It’s been a hot topic in industry for decades. One of the most frequently quoted studies related to electric motor reliability is a 1983 Electric Power Research Institute (EPRI) project performed by General Electric (GE)  that's been used to support a variety of programs, equipment and other motor strategies. While searching for a copy of the original study, this author has often cited other papers that reference its findings. Such a paper was recently made available by the Institute of Electrical and Electronics Engineers, Inc. (IEEE). A review of it indicates some statements that have discussed or tried to interpret EPRI's findings over the years have been incomplete and/or incorrect. The good news is that the EPRI project wasn't the only major study to target electrical and electric-machine reliability. Studies conducted by others from 1962 through 1995, including those of an IEEE Power Engineering Society group, were supported by various industry groups as late as 2010.
What’s particularly interesting about the non-EPRI research efforts is that despite their focus on different industries (such as utilities, petrochemical, general-industry and commercial-building sectors), they report similar results. Here, we review what these studies really mean in regard to their primary emphasis—larger machines. This includes identified reliability issues and recommended improvement strategies. While the full breadth of the related studies is more than we can cover in this article, the information presented in these pages will significantly impact how you look at your motor systems.
Regarding the EPRI study
The percentages shown in Fig. 1 are often cited as the conclusion of the EPRI study—which is correct. However, the details behind those percentages are intriguing. This includes the number of motors that failed more than once (and the apparent causes of those failures), as well as the general reliability of electric motors in utilities.
First it was noted that more than 90% of the failures occurred in 54% of the facilities evaluated, and half of the failures occurred in 17% of the facilities. This means that a majority of the failures occurred in less than half of the evaluated sites. The average failure rate of motors across all the facilities was 3.4% per unit annually, with some operations having a higher rate, and 46% of them having very low failure rates. In all, the study found that those plants at the extremes had a failure rate of 9.3% per year (17% of facilities) and 13% of the sites had about a 0.8% failure rate.
There were 4797 motors evaluated in the study, with a total of 1227 failures on 872 units. This means 335 of the 1227 failures were repeat failures. The best sites saw some of their motors fail two to three times versus results from the median group, which saw failures of four or more times, and the worst group, which saw an even higher repeat-failure rate.
The apparent causes of failure were also surprising: Only 34.1% were blamed on misapplication and misoperation. (As noted, though, causes weren’t specified for more than half [50.2%] of the failures.) The failure modes were correctly identified, with repeated incidents being the same as the original failures. Table 1 identifies the failures and the percent of each failure mode.
Of these failures, design was determined to be 39.1% and workmanship was 26.8%. In effect, the survey determined that 65.9% of the motor failures were related to the manufacturer and rebuilder.
The failure rate by manufacturer was found to range from 0.84% to 5.27% for the top seven OEMs: 16.44% for one manufacturer, and a combined total of 6.50% for all the others. The manufacturers were not identified.
One issue brought to light by statistics in the 1983 EPRI study is that insulation-to-ground faults are the majority of winding faults. Quotes related to this study and other industry statements identify turn faults as the initiation of failure—but such a statement is not found in the EPRI research or follow-up studies.
Review and comparison of studies
In the motor-failure studies of the 1980s, it was determined that a given population of motors had an average failure rate of 0.0708 failures per unit, per year (FPU) for general industry  and 0.035 FPU for maintenance-intensive industries such as utilities . In 1995, new research would support the original assumptions. These industry studies found that in machines with required minimum protection, such as fuses or breakers, the failure rate was 0.0707 FPU, while those with embedded thermal protection had a failure rate of 0.0202 FPU, or less than 1/3rd of the failures. 
Maintenance was also found to have a significant impact through all of the studies. When maintenance frequency was involved, the post-EPRI studies all found that frequencies under a year had the best impact. The 1985 IEEE study identified that maintenance performed with a frequency under 12 months resulted in 0.0124 FPU, from 13 to 24 months 0.0506 FPU, and greater than 25 months 0.0881 FPU. Machines that were maintained within the 12-month period were also found, within the survey, to have excellent practices resulting in the failure rate of 0.0124 FPU while all others had failure rates in excess of 0.0681 FPU.
A key difference between the EPRI and IEEE studies is that the 1985 IEEE research didn’t just look at general failures, but also broke out service factor, speed and maintenance. The 1995 IEEE survey further modified the findings by identifying size and voltage to determine factors that relate to each. A 2010 paper on root cause failure analysis supported the findings of the 1995 study .
One consideration for these studies is their ages. From the first study published in 1974 (relating to electrical reliability of electrical equipment in industrial plants) through the 1995 IEEE study, fundamental facts had not changed: Machine reliability based upon application, enclosure, service factor, speed, protection and level and type of maintenance was explored, and the combined studies covered virtually all industries—from petrochemical, chemical and utilities to general-industrial and commercial applications.
Application of the studies to large machines
As the studies provided similar data based upon failure rates, (and it can be assumed that variations in failure rates and reliability of machines by facility in the EPRI study relate to the level of maintenance), we will focus on information in the IEEE studies. This is broken down by size, enclosure and speed, providing the ability to demonstrate the importance of maintenance on large machines.
The primary difference is identified in Fig. 2 where various faults found in the machines were substantially different. It is noted that the EPRI study focused on utility motors that were 100 hp and larger, while the IEEE study related to machines of 10kW (~15 hp) and larger at 50 Hz and 60 Hz.
