This action might not be possible to undo. Are you sure you want to continue?

How to determine the MTBF of gearboxes Dr. Gerhard G. Antony Neugart USA LP Paper for 07 AGMA FTM Introduction In the past decades, “Mean Time Between Failures” (MTBF) has become a very frequently and broadly used characteristic measure of reliability for components, systems and devices used mainly in conjunction with electrical and electronic equipments. From the engineering point of view, assessing the life and reliability of products is a vital part of product design, development, and selection. Life and reliability of a product are also important characteristics for the user (customer) in comparing the widgets (gearboxes) to assess their useful value or life for a certain application. The reliability of a product becomes a frequently used marketing and sales feature. The life characteristics of different products and components depend of a wide range of factors from type and condition of material to type of exposure to loads, magnitude of the loads and other effects such as environment. Products are designed for a certain purpose, function, duty, load etc. The life and reliability is a characteristic statistical value and it can be only approached and assessed by statistical methods. Today an increasing number of manufacturers combine a mechanical device, such as a gearbox, with an electromechanical, such as an electric motor, and electronic drives, logic controllers sensor into a compact integrated “mechatronics” product. The (Mean Time Between Failures) MTBF value of many electronic components and systems is obtainable from the manufacturer. The design life of mechanical components and systems is mainly based on the endurance characteristics of the components on the statistical life expectancy under a certain load such as the L10 design life. Can the reliability of a mechanical device such as a gearbox be expressed in terms of MTBF? How many hours MTBFequals a known L10 life to? The here presented paper is an attempt to find some answers for these and similar MTBF related questions. Life and reliability related issues are based on tests, observation of populations of “widgets” applying mathematical (statistical) evaluations, approximations, using appropriate functions and formulas. It is common practice to define characteristic values, and choose “scientific” statistical methods to analyze and evaluate them. It is fairly easy to define certain characteristics of selected number of test specimens apply statistical methods on an observed population / group and come up with a “scientific statistical conclusion”. However to have realistic, meaningful conclusions the definitions and the evaluation methods must be clear and transparent. This paper is an attempt to express the life and reliability of gearboxes in terms of MTBF without going into an in-depth discussion of statistical methods. As mentioned MTBF is a widely used characteristic value to quantify reliability of electronic components and systems, but it is not commonly used for reliability assessment of mechanical components and systems. For the correct interpretation of the MTBF value it is necessary to understand some basic concepts of probability of failures and the methods of their evaluation. Failure rate The fundamental first step in determining the reliability and life of a widget is to observe a representative set of samples, a “population” of widgets and record the failures over a certain time frame.

this would mean. the actual failure rate should yield to FR = 1400 / 9000 = 0.7% = 1. population of products. However. Example iPod: (.000 iPods owned by more than 4. related to the size of the observed tested number of units. .000 respondents.5% for the observed time interval. it always has to be related to the observed population size.days . An absolute number of failures has no real practical meaning.8% failed for some other reason than the listed. failures / miles (in automotive field). more than 1.miles . it is not an absolute number of units failed but rather a relative number.efunda. . years or within a certain time related measure such as distance i.e.2 The collected data will show a certain number of failures over an observed time period. under headline “iPods built to last 4 years” ) “Apple spokeswoman Natalie Kerris recently told the Chicago Tribune that iPods have a failure rate of less then 5%.source: “AppleInside” July 27 2006 reporting. we assume that the failure rate is constant during this period. This is expressed by the failure rate. it assumes a value between 0 (0% failures) and 1 (100% failures per hour).400 of the players had failed.e. a survey conducted by “Macintouch” last year found that out of nearly 9. hours.com/designstandards/bearings/bearings_linear_ld_life. The survey concluded that the failure rate was 13. which can be interpreted that 15.large population of units in field Lab test ? Field test ? How far are the loads and op. stemming from an equal mix of hard drive and battery related issues” Remark: Based on the numbers. because the value is related to the overall observed population.155 or 15. This relative failure rate can be recorded at regular time intervals (important in where we want find out if and how the failure rate is changing over the life time) or just record it for a predefined period such as the “design life” the widget. Failure rate FR Number of Failures (NF) FR = ----------------------------------------------------Observation Time (OT) x Population Size (N) NF. at a certain time / time interval Failures: .7 %.conditions consistent ? Fig 1 Source: :www. or per operating cycles such as failures in 1 million revolutions (bearings) etc.Observation Time (till a certain amount or all failed) hours . failures / minutes.which type of failures ? failure mode ? which part ? repairable non unrepairable ? loading conditions environmental effects consistent ? OT. Failure rate is the relative frequency at which a component or system fails in a given time i.umber of Failures (during the overall time period.55% – 13. As we can see.cycles N.. Other explanations are also possible however the survey does not list any reasons.Population Size – overall number of units observed selected test units .cfm It would be more exact to call it the “relative failure rate”. which she said is "fairly low" compared with other consumer electronics.

