You are on page 1of 32
AN AMERICAN NATIONAL STANDARD Design of Transmission Shafting ANSI/ASME B106.1M - 1985 (SECOND PRINTING) SPONSORED AND PUBLISHED BY THE AMERICAN SOCIETY OF MECHANICAL ENGINEERS United Engineering Center 345 East 47th Street New York, N. Y¥. 10017 ‘Second printing ~ March 1988. Incorporates the following editoril chonges Pose Location Change 1 2 Definition of Sy revised 7 abies Note (2) and reference added 8 Fig. 4 Acknowledgmont correted 8 Fig. (1 Caption revised (2) Ganoral Note deleted 16-19 Table 1 vised in ite entirety a Table C1 “Third column heading and fist entry revised Dato of Issuance: August 30, 1985 This Standard will be revised when the Society approves the issuance of a new edition. ‘There will be no addenda or written interpretations of the requirements of this Standard issued to this Edition, ‘his code or standard was developed under procedures accredited as mecting the criteria for ‘American National Standards. The Consensus Committee that approved the code or standard was balanced to assure that individuals from competent and concerned interests have had an oppor ‘tunity to participate. The proposed code or standard was made avaliable for public review and ‘comment which provides an opportunity for additional public input from industry, academia, reg Ulatory agencies, and the publo-et-arge. [ASME doss not “approve,” “rat ‘ASME does not take any position with respect to the validity of any patent rights asserted in ‘connection with any items mentioned in this document, and does not undertake to insure anyone sndard against lablty for infringement of any applicable Letters Patent, nor assume y. Users of a code or standard are expressly advised that determination of the validity of any such patent rights, and the risk of infringement of such righte, le entirely thelr own responsibilty, Participation by federal agoncy represontatvels) or personts) affliated with industry Ie not to be interpreted as government or industry endorsement of this code or standard ‘ASME accepts responsiblity for ony those interpretations issued in accordance with governing ASME procedures and policies which preclude the issuance of interpretations by individual vo: sndorse"” any tem, construction, proprietary device, oF No part of this document may be reproduced in any form, in an electron retrieval system of otherwise, without the prior writin permission ofthe publisher. Copyright © 1986 by ‘THE AMERICAN SOCIETY OF MECHANICAL ENGINEERS ‘Al Rights Reserved Printed in U.S.A, FOREWORD (This Foreword is not pert of ANSVASME B106.1M-1985.) Since 1954, when the Code for Design of Transmission Shafting (ASA-B17C-1927) was with- Grawn, theze has been a need for a design method which recognizes advances in shaft design technology. ASA-BITC provided a design method covering a wide range of machine shafting applications, It was based on the static yield strength of the shaft material under combined bending and tor- sional shaft load, and made allowance for service conditions. It was found to be overly conser- vative in some cases; in others, it was believed to be incomplete. It is now well accepted that ‘most shaft failures are caused by progressive crack propagation resulting from fluctuating load, ‘commonly called “fatigue failure.” A shaft design method based on the fatigue strength of the shaft is needed. The purpose of this Standard is to present such a method for the most common shaft loading condition of combined reversed-bending and steady torsion. ‘American National Standards Committee B106 held its organizational meeting on May 23, 1968. On October 6, 1981, the Committee was reorganized as ASME Standards Committee B106 on Design of Transmission Shafting ‘Suggestions for the improvement of this Standard will be welcomed. They should be sent to ‘The American Society of Mechanical Engineers, United Engineering Center, 345 East 47th Street, New York, New York 10017. ‘This Standard was approved by the American National Standards Institute on May 17, 1985. ASME STANDARDS COMMITTEE B106 Design of Transmission Shafting {The following isthe rostor of the Committee st the une of approval of thie Standard.) OFFICERS