As shown in Fig. 3, actual failure modes for each industry group were also quite different. A majority of faults were a result of bearings, with windings second, followed by the rotor, then all other faults combined. The 1985 IEEE survey covered industrial and commercial facilities, while the 1983 EPRI study only covered utilities. The 1995 IEEE survey covered petrochemical and similar industries. Other differences with the 1985 survey include a focus on machines ranging from 200 to 10,000 hp, voltages to 13.8kV and induction, synchronous, wound-rotor and DC motors.
Based upon the breadth of industries covered, we will review the following data as it relates to the 1985 IEEE survey and machines over 1000V. From an overall industry standpoint, 2300 and 4160 Vac machines have a median failure rate of 0.0714 FPU for induction motors; 0.0762 FPU for synchronous motors; and 0.0319 FPU for wound-rotor motors.
By further breaking down the survey information, it can be seen that motors from 500 to 5000 hp had a median failure rate of 0.0730 FPU, and those from 5001 to 10,000 hp had a median failure rate of 0.2169 FPU. In relation to motor speed and failure rate: 0-720 RPM is 0.1004 FPU; from 721-1800 RPM is 0.0721 FPU; and 1801-3600 RPM is 0.0519 FPU. In effect, larger, slower-speed motors had a higher failure rate (most in the survey being induction and synchronous). The wound-rotor machines that were covered tended to be a smaller horsepower.
Based on the IEEE research, use of continuous monitoring—such as temperature and vibration—can reduce the failure rate by about two-thirds. None of the studies has identified the effect of the use of partial discharge testing on machines over 6000V. It can, however, be assumed that, in most cases, such practices and technologies are used for fault detection rather than for winding protection. The question is, “Does this have an impact?”
The IEEE studies identified the number of faults detected by a variety of technologies and maintenance practices—with the median downtime hours per failure based on the fault being detected as part of a maintenance practice or during operation. As reported by the 1985 IEEE survey, Fig. 4 shows at what point the failures were detected.
The level of a maintenance program and the frequency of maintenance practices also had a marked impact on not just the failure rate, but also the median hours of downtime per failure (Table II).
The maintenance practices that encompassed "excellent" maintenance included:
One explanation for the higher failure rate and lower average associated production disruption was that potential faults were detected as part of the maintenance practice.
Past electric motor studies have been incorrectly quoted for many years. A review of the associated studies has identified that actual opportunities are far greater than what previously had been discussed. The purpose of this article was to demonstrate some of the information in relation to large, medium-voltage machines. Primary opportunities include the use of continuous-monitoring systems (i.e., temperature and vibration) and the application of technologies and maintenance practices that will avoid or detect electrical and mechanical faults. The result is about a two-thirds reduction in failure rate and a significant decrease in production downtime.
While the research referenced in this article may have been conducted and published from 1973 through 1995, the information on failure rates was similar. Furthermore, papers published as late as 2010 continued to support the original findings. The primary differences in the research were the targeted industries and distribution of the failure modes listed by each study. That, in short, is what the studies really “say.” MT
1. Albrecht, et.al., “Assessment of the Reliability of Motors in Utility Applications,” IEEE Transactions on Energy Conversion, Vol. EC-2, No. 3, September 1987
Albrecht, et.al., “Assessment of the Reliability of Motors in Utility Applications – Updated,” IEEE Transactions on Energy Conversion, Vol. EC-1, No. 1, March 1986
2. “Report of Large Motor Reliability Survey of Industrial and Commercial Installations, Part I,” IEEE Transactions on Industry Applications, Vol. IA-21, No. 4, July/August 1985
“Report of Large Motor Reliability Survey of Industrial and Commercial Installations, Part II,” IEEE Transactions on Industry Applications, Vol. IA-21, No. 4, July/August 1985
“Report of Large Motor Reliability Survey of Industrial and Commercial Installations, Part III,” IEEE Transactions on Industry Applications, Vol. IA-23, No. 1, January/February 1987
3. Thorsen and Dalva, “A Survey of Faults on Induction Motors in Offshore Oil Industry, Petrochemical Industry, Gas Terminals, and Oil Refineries,” IEEE Transactions on Industry Applications, Vol. 31, No. 5, September/October 1995
4. Bonnett, Austin H., “Root Cause Failure Analysis for AC Induction Motors in the Petroleum and Chemical Industry,” Proceedings, 57th Annual Petroleum and Chemical Industry Conference, 2010
Howard Penrose is the Vice President of Engineering and Reliability Services for Dreisilker, the Webmaster of the IEEE Dielectrics and Electrical Insulation Society, and the Director of Outreach for the Society for Maintenance and Reliability Professionals (SMRP). He has won five consecutive UAW and General Motors People Make Quality Happen Awards (2005-2009) for energy, conservation, production and motor management programs developed for GM facilities globally. An SMRP Certified Maintenance and Reliability Professional (CMRP), Dr. Penrose is the author of the Axiom Business Book Award-winning (2008 Bronze and 2009 Bronze) book Physical Asset Management for the Executive (Caution: Do Not Read This If You Are on an Airplane), and the 2008 Foreword Book of the Year Finalist textbook,