failure 1% (200000 x 2 x 0. On the other hand. return. Failure rate values result mainly from “field tests” i. and life characteristics. Bathtub curve .01) = 4000 Estimated average operating hours 8 hours /day 4000 warranty returns FR = ------------------------------------------------------------------------------------------------------------------------400000 (unit shipped in 2 years) x 2 (years warranty) x 365 days x 8 hrs FR = 0. ISO. and far less controlled conditions.000001712 / hour Electronic components and system such as a simple LED or complex processor chips are used in millions of computers or other devices under exactly defined and controlled conditions (certain voltage and clock frequency rate. Example K-gearbox: The right angle bevel-helical K-boxes have 2 years warranty 200000 K boxes are shipped / year Warranty repair. It is economically impractical to life test a large population of gearboxes or other mechanical components.e. to find the mortality rate. neither about the conditions under which the data were collected.) On the other hand. and environments. The shafts gears fasteners etc are all based on the endurance limit hence there is theoretically no life limitation under the nominal load. temperature etc. Electronic components are routinely lab tested in high volumes. Kerris from Apple correct with the 5% value over the design life or are the conclusions of the survey with 13. loads. Many components such as the vital bearings have well known statistically proven reliability.3 Is Mrs.) What was the real usage time or reference time of both observations? Does the statement “built to last 4 years” means 4 years 24/7 usage.7% correct? Here are some important questions before we take sides: Are both talking about the same type of failures? Does Apple consider the necessity to replace the battery a failure? How many units were surveyed by Apple? Is the population of 9000 samples in the survey representative? (Apple ships in excess of 10 million units a quarter.. the observed time interval and failure mode etc. the failure rate. securing exact same load conditions etc. or for instance only 4 hours a day usage say 6 days a week (duty cycle) ? None of the above two failure rates gives any specific information about these important basic factors. statically proven methods frequently regulated by standards such as AGMA. gearboxes are subjected to a wide range.000017757 / hour 9000 units x (2 years 365 days x 12 hrs) The example above highlights the main difficulty of the reliability /life calculation and the main source for misrepresentations namely the selection of the population. etc. Also the population size is substantially higher for electronic components than for gearboxes. gearboxes are designed based on well-established. should be calculated as: 1400 failures FR = --------------------------------------------------------------------------------.= 0. observations in real world applications. DIN. Let us assume that the survey was based on a 12 hour daily usage over a 2 year period. This issue will be discussed further in relation to the two proposed methods of gauging gearbox MTBF.

here the probability of a certain failure rate. the “infant mortality” period with decreasing failure rates. hence the name “bathtub curve”.0 Increasing failure rate Decreasing failure rate “Infant mortality” Low constant failure rate normal useful life (design life) 0. in most practical applications. Electronic components have a very distinctive infant mortality. to minimize this impact on the reliability in practical applications. is mathematically described. approximated and analyzed by defining a suitable Probability Density Function (PDF. “wear out” . The area under the PDF i. however. For obvious reasons. not the reliability during the infant mortality period or during the period exceeding the design life namely the wear out period. (shown below). such as humans also resembles the bathtub curve. It has 3 characteristic areas. Probability density function PDF and Cumulative distribution function CDF The probability of an occurrence. the integral of the PDF is called the Cumulative Distribution Function (CDF).e. Mechanical components.0 Failure Rate Time (hours. miles cycles …) Fig 2 The plot of the failure rate over time for most engineered components and systems resembles the form of a bathtub. such as gears and gearbox-components behave differently. the wear out can be significant. On the other hand. followed by a almost flat. the wear out of solid state electronic components is far less significant.4 Most components follow the characteristic plot of failure rate over time. The most common and well known PDF is the normal probability distribution (Gauss distribution) applicable to many natural phenomena (Figure 3). The failure rate of living creatures. There is no significant infant mortality. electronic components are frequently subjected to a “burn in” which separates the early failures from the population. the useful life period is of interest. nearly constant failure rate period frequently called the “useful life period” and finally the third part with increasing failure rate the “wear out”. Bathtub Curve 1.

