Final echnical e

1 Final echnical e .e 7Z-~C-j'I &ECTRIC OOTOR ELIABLT .iorTE/J?~.?R/SmithShaker Research CorporationIb m400"ft Ak Nifp o*, ?4~w Vawk2 -AThis report hat been reviewed by the RADC Information iff ice (01) erd Isreleasable to the National Technical Informuation Service (NTIS). At NTIS itwill be releessr'le to the generAl rublic, including foreign ,'-408 has been reviewed and is appeoved for : J 4 ALESTER .1. GUBBINSP roject EngineerAPPROVED:JLSEJH J. -NRESKYChl~ef, Rteliability and Compatibility Divis~ionVOR TME COMN2 Ilk: cF, ftl~~VON* 0N", 0'4 SICUIRITY OF THIS PAGE (Whe~n Date Entered)REPOT DCUMNTATON AGEREAD INSTRUCTIONSREPORT____DOCUMENTATION_____ PAGE _ BEFORECOMPLETINGFORM7 EONT P(UMMIR 12. GOVT ACCESSION No. 3. RECIPIENT'S CATALOG NUNUERMC-TR-77-408 _____I_____4.:,4 TLE (isid Subti tle) 11. TYPE OF REPORT & PERIOD COVEREDELCRI OTOR RELIABILITY model -See 7fi -Sep 770. PERFORMING ORG. REfPnRT NUMSER7.)

2 AUTHOFI(s) S. CONTRACT 0OR GRANT NUMISER(s)D. S. WilsonR. Smith F30602-76-C-04l3 "49. PERFORMING ORGANIZATION NAMZ AND ADDRESS 10. PnOGft'M ELEMENT, PROJECT. TASKAREA A WORK UN IT NUNSERSS haker Research Corporation 672 Bali1ston Lake NY 12019 672_____ OFFICE NAME AND ADORISS OATt December 1977 1 Rome Air Development Center (RBRT) 13. NUM9 OER OF PAGESG riffiss AFB NY 13441 107/ ~ .ONTOIN AJECY NAME A ADDRESS~it different Item Contuofitnd Offie) IS, SECURITY CLASS. (of this report)SamelIS. ITRIOUTION (Ofthis Report) /Approved for puwlic releass; , DISTRISUTION STATEMENT (of the abstract .,.tsied in Bloek 20, it difirran: fromy Report)SameIS. SUPPLEMENTARY NOTESRADC Project Engineer: Lester 3. Gubbins (RBRT)1S. KEY WORDS (Coeuit"u an reverse side it qscoo-sar,' mid Identify by btock number)Reliability BearingsElectric Motors Electrical WindingsFaiIlsr. Rate80. A-11110 WRAa, (eG olan rverse e de It micessairy and identify by &Jack )"An existing data failure bank consisting of over 2000 actual fractionalhorsepower (FHP) motor failures was upgraded and utilized to develop a failuremodel for HP1 motors.

3 Motor fal-lures were dominated by two failure modes; , vindisug failures and bearing mathe~matical reliability model was developed for each failure modeutilizing a Weibull cumulative distribution function and regression NVSIQLT UNCLASSIFIEDSEicURITY CMLI SIFICATION OP~ THIS PAGt (1111en Date shlerim!)UNCL&SS IFIEDSECURITY CLASSIFICATO FTI AEW~D~ )R-'5esults of a literature survey of both failure modes were used to supplementfailure data where required and verify the developed modes were recombined into a competing risk cumulativedictribution model for use in predicting overall motor life and CLASSIFICATION OF THIS PAOE(Vten )PREFACEThis Final report was prepared by the Shaker Reseirch Corporation ofBallston Lake, New York under Contract F 30602-76-C-0413. The work reportedcovers the period from September 1976 through August 1977. The report wasprepared under direction of Mr. D. S. Wilson with major technical contribu-tions from Mr.

4 R. Smith of Shaker Research. Failure data collection wasassisted by Mr. A. Fries of the Rotron Manufacturing Company. The RADCP roject Engineer was Mr. L. | ..IViiS.."[. Introduction ..1 ..3 Discussion .. Data Bank Analysis .. Bearing Reliability model .. Literature Review of Bearings .. Development of Bearing Life Prediction model .. Motor Winding Reliability model ..# .. Literature Review of Electrical Failures ..- Regression Analysis of Winding Failures.. Influence of Altitude on Bearing and WindingTemperat ,re..0 .#.. FHP Motor Failure model .. Failure Rates of FHP Motors.. Use of FHP Motor Failure Models .. Conclusions .. References ..0.. List of Symbols ..75 Appendix A Multiple Linear Regression Analysis Techniques andExamples .. RegressAion Analysis .. Use of. Regression Analysis.. Correlation Coefficients, ..0 . Correlation Matrix .. Correlation Coefficients of Transformed Data.]