V.R. Lal, Chairman S.H. Loewenthsl, Vice Chairman Westy, Secretary COMMITTEE PERSONNEL AMERICAN GEAR MANUFACTURERS ASSOCIATION RC. Brown, American Gear Manufacturer Astocation, Avington, Virginia AMERICAN SOCIETY OF MECHANICAL ENGINEERS, THE ‘A.L.Hltcheox, Penton/IPC, Cleveland, Ohio MECHANICAL POWER TRANSMISSION ASSOCIATION C.L. Vroom, Ven Gorp Corp, Pela, ome NATIONAL AERONAUTICS AND SPACE ADMINISTRATION S.H. Loewenthal, NASA Lewis Research Center, Cleveland, Ohio V.R. Lali, NASA Lewis Research Center, Cleveland, Ohio INDIVIDUAL MEMBERS . A. Hughes, Production Engineoring Magazine, Clevolond, Ohio D-H. Wright, RJ. Dick Ine, Muscatine, lowe CONTENTS La pogbanGadone Standards Committee Roster Scope 6.2... Nomenclature , Design Formulas... || Fatigue Modifying Factors Factor of Safety (FS) Other Considerations - Figuros 1 Surface Finish Factor k, as a Function of @ Surface Condition and Tensile Strength (From Several Sources) 2 SieFactor ...., ae SqOHORpnOHE Recommended Size Effect Factors for Unnotched Steel Shafts (From Several Sources) . Bagnbeoosn 4 Notch Sensitivity g o.oo TT) 5 Round Shaft With Shoulder Fillet in Bending 6 Grooved Round Bar in Bending .......- 7 Round Shaft in Bending With a Transverse Hole Tables 1 Reliability Factor eee 2 Temperature Factor kg — Fatigue Properties a2 Related io Room ‘Temperature Properties [70°F (23°C)}. ac 3 Fatigue Stress Concentration Factor ky — Typical Values for Keyways in Solid Round Steel Shafts Appendices ‘A’ Theoretical Background B Materials and Properties . C Sample Problem . . Figures AI Combined Stress Fatigue Test Data for Reversed Bending in iComtavation With Static Torsca| esas enieneee a A2 Combined Stress Fatigue Test Data for Reversed Bending in ‘Combination With Reversed Torsion... .- 3. Comparison of Combined Stress and Separate Stress Design Methods. C1 Load Diagram... , C2. Sample Problem _ vil ul 15 a 2 2 B 23 23 Tables 6 B1 _ Representative Mechanical Properties of Shafting Steels . . .. . . 16 C1 Strength Properties of UNS G-10450 Cold-Drawn Steel 2 References 25 ANSI/ASME 8106.1N-1985 AN AMERICAN NATIONAL STANDARD [AN AMERICAN NATIONAL STANDARD DESIGN OF TRANSMISSION SHAFTING 1 SCoPE ‘This Standard takes into account many modern theo- ries and has been substantiated, in part, by laboratory testing. ‘This Standard is not intended as a textbook. It is writ- ten for use by those skilled in the art of shaft design and stress calculations. This Standard provides a common basis for discussion and understanding between persons involved in shaft design ‘A design procedure is presented for computing the diameter of either hollow or solid rotating steel shafts under combined cyclic bending and steady torsional loading for unlimited life. The method is based on an elliptical variation of fatigue strength with increasing torque as exhibited by combined stress fatigue data. The corrected reversed-bending fatigue limit of the shaft to be designed is used in the recommended design formula, It is calculated from the fatigue limit of the ma- terial from rotating beam specimen data, such as that listed in Appendix B. Fatigue modifying factors are cited to correct the polished, unnotched specimen data for a umber of service factors. These service factors include surface condition, size, reliability, temperature, duty cycle, fatigue stress, concentration, and miscellaneous effects. Although the present design procedure ensures that the shaft is properly sized to provide adequate service life, it is not the only shaft design consideration. A shaft must also be stiff enough to limit deflections of key power transfer elements such as gears and pulleys, and also have sufficient stiffness to minimize misalignment through bearings and seals. Furthermore, the shaft must have suf- ficient radial and torsional stiffness or lack of stiffness to avoid unwanted vibrations at the operating speed. These factors have not been included in this Standard although checking them is good design practice. Calculation pro- cedures for these factors may be found in most machine design textbooks. An example of how to calculate shaft deflection and critical speed is given in Appendix C. 2 