Listen, look, but keep them running
Excerpted from 100 Years of Maintenance and Reliability: Practical Lessons from Three Lifetimes at Process Plants by V. Narayan, James W. Wardhaugh & Mahen C. Das (Industrial Press)
The chain of habit coils itself around the heart like a serpent, to gnaw and stifle it. - William Hazlitt, Writer and Critic.
Author: Jim Wardhaugh
Location: 2.3.3 Corporate Technical Headquarters
39.1 Background
I left my Far Eastern idyll for cold dark Europe. Those who move the pawns in the corporate headquarters noticed my success in my assignment in the Far East. So they head-hunted me and I joined this elite group. The job was to help locations that were less well performing to make significant improvements. I met Vee Narayan, who was already in this group; later, Mahen Das also joined us.
As an electrical engineer, I had been extensively involved in previous lives for the implementation of computerized maintenance management systems (CMMS). So I became responsible for electrical maintenance and CMMS implementation.
An Electrical Engineers’ Conference gave me the opportunity to prompt the collection of some data. In the Far East, I found that motor repairs were a significant ongoing manpower consumer and, hence, cost. I decided to put some effort into collecting data on this subject so that we could make informed decisions.
39.2 Motor Failure Data
A few of our bigger locations and the Institute of Electrical and Electronic Engineers (IEEE) had been collecting data for many years. We captured this data and collated them (see Tables 39.1 and 39.2). We found large differences between our numbers and those from IEEE. We can account for these in part by the difference in the ways of collecting data and in definitions. We believe that IEEE figures give a pessimistic view of overall failure rates that one might expect from a professionally-managed site.
The information shows that motors are generally very reliable. There are few electrical failures and most failures are bearing related. The most significant underlying causes of failure are:
• Defective components
• Poor installation/maintenance
• Poor lubrication
• Water ingress
39.3 Electrical Failures
Scrutiny of the data showed three prime causes of electrical failures:
• Catastrophic failure due to bearing collapse and rotor rubbing on the stator
• Water ingress due to cooler leaks or cleaning with high-pressure water hoses
• Breakage of connections (due often to inadequate bracing)
There did seem to be some deterioration of motors with age. This was most apparent in large motors with a high starting frequency. We found that the larger the machine, the higher it was stressed. Manufacturers had algorithms which could predict the end of useful life with reasonable accuracy. For this they needed service and operational data, such as frequency of starts. This might be worth doing for older machines in critical services.
39.4 Bearing Failures
Scrutiny of the data showed four main causes of (premature) bearing failure:
• Wrong (or inadequate) bearing installed
• Poor installation practice causing initial damage to the bearing
• Poor lubrication regime
• Poor alignment of driver and driven
39.5 Performance of Seven Motors at Locations
We looked at the motor maintenance activities in seven of our companies in six different countries. Each was a company plant built to corporate standards, with most rotating equipment having an installed spare. However, there were a variety of maintenance strategies in place.
Figure 39.1 summarizes our findings. It gives the percentage of each site’s inventory of motors removed to the workshop for significant repair each year. These percentages have been broken down by reason for removal:
• Breakdown (i.e., the motor had been run to failure)
• Condition monitoring had indicated imminent failure
• Time-based overhaul regime in place for some or all of the motors
Considering this, we found that:
• The large proportion of time-based overhaul activities of Location 7 did not seem to reduce breakdowns significantly.
• Location 6 did seem to be somewhat more effective, but arguably was still not cost effective.
• Location 5 had many breakdowns, even though a significant percentage of motors were repaired because condition monitoring was predicting imminent failure.
• Their condition monitoring did not seem very effective in predicting and/or pre-empting failures.
• The site had an extreme blame culture.
• Location 1 had minimized repair efforts by using run-to-failure as a default strategy.
• The small amount of time-based maintenance was for a very few un-spared furnace fans which were overhauled when the plants were shut down every four or five years.
• This location had a fairly skeptical view of the merits of condition monitoring. They would keep motors running until imminent failure was very apparent.
• What they also had found was that running motors less than 30 hp to failure did not result in significant additional consequential damage and cost compared to pre-emptive action.
• The proportion of breakdowns is fairly constant whether you do condition monitoring and/or overhauls or just let things run to failure.
39.6 Summarized Findings
From our review, we learned that electrical motors are very reliable. In more detail, we found that:
• Windings do not exhibit significant wear-out unless they are:
• Too frequently started or
• Large and highly stressed
• Winding connections are prone to breaking if not well braced against movement.
• Bearings do wear out, but long life is a function of a few simple things:
• Correct bearing selection (not necessarily the same as found in the machine)
• Correct installation, using bearing heaters etc. to minimize damage
• A good lubrication regime (correct type and quantity of lubricant)
• Correct alignment
• Smaller machines (less than 30 hp) make up the bulk of the population and can be run to catastrophic failure without significantly increasing consequential damage to shaft or windings.
• In many locations, there is a high level of installed sparing so the consequences of failure are low.
• Motor condition monitoring has to be quite cheap; else it is not an economical strategy. We concluded that:
• Vibration monitoring could not be justified for most motors; it became viable (just) if you were already going to check the driven equipment.
• Ultrasound was potentially useful if you had many close packed fractional horsepower motors.
• Winding monitoring could not be justified for most motors; a special justification was needed for critical applications.
• Winding monitoring is improving, so this would be kept under review.
39.7 A Maintenance Strategy for Refinery Motors
We carried out an evaluation of possible strategies and monitoring techniques. There were many magic bullets being advocated by credible universities, consultants, and large companies.
We concluded that for very large critical machines:
• A proprietary monitoring installation that continually monitors vibration, axial displacement, etc., should be the norm.
• Information should be centrally monitored.
• Alarm and trip parameters should be set after agreement with the manufacturers.
• We should do regular (annual) internal inspections, using a boroscope.
For the bulk of industrial motors, we concluded that regimes described briefly in Tables 39.3 and 39.4 were justifiable and should be the norm in most environments. Obviously in environments which are extremely arduous, additional steps may be needed.
If ultrasound and thermographic tools are available, a minimum effort periodic inspection could be beneficial. They are particularly useful if you have a large number of motors very small, and closely packed.
39.8 Lessons
• Buy reasonable quality motors which are inherently reliable enough for your application.
• Ensure good lubrication.
• Select correct bearings and install correctly.
• Run small spared motors to failure. Run to failure or at least imminent failure is a very respectable strategy for equipment where the consequential loss is low.
• Predicting and pre-empting failure, however cheap, is only cost effective if you can pre-empt catastrophic failure or major production loss. In general, pre-emptive actions must cost less (and probably significantly less) than the consequential loss due to failure.
• Condition monitoring can be costly and ineffective so you need to audit the effectiveness and cost effectiveness of the system.
• When motors fail, investigate to find root cause and try to eradicate repeat failure causes. (Think correct lubricant, correct lubrication regime, correct bearing, starting frequency.)
39.9 Principles
Preventive and predictive maintenance are generally sound strategies, but not universally applicable. Run-to-failure strategies are perfectly acceptable in many common situations. Maintenance strategies must be based on a rigorous understanding of the risks of failures, not on current fashions.
Recent Comments