5 Selecting an "Optimum Set" of ExplainingParameters, ..9 .. A Large of Independent Variables ..o .. Regression Tables Explained ..o .. Examples of Data Analysis Reduction ..85 Appendix B Recoamended Repl2cement of Reliability model forElectvonic Cooling Devices and FHP Motors forMIL-RDBK-217B .. Motor Reliability Prediction model .. Predicting Failure Rate Without Utilizing Tables..98$ Bearing Characteristic Life Resultingfrom Thermal Cycling .. 'Winding ChavaLtetristic Life Re3ultIng from ThermalCycli ng ..0. *..0..0 .. Examples of Motor Failure Rate Prediction..0 .a.. Example No. 1 .. Example No. 2..104-,*a4A:?.V LIST OF FIGURESI Fractional Horsepower Motor Reliability Data Life 22 Versus Temperature ..a ..222 Motor Temperature Rise Distribution ..273 Bearing Life at Low Temperature..354 Bearing Life over Temperature Range.. 375 Bearing Test Data -- All Greases ..0 ..396 Winding Life Versus Ambient Temperature.

6 A ..487 Winding Life Versus Bearing Life at 1000C Ambient .. 508 Influence of Altitude on Life..0 ..559 Temperature Rise Characteristics Internally Fan Cooled andNon-Fan Cooled Totally Enclosed Motors Versus Altitude .56A-1 Schematic Plots Depicting Various Amounts of Scatter, asWall as Correlation Coefficient (R) between Variable Xand Response Y .. 88A-2 Correlation Matrix for over 1000 FHP Motor Failures..89A-3 Schematic Data Plot Showing Observed Data Points (Xi, Y )IBest Fit Points (X i Yi), Average x -X IN, Y -YtINError Values (Yi-Yi), and Residuals (Yi-Y) .. 90A-4 Three Criteria for Testing the Relative "Best Fit" of aCombination of Variables in a Hypothetical RegressionModel ..91A-5 Regression Table Description ..92A-6 Regression Table for Bearing Temperature .. 93A-7 Regression T3ble for Ambient Temperature .. 94A-8 Correlation Matrix for over 1000 FHP Motor Failures .. 95A-9 Correlatien Matrix Winding Failures.

7 96B-. Thermal Cycle ..102vi ' , ""7 .._ _-_ _-_,_ _".. _-- - 7-LIST OF TABLESI Life Test Summary Grease 1 ..92 Life Tesg Summary Grease 2 .. ii3 Life Test Summary Grease 3..124 Life Test Sumnmary Winding Failures ..145 Typical Output from 45 Test Populations .. 156 TesL of Bearing Failure Data versus Predictivu Equationsfor .35 Standard Deviation.., ..257 Grease Constant K .. 24a Typical Temperature Rise Data .. *.. 289 Bearing Size Vs. Motor Input Power..#.. 3010 Bearing Life Test Summary Cyclic Temperature Tests ..3111 Thermal Cycling Tests Duty Cycles ..* 3212 Low Temperature Bearing Life.. 3313 Literature Review of Temperature Influence on Life .. 4114 Mntor 1.'4 ding Failiros..4515 Winding Failure Weibull Slope Mean Values .. 4716 (Reference 21)Influence of Altitude on Motor Life for Fan-Cooled Motor4 .5717 FHP Motor life Prediction .. 59A Motor Failure Rate Table .. contractual effort is part of the broad RADC ReliabilityProgram intended to provide reliability predictionl procedures formilitary electronic equipment and systems.

8 These prediction proceduresare contained in MIL-HDBK-217B for which RADC is the preparinq Final model developed in this study will replace the presentelectric motor reliability models that are now in J. GUBBINSP roject EngineerViliAI- q INTRODUCTIONThis report sunmarizes the results of a study directed towards the de-velopment of a failure rate prediction model for fractional horsepower electricmotors. The objective of the study was to completely revise the informationand failure rate models for fractional horsepower electric motors an used inelectronic equipment currently included in Section of MIL-HDBK-217 Bdated 20 September accontplish the program objectives, an existing failure data bank(Reference I) of over 2000 failures of fractional horsepower motors, was up-dated and supplemented from a survey of electric motor manufacturers. Themajor portion of the failure data history was furnished b7 the Rotron M~anu-facturing Company.

9 A regression analysis was conducted on all failuresutilizing such variables as motor power, speed, temperature, failure mode,insulation type and motor type to establ~sh the dominant motor life influencin3parameter.,q and their competing failure modes were found to dominate motor life; ,bearing failures and winding failures. All becrings in the failure bank vereof the greace-packed ball bearing type. Separation of by failure modeindicated that both modes could be describec by a Weibull CommulativeDistribution. A mathematical model for each failure mode dominant varia-bles was developed and a ' rik failure model literature sur' ey of bearing failures and windina failures was con-ducted to verify failure modes and dependent failure variables. The follow-ins sections of this report describe the data failure bank, failure modes,t literature review and development of the mathematica', prediction models.

10 Ap-pendix A describes the rigression analysis, and Appendix B outlines therecommended replacement for Section of ? SUMCARYAi, existing data failure bank consisting of over 2000 actual fractionalhorsepower (FHP) motor failures was upgraded and utilized to develop a failuremodel for FHP motors. Motor failures were dominated by two failure modes; , winding failures and bearing mathematical reliability model was developed for each failure modeutilizing a Weibull cumulrtive distribution function and regression of a literature survey of both failure modes were used to supplementfailure data where required and verify the developed modes were recombined into a competing risk cumulative distri-bution model for use in predicting overall motor life and failure ''' a. a . DISCUSSIONThe development of a "Reliability model " for fractional horsepower(FHP) .-tcrs ncccsiit&es a well-documented data bank of motor failures onwhich to base failure predictions.