NOMENCLATURE shaft diameter, in. (mm) inside shaft diameter, in. (mm) FS = factor of safety K = hollowness ratio (afd) Ky = fatigue strength reduction factor ky = sie factor reliability factor temperature factor duty cycle factor y= fatigue stress concentration factor = miscellaneous effects factor ‘M = reversed-bending moment, Ibn, (N +m) .N-= frst lateral critical speed of shat system, xpm ‘4 fatigue notch sensitivity factor ‘Sp = reversed-bending stres, psi (N/m?) @2M/na) '5;= corrected fatigue (endurance) limit of shaft in reversed bending, psi (N/m?) ‘Sf = fatigue (endurance) limit of polished, unnotched ‘test specimen in reversed bending, psi (N/m*) Spa Allowable corrected fatigue (endurance) limit of shaft in reversed bending, psi (N/m?) (SylFS) ‘S}, = fatigue (endurance) limit of polished, unnotched test specimen in reversed bending with steady torque, pi (N/m?) static-orsional stress, psi (N/m?) (16T}nd*) Spy = reversedtorsional stress psi (Nim?) Surg fatigue (endurance) limit of polished, unnotched test specimens in reversed torsion, psi (N/m?) Syy = torsional yield strength, psi (Nim?), (S,y ~ SyhV/3 for most stels) Szyq = allowable torsional yield strength, psi (N/m?) = GyyIFS) ‘5, = ultimate tensile strength, psi (N/m?) 'S, = tensile yield strength, psi (N/m?) T= mean stati torque, bin, (N »m) Se 3 DESIGN FORMULAS ‘The shaft diameter formulas appearing in this Section ‘were theoretically developed in combination with experi- ANSVASME 8108.1M-1985, ‘AN AMERICAN NATIONAL STANDARD ‘mental data, The theoretical background of these deriva- tions appears in Appendix A. The variables to be used in these formulas are described in Sections 4, 5, and 6. 8.1. Solid Shafting a= VG2FSi) JOISGF RTS, (1) ‘where the corrected shaft endurance limit is Sp KakykckakekphyS} ‘The k factors in the above formula are the fatigue modi- fying factors discussed in Section 4. Sf is the uncorrected fatigue limit of the shaft mate- tial obtained from rotating beam specimens. To obtain ‘S}.,see Appendix B, 3.2 Hollow Shafting LAITY) GRR |a) VMI TTS) 2 where the shaft hollowness ratio is Keaya NOTE: The applicability of Bq. (2) to the desien of thin-val shafts of tubes (K > 0.9) has not yet been established 4 FATIGUE MODIFYING FACTORS In shaft design Eqs. (1) and (2), the fatigue limit of the shaft Sy differs from the fatigue limit of the highly polished, notch-re, rotating beam test specimen $7. Before a shaft size can be determined, the uncorrected fatigue limit of the test specimen, commonly listed in design tables, must be modified to account for diffe. ences between the shaft tobe designed and the test speci men. These differences include surface finish, size, rll bility, temperature, stress concentration, and other mis- cellaneous factors. 4.1 Surface Finish Factor Ky ‘This factor accounts for the difference in surface con: dition between the shaft to be designed and the highly polished (mirror finish) test specimen. Experiments have shown that surface condition can have a considerable effect on fatigue strength, since fatigue cracks are gen- DESIGN OF TRANSMISSION SHAFTING erally initiated at the surface of the shaft, where the stresses are highest [1]. Surface treatment affects fatigue strength in several ways. Forging, surface rolling, and other work-hardening processes tend to set up beneficial compressive residual stress in the surface layer. However, the surface decar- burization that often accompanies forging can cause a severe reduction in fatigue strength. Fatigue strength also can be reduced by surface roughines, which intro- duces surface irregularities that act a8 local stress raisers. ‘The surface finish factor ky shown in Fig, 1 [1] eor- rects the fatigue limit ofthe test specimen, Tt is based on 2 compilation of test data from several investigations for a variety of ferrous metals and alloys compiled in [1] Based on [1], the ground surface category includes all types of surface Finishing which does not affect the fatigue limit by more than 10%. Polished, ground, hhoned, lapped, or supersfnished shafts are included in this ground category as well as commercial shafts that axe turned, ground, and polished, or turned and