wikipedia.exp(-(t/η)β) This distribution function with the parameters slope or shape parameter β and parameter characteristic life or shape parameter η is very flexible for modeling failure behavior such as the bathtub curve. Mr.com However the Gauss normal distribution function is not applicable to “the bathtub curve type” distributions. This inherent shape makes it of limited usability for failure distribution modeling. Waloddi Weibull . Weibull probability distribution function Weibull PDF Weibull CDF f(t) = (β/η) (t/η)β-1 exp(-(t/η)β ) Note: exp(x) = ex F(t) = 1.5 Gauss normal probability distribution Function PDF CDF The PDF is the derivative of CDF (the CDF is the integral of PDF) The Gauss normal distribution function is symmetric in relation to the mean (median) μ with a variance of σ. Fig 4 Source: www.html Whereas the normal probability distribution function has the same basic shape for all parameters. Fig 3 Source: :www.com/hotwire/issue14/relbasics14.weibull. Weibull is well known to us “gearpeople” familiar with the bearing design and associated ”B-life” ratings. the Weibull 3 or 2 parameter distribution function allows to ”model”(describe) widely different shapes of PDF’s depending upon the shape parameter β.

6321 line” can be interpreted as the mean time between failures.weibull. since F(t) @ t= η 1-1/e = 0. decreasing failure rate of the bathtub curve.. R(t) is “reliability”. Furthermore at a time t = η..html In the Weibull plot the horizontal line at 0.e. corresponds to beta values <1 .6231 failure rate has a special meaning. Also Note F(t) = 63.the infant mortality.21. corresponds to beta =1 .6 suggested bearings should be compared at a life corresponding to 10% failure probability.21% of the population will fail independent from the β value.the useful life period.78. constant failure rate – useful life region which is the scope of most practical considerations.e-(t/η) β or R (t) = e-(t/η) β F(t) the Weibull cumulative distribution function CDF (here the widely used 2 parameter distribution) provides the probability of failure.2% Fig 5 Source: www. the B10 (L10) life. the t value corresponding to the intersection point of the F(t) line and the horizontal “ 0. have the shape of a simple line where the slope is equal to the parameter β.% survival probability . % failure probability means R(t) = 36. 63. η – characteristic life. the distribution functions.6321.or scale factor β – shape parameter or slope e – Euler’s number or Napier’s constant (the base for natural logarithms) For the 3 characteristic areas of a bath tub curve . the complement of F(t) where: t – failure time. In the Weibull probability plot which is using an adjusted logarithmic scale. For failure probability distributions with β =1. Note this is only correct when β =1 i. increasing failure rate. Weibull plot 63. constant failure rate. F(t) = 1. corresponds to beta values >1.com/hotwire/issue14/relbasics14.the wear out.

umber of Failures (during the overall time period..7 Men Time Between Failures MTBF It should be emphasized that in all practical applications of widgets the reliability during the “useful life” (also called the design life) is what matters..Population Size – overall number of units observed selected test units . just compare it to the definition of Failure Rate.days . return.. This period is characterized by β =1 in the Weibull distribution.miles ...which type of failures ? failure mode ? which part ? repairable non unrepairable ? loading conditions environmental effects consistent ? OT. Example iPod: the Apple iPod example MTBF = 9000 x 2 years x 365 days x 12 hours / 1400 failures = 56 314 hrs Obviously we cannot expect that an iPod will last 56314 hrs equivalent to over 6 years uninterrupted operation. at a certain time / time interval Failures: .com/designstandards/bearings/bearings_linear_ld_life. op. The basic definition of Mean Time Between Failures is simple and logical.efunda..conditions consistent ?.Lab test ? Field test ? … are the loads and op. The MTBF is the actually reciprocal value of the Failure Rate (FR) MTBF = 1/ FR MTBF Observation Time (OT) x Population size N MTBF = ------------------------------------------------------------Number of Failures (NF) NF.Observation Time (till a certain amount or all failed) hours .e.. 200000 unit x 2 years x 1% = 4000 units Estimated average operating hours: 8 hours /day MTBF = 200000 / year shipped x 2 years x 2 years warranty x 365 days x 8 hrs / 4000 warranty MTBF = 584 000 hrs .- Fig 6 Source: :www.large population of units in field - MTBF = 1 / FR …. failure 1% i.cfm Lets calculate the MTBF for the 2 examples presented above. Example K-Gearbox: Right angle helical bevel K-boxes have 2 years warranty 200000 K boxes are shipped / year Warranty repair.cycles N.

