asked for our opinion on the matter.

Answer.

Well, the manual didn’t exist at MNXC when I opted for early retirement in 1986. In the intervening years, I have had ample opportunity to research the issue and to observe prevailing practices in many countries. I subsequently alluded to pump “swap-over” topics in books and articles on many occasions.

As to the issue raised by the reader, there are several possibilities. Some, I will admit, are alluded to with a twist of irony, but they deserve to be mentioned here:
a) There is confusion with the terminology “swap” vs. “test run.” The old manual might advocate a “swap” once a year, but (elsewhere) asks the operators to test run the second pump for at least four hours every month.
b) The old owner company, MNXC and the person(s) compiling the manual have overlooked the fact that letting a pump sit idle for 52 weeks and to then expect it to run flawlessly is stunningly naive and will—ultimately—cost the owner dearly. Experienced reliability professionals know well that, after a year, the bearings of the stand-still “spare” pump will have degraded due to two actions:
• One action is micro-vibration transmitted from adjacent running equipment; such vibratory motion causes the oil film to be wiped off. The result is metal-to-metal contact and false brinelling (Fig. 1), usually at the race near the lowermost bearing balls.
• The second typical action results from corrosive damage, unless, of course, dry-sump oil mist is used for both lubrication and stand-still protection.
c) Whoever made the rules is inexperienced and/or bases this advice on the premise that “one has to do something in order to justify being on the company’s payroll.”
d) The owner company is planning to save money and will later sell the asset to some unsuspecting buyer. The enterprise will look good—low operating cost, high profits. Two years later, the facility will be well on its way to becoming maintenance-intensive and unprofitable.

Fig. 1. False brinelling caused by vibration of   non-running equipment

Personally, I believe what a leading producer of petrochemicals (LPC) advocated and practiced in 1986 is still correct: Swapping every four weeks will protect the bearings and will keep the stagnant liquid product in the piping and seal regions from partially vaporizing. Periodically swapping pumps will also serve as a training exercise to keep operators knowledgeable and alert.

An LPC engineer told us, at a conference in 2008, that LPC is now swapping pumps every six weeks. That’s sufficiently close to our old four-week rule, and I can live with that. The “new rule” that the reader mentioned in his note makes no sense, but I’m past arguing.

The wheel is being reinvented even as we speak. Interestingly, pump MTBFs at some plants have declined since 2002. The “new rule” undoubtedly contributes to the MTBF decline. In any event, the important answer and technical explanation is found in b), above. You can disregard the rest. HP