polished. The machined surface category includes shafts that are either rough oF finished machined, or unfinished cold- drawn shafts with roughness ranging between 62 yin, and 250 pin. (1.6 umn and 6.3 um). The hot rolled cate- sory covers surface conditions encountered on hot rolled shafts which have slight irregularities; some include oxide and scale defects with partial surface decarburiza- tion [1]. The asforged category includes shafts with large surface irregularities, included oxide, and scale de- fects, with total surface deeazburization, Figure I shows the endurance characteristics of higher tensile strength steels to be more adversely affected by poorer surface conditions. 42 Size Factor ky Experience and testing have shown that the fatigue limit generally decreases as shaft size increases. It is be- lieved that this size effect is related to the preater likeli hhood of encountering « potential fatigue-intisting defect. with 2 lagershaft. Larger shaftshave a lower stress gradi- cent than smaller shafts and thus have a greater volume of ‘matorial under higher stress. Another factor is that the hhoat treatment of large perts may produce # metal- lurgical structure that is not as uniform nor having as fine a grain structure as that obtained with smaller pars ‘The results of reversed bending tests on unnotched, polished steel specimens of up to in, (50 mm) in diam- eter are shown in Fig. 2. These data have been compiled Numbers in brackets correspond to the referenced documents in the References section following the Appendices. 6 ANSITASME 8106.16-1985 DESIGN OF TRANSMISSION SHAFTING AN AMERICAN NATIONAL STANDARD Mirror polished ~ Cd Fine ground A, 7 = aL en ee ~N ~~ Surface Factor k oa o2[e Astoreed 0406 08S BAO Ultimate Tensile Strength Sy, MPa © 109 140 180 200 260 Utimete Tensile Strongth, Sy ksi FIG.1 SURFACE FINISH FACTOR k, AS A FUNCTION OF A SURFACE CONDITION AND TENSILE STRENGTH (FROM SEVERAL SOURCES) ("sprodvesd from Juvinal, RC, Englneoring Consideration of Stress, Stain and Strength, MeGrow: Hill Book Compsny, 1967, by permission) ANSU/ASME 8106.1M-1985 AN AMERICAN NATIONAL STANDARD 12 1 Size Foctor ky, oa DESIGN OF TRANSMISSION SHAFTING ‘Specimon Dismetor, nm ee ° 04 08 12 16 20 ‘Specimen Diameter, in GENERAL NOTE: Data for unnoxched, polished ste! specimens having 50 ks 10165 ksi were compiled from several sources in [2] FIG,2 SIZE FACTOR (Reproduced from Juvinal, A. C., Enginesring Consideration of Stress, Strain and Strength, MeGrave from many investigations [2]. Included in Fig. 2 is an analytical expression based on the concept of Kuguel [3]. Kuguel hypothesizes that a decrease in fatigue strength can be mathematically related to an increase in the volume of material subjected to at least 95% of the maximum stress. ‘The Kuguel? expression can be written Gch units) (millimeter units) ky (aj0.3)"9955 gy = (ay7.6y 0088 where = the diameter of the shaft, in. (mm) Although there are several other sourced for deter- ‘mining this size effect factor [3-9], there are only minor differences between these other sources and the Kuguel expression for shafts up to 2 in. (50 mm) in diameter. Excerpted from Kugel, R,, “A Relation Between Theoretical Suress Concentration Factor and Fatigue Notch Factor Deduced From the Concept of Highly Stressed Volume,” in Proceedings ‘of ASTM, 1969, with permission, from ASTM. Copyright, ASTM, O16 Rave Street, Philadelphia, PA 19103. Because the data fit reasonably well, the Kuguel expres- sion is adopted here. ‘However, for shafts larger than 2 in. (50 mm) in diam= eter, very few test data exist (10, 11]. There are insuffi- cient data for establishing any definitive formula for these larger shaft sizes. The few relevant tests conducted indi- cate a considerable decrease in fatigue limit for shafts having very large diameters [6]. A collection of recom- ‘mended size factors for larger shafts from several special- ists (2, 6, 7, 9] is shown in Fig. 3. Also shown in Fig. 3 is an arbitrary composite expression, inch units) (altimeter units) Kye dt? y= 1.g5d7829 Which is basically a compromise of these other suggested approaches. Fatigue test data from [11] for 6 in, and 8.5 in (150 mm and 213 mm) diameter, plain carbon (0.4% to 0.5%) steel specimens in rotating bending have also been included for comparison. The range of fatigue limit values found in these tests [11] are denoted by the length of the arrows. In the absence of actual data for the shaft to be designed, this composite expression will Provide an estimate of the size effect for shafts larger a DESIGN OF TRANSMISSION SHAFTING. 