comparable and representative MTBF value. the populating size. The main load carrying (transferring) components of a gearbox are the gears. rated load conditions unlimited load cycles.e. the life and reliability of mechanical system such as the gearbox depends on the life/reliability characteristics of its components at a certain defined load (called design load. However in the real world applications these components do fail but mainly because overloads occur i. are not participating direct in the torque transfer. they are “statistical”.e. shafts. In case the number. This fact makes it possible to approximate the life / reliability equivalent of a bearings in terms of MTBF. surface-. Therefore their influence on the gearbox life is practically impossible to quantify just from the design data. if they are loaded beyond the design specifications. Since they are based on endurance limit. These components should be selected and shaped to endure under the “nominal” i. fasteners etc. life. The stresses under the nominal load. For this reason the design standards include a number of sizing factors (size-. “secondary parts and accessories” such as seals. The endurance limit values itself are not “exact” .factor …etc) to adjust the endurance limit “to be on the safe side”. The main goal would be to determine an MTBF values based on the design parameters. . Gearbox MTBF determination Obviously. (theoretically unlimited life) it is permissible to suggest that components designed based on endurance limit will not influence the MTBF. Their selection / design is based on standardized calculations rooted in statistical evaluations / methods.8 Here again the expectation that a K-box will last about 67 years under continuous operation would be a false interpretation of the MTBF value. Gears. Other. Bearings The rolling element bearings. shaft / hub connecting devices and bearings. since products are designed and made for certain nominal loading (usage) conditions. consistency of loads and real operation time all influence the MTBF. using the reliability characteristics of the components. by using calculation methods such as the Palmgren-Miner linear damage hypothesis. rated load nominal load). shafts and hub / shaft connections are generally designed based on the endurance (fatigue characteristics) design standards. as they are not selected based on some endurance limit and their life is inherently limited. the observed time frame. duration an magnitude of the load cycles above the nominal load are known. Since testing a large number of gearboxes to have statistically well backed up life / reliability values is economically not practical especially for custom gearboxes build in low volume or even as single piece. The expense of running lab tests on hundreds of gearboxes for the period of their design life is not even justified in high volume products such as automotive transmission. the other main component of a gearbox has a different life characteristic. it is possible to estimate (approximate) the life. Gears and Shafts Remember. for instance the bending stress at the tooth root must be below the endurance limit. If the loads are above the nominal value. (even if only occasionally) the life of these parts is limited. The above examples calculated the MTBF based on field survey data using a number of assumptions. As pointed out in the paragraph about Failure Rate. the MTBF generally is referenced to these “normal” conditions. Ideally we should run lab-tests on a large population of products ensuring same conditions to have an objective. It is economically not feasible to carry out extensive lab tests on products like industrial gearboxes. Comparing the 2 values we can certainly say the K-box is about 10 times more reliable then an iPod.

78 % reliability (survival probability). comparable MTBF values it would be beneficial to develop certain guidelines. standards for the collection of the above mentioned data.00 x L10 Fig 7 Source: :www. Selection of bearings is based on this value.e.cfm .e 36. Discussing the Weibull plot at β =1 we concluded that the MTBF value corresponds in to 63. Key mechanical components of countless mechanical systems are the rolling bearings. The B10 (L10) life of bearings is well defined. means 10% failure probability or i. Proposed Alternative 1 Gearbox MTBF determination – based on warranty / repair figures The calculated MTBF value of gearboxes based on: a) Observation time equal to the warranty time b) Population equal to average amount shipped during the observation time c) Number of warranty returns or % of the warranty returns as number of failures is a valid approach to determine the MTBF.based on L10 life As discussed above Weibull distribution function at β =1 the η value corresponds to the MTBF. Ln values (L1 to L50) for bearings are listed in terms of the L10 in a number of engineering literatures such as as Ln = fr x L10.9 Based on the above discussed two alternatives are suggested here for the determination of the MTBF of a gearbox. ( see example K-box above ) To have a honest.21 % failure probability i.com/designstandards/bearings/bearings_linear_ld_life. Most manufacturers have these or similar values many of them collected by their quality management system such as ISO 9000.e.efunda. its complement 90% reliability (survival probability). This is based on many years of tests and field data. i. Say a gearbox has bearings designed / rated for a 100 000 hrs L10 life. Proposed Alternative 2 Gearbox MTBF determination . Relationship between Bearing Ln and L10 life Example: L50 = 5.