10 7 Pet (3) 7 Pet (6) 09 ~~ < 08 Bol \ : 5 os Leet 1) roe ANSU/ASME 8106.16-1985, [AN AMERICAN NATIONAL STANDARD 7 Composite expression ki Pot. 17) OP el 2) “pF ae / Rot. (11) Test ava Ret. [11] ol tof | oj | | | | ow | o om 2 3S 4 5 6 7 8 9 1 Shaft Diameter in, L ° 50 100 150 "200 250 ‘Shaft Diameter dam FIG.3 RECOMMENDED SIZE EFFECT FACTORS FOR UNNOTCHED STEEL SHAFTS (FROM SEVERAL SOURCES) than 2 in, (50 mm) and less than 10 in. (254 mm) in diameter. 43 Reliability Factor ke Even under well-controlled test conditions, it is clear that the unavoidable variability in the preparation of test specimens and in their metallurgical structures will ‘cause variations or scatter in their measured endurance strengths. Fatigue limit data published in standard design references usually represent some mean value of endur- ance for the sample of test specimens. Most designs re- ire a much higher survival rate than 50%, that is, the probability that at lest half the population will ot fil in service. Consequently, the fatigue limit must be re- duced to increase reliability. A strength and a stress distribution can be associated, with each part. In the absence of test data, a good rule is to assume a Gaussian failure distribution with a standard deviation of 8% of the mean fatigue limit strength for determining k,. These values are given in Table 1 Values of Kk, for reliability levels 0.99 and above are ‘quite sensitive to the failure distribution assumed, In this range, ke values listed in Table 1 are not as accurate and. should be used only as a guide. 4A Temperature Factor kg Extreme operating temperatures affect the fatigue limit of steels. However, between normal operating tem- peratures ranging from about ~70°F (-57°C) to 400°F (204°C), the fatigue strength characteristics of most steels are essentially unchanged. For this temperature range, a temperature factor kg = 1 is recommended. For applications outside this temperature range, the fatigue properties at the appropriate temperature for the shaft material in question should be ascertained from actual test data, either published or user-generated, and compared with room temperature properties to determine appropriate Kg. Table 2, which lists variation in k with temperature change for several steel compositions, based on a number of investigations, should provide some suidance 4 Duty Cyele Factor ke Shafts are seldom exposed to constant amplitude load- ing in service, The designer must consider start-stop cy- cles, transient overloads, vibrational or shock loading, and changes in the load spectrum of the equipment driven, by the shaft. An important question is: how much fatigue ANSW/ASME 8106.1M-1985 [AN AMERICAN NATIONAL STANDARD DESIGN OF TRANSMISSION SHAFTING TABLE 1 RELIABILITY FACTOR ‘Shaft ke 10 0.897 oat TABLE 2 TEMPERATURE FACTOR kg ~ FATIGUE PROPERTIES AS RELATED TO ROOM TEMPERATURE PROPERTIES [70°F (23°C)] ‘Temperature, °F ("C) UNS Number [Note(1D] -200 100 «0 #70—= 200 «400-==««8DS8D=S«000 sand Condition (12) 73) 18) 23) (83) (204) 16) 427) 538). (6-10350 17 1 1t) 1OAC G-10900 15 120049 19D tna 6-43400, 13 111900199 G-43400 {notched 19 09 09st 0.7% carbon ol 0 10 tas zs on is) 6.43400 : 10 : rr (Carbon sts! 13. 