e. it can be certainly used to describe reliability of mechanical components and systems if the basic rules are followed and interpreted correctly. Whereas the first suggested method.5 = 255 000 hrs Conclusions / Suggestions MTBF a frequently used value to quantify reliability of electronic component and systems. However extrapolating graphical curves Ln = f (L10) indicate that the “fr” value at 63. no explicit “L 63. Using predictive maintenance techniques bearings can be kept in operation far longer than the designed L10 life. when listing an MTBF value it should be noted which approach was used. The proposed two alternatives determine the MTBF of a gearbox using data which in many cases are readily available to the gearbox manufacturer and designer.21” value could found. The above approximation of the overall gearbox MTBF based on the L10 value is rather conservative.5 Therefore we conclude. we have to remember that in many gearboxes the bearings are considered as wear parts which can and should be periodically replaced. Predictive maintenance can also tell to replace a bearing even though before its L10 life.21% reliability is around 8. On the other hand some gear types such as the epicyclical (planetary gears) the bearings are direct involved in the torque transmission such as the needle bearings of the planet wheels. to make the used data consistent. In many gearboxes the bearings are not members which are actively involved in the torque transmission.10 The literature is lists values up to L50. based on warranty figures i. .5 That said. The MTBF value determined method by 1 and method 2 for the same gearbox will significantly differ in most cases. Example PLE Planetary gear head: The needle bearings of a PLE planetary gearbox are designed for 30 000 hrs L10 life at rated torque. but still having a very vital function of supporting the torque transmitting components.. and hence the MTBF values of different gearboxes comparable. The MTBF of the PLE gear head can be calculated as: MTBF = 30 000 x 8. so avoiding consequential damages to the gears and / or other components. The bearing base method would be only recommended if field test based values are not available. the MTBF is be equal to MTBF = L10 x 8. Therefore. The gears are designed based on the endurance limit at rated torque. recommendations or standards. that systems (such as gearboxes) where the rated life is mainly based on the bearing L10 value. In planetary gears the planet gear bearing is a vital part in the torque transmission subjected to loads proportional to the transmitted torque. de facto field tests give a more balanced more complete realistic reliability assessment than the suggested second alternative base on the Bearing L10 life. It would be beneficial to develop appropriate AGMA guidelines.

& Winter H. AGMA 2001 ISO 6336.Machine Elements Springer Verlag www. .“Reliability and MTBF Overview” .com ANSI / AGMA 2101.Zahnraeder Berlin Springer Verlag Nieman G. NY Dudley D. . .Machine Design McGraw-Hill Book Company Inc.weibull. DIN 3990 . Wilkins . VICOR Reliability Engineering Shigley J.Reliability Engineering and Weibull analysis resources wikipedia. .com . W.11 References Abernathy Robert B.efunda. E. .Gear Handbook McGraw-Hill Book Company Inc NY Dudley D.W.“The New Weibull Handbook” Reliability Basics Reliability HotWire. & Winter H.The Magazine for the reliability Professionals Issue 22. . April 2002 Dennis J.com .“Bathtub Curve an Product Failure Behavior “ Reliability HotWire .Engineering Fundamentals www. Dr. December 2002 Speaks Scott .The Magazine for the reliability Professionals Issue 14.

- Antony How to Determine the MTBF of Gearboxes
- Reliability Engineering
- Reliability Models
- Reliability Presentation
- ASA 96 Trends in Repairable System Reliability
- Reliablity Assignment
- Calculating Reliability
- Mining Equipment Maintenance
- 15 Reliability 1 - Student
- Reliability
- 4 the Assessment and Control of Product Reliability
- Machinery Selection-Books23
- Review Rel Meas
- Rr410211 Reliabilty Engineering and Application to Power Systems
- DFR - Presentation
- 36
- ENGG ZC242-L7
- Cost Optimized Reliability Test Planning Rev 11
- Availability
- Bayesian Updating of Reliability of Civil Infrastructure
- QRMEC-Elsayed May 28 2013
- Basic Reliability Concepts and Analysis Chapter 2
- DCDS2009_Medina_Weber_Simon_Iung Review HAL
- Reliability
- reliabilityengineering-chapter-2reliabilityofsystems-141007073755-conversion-gate02.pptx
- QM ZG528-L6.pptx
- NR 420204 Power System Reliability
- 14_14-0181.PDF
- Limit State Handout[1]
- TQM_BBA_8_WK_10
- MTBF

Are you sure?

This action might not be possible to undo. Are you sure you want to continue?

We've moved you to where you read on your other device.

Get the full title to continue

Get the full title to continue reading from where you left off, or restart the preview.

scribd