19 a 3) Carbon see (wotched) 4 19 131 Alloy ste (rotehed) a a1 nore: (1) The Unitied Numbering System (UNS) ws jintly developed by ASTM and SAE to provide for a consistant and uniform numbering system for commercially ible metals and alloys. The prefix letter designtos the family of metals, suchas G for ste! The frst ‘our numbers are generally taken from the traditional AISIGAE number system. For examole, G-10260 in the UNS AISI-SAE 1025 fred, strength is left in the shaft material which has already been exposed to cyclic stress for a given number of shaft cycles? Because fatigue damage is particularly sensitive tohigh stress amplitudes, startstop cycles, transient overloads, and petiods of high loads can have significant effect on fatigue life. For example, according to [13], a 20% in- crease in loading acting only 20% of the time can cause a life reduction from 30% to 6496 relative toa shaft with only constant amplitude loading. ‘There is considerable experimental evidence 2, 5,13] that the sequence of loading (that is, high amplitude loads following low ones or vice versa), can also have significant effect on the fatigue process. A number of experimental investigations [14] indicate that repeated application of stresses below the fatigue limit (that is, understressing), may actually improve the materials fatigue limit. Thus, for applications where the cyclic stresses vary in magnitude, but do not exceed the fatigue limit Sf of the materia, Ke =1 would produce a con servative design. However, subjecting shafts to stresses areater than Sf (that is, overstressing) fora significant number of stress cycles, would adversely affect the ma- teria’s endurance properties. For example, a 3% to 10% reduction in the fatigue limit occurs when a reversed- bending stress that is 10% greater than the origina fatigue limit is applied to the mild steel test specimens forthe first 20% of the cycles to failure [15]. When this stress is increased to 130% of the normal fatigue limit, £8 7% to 20% reduction inthe fatigue limit results. At present there is no single comprehensive method to determine factor ke for the effects described above, However, «shaft design method appearing in (13, 16] accounts for the effects of a variable amplitude Toad- ing history using a Palmgren-Minor linear damage law approach. On occasion, shafts are designed fora limited service life for purposes of economy. In a limited fatigue life design the number of eycles at stress amplitudes greater than the fatigue limit that a material can withstand with- out failing depends on the following: (@) the magnitude of the alternating tresses; DESIGN OF TRANSMISSION SHAFTING ANSVASME 8106.1M-1988 |AN AMERICAN NATIONAL STANDARD TABLE 3 FATIGUE STRESS CONCENTRATION FACTOR k; — TYPICAL VALUES FOR KEYWAYS IN SOLID ROUND STEEL SHAFTS! (Reproduced fom Juvinall F.C., Englagering Consideration of Stree, Strain and Strength, McGraw-Hill Book Company, 1967, By pormizion) Stoo! ‘Annealed {Note (21) (less than 200 BHN}) ‘uence and dawn {over 200 BHI) Nore: Profiled Keyway ‘Slod-Runner Keyway Str 050 063 (1) Nominal stresses shouldbe based on the section modulus forthe total shaft section. (2) May also apply to hot rolled shaft, (0) the slope of the stress vs cycles-o-‘failure curve for the material used; (©) the sitess history that preceded the current and subsequent stress. When the constant amplitude fatigue properties of a member and the actual force-time history to which it will be exposed are known, the current state of knowl- edge enables predicting its fatigue life within about 033, to 3.0 times the experimentally determined fatigue life [17]. In shaft design problems, itis usually possible to obtain or determine reasonably accurate information on ‘the constant amplitude fatigue properties of the material from which the shaft will be made. However, while shafts normally are exposed to many millions (or even billions) of stress cycles within their designed lifetimes, the actual force-time histories to which they will be exposed are rarely known very accurately. Therefore, this Standard does not recommend attempting to design shafts for fi- nite fatigue life without obtaining prototype fi data under simulated operating conditions. 4.6 Fatigue Stross Concentration Factor ky Experience has shown that a shaft fatigue failure al- most always occurs at 2 notch, hole, keyway, shoulder, or other discontinuity where the effective stresses have bbeen amplified. The elfect of stress concentration on the fatigue limit of the shaft is represented by the fatigue stress concentration factor ky where gp=__fatigue limit ofthe notched specimen __1_ "1 fatigue limit of a specimen free of notches — Ky and where K, Experimental data indicate that low strength steels are significantly less sensitive to fatigue at notches than are high-strength steels. The notch sensitivity q of materials fatigue strength reduction factor can be used to relate the fatigue strength reduction factor Kyo the theoretical (static) stress concentration factor Kas follows: Ky=14a(K- 1) ‘The appropriate theoretical stress concentration fac- tor K; to be used isthe value for bending. Thisis because the fatigue stress concentration factor Ky is used to mod- ify the specimen’s bending fatigue limit S7. Thus, eom- Dining the above equations yields ky /Ky= [1 +4a®- D) Recommended values for q and K; are given in Figs. 4 through 7 and Table 3, 4.7 Miscellaneous Effects Factor ky Since fatigue failures nearly always occurat or near the surface of the shaft where the streses are the greatest, surface condition strongly influences fatigue life. A num: ber of factors affecting the fatigue limit have values not readily found in design texts. Some of these factors (@ residual stresses (cold rolling, peening, welding, te.) (8) heat treatment (case hardening, decarburization, etc.) (©) corrosion (stress corrosion cracking, fretting cor- rosion, ete.) (@) plating or surface coating (e) interference-tt (collars and splines) (A) vibration G@) environment (thermal fatigue environment, ele trochemical fatigue environment, ec.) (1) unusual loading (axial loading, eyelic torsional and axial loads, thermal expansion loads, etc.) ANSI/ASME 8106,1M-1085 ‘AN AMERICAN NATIONAL STANDARD DESIGN OF TRANSMISSION SHAFTING ee as Co (0) For set subjects to evra bending rane aa lad. For area ute the values FIG. 4 NOTCH SENSITIVITY ¢ ‘Reproduced from Sines, G.,and Waltman, J. L., Metal Fatigue, MeGra-Hill Book ¢ ‘Company, 1959, by permission) “a Ty e ro EL Perot titi iid GENERAL NOTE: Sag *MefE wher a= ne8 FIG.5 ROUND SHAFT WITH SHOULDER FILLET IN BENDING o (eproducsd from Peterion, RE. Serees Concentration Factors, John Wiley & Sons, LU Copyright ©1974, by permission) 8 ANSI/ASME 8106.1M-1965 DESIGN OF TRANSMISSION SHAFTING [AN AMERICAN NATIONAL STANDARD tie 4 ae aresorinate) whol bitin © 0 010 015 070 028 020 FIG.6 GROOVED ROUND BAR IN BENDING (Roproduced from Paterson, RE, Stress Concentration Factor, Joh Wiley & Sons, Copyright ©1974, by permission) oe FIG.7 ROUND SHAFT IN BENDING WITH A TRANSVERSE HOLE (Reproduesd fram Peterson, R. E, Stes Concontration Factors, John Wily & Sons, ‘Copyright ©1974, by permission) ANSWASME B106.16-1095 ‘AN AMERICAN NATIONAL STANDARD Although only limited quantitative data have been published for these factors [1,2,5, 14, 16, 18-21] , they should, nonetheless, be considered and accounted for if applicable. Some of those factors can have a considerable effect on the shaft's endurance characteristics. In the absence of published data, itis advisable to conduct fe tigue tests that closely simulate the shaft condition and its operating environment. 5 FACTOR OF SAFETY (FS) In any design computation there is always some un- certainty as to how closely the analysis approximates the actual application and its resulting accuracy. The design procedures presented here assume that exact values will bbe used for the fatigue limit of the shaft and its tensile yield strength, However, in most cases nominal values for these material properties are all the data that are avuil- able. These values may differ slightly from the true val- vues for the batch of material on hand. Also, the rec- ommended fatigue modifying factors used to correct material fatigue properties may only approximate the true factors for the given application, Furthermore, appli cation factors which would adversely affect shaft life may be present, but are either not accounted for in this ‘Standard or are overlooked by the shaft designer. Un- anticipated loads, uncertainty in the load spectrum, or environmental factors can also shorten shaft life. In view 10 DESIGN OF TRANSMISSION SHAFTING of these considerations, itis prudent to introduce some ‘measure of conservatism into the design process to com: pensate for these uncertainties. Traditionally, this con- servatism takes the form of a factor of safety FS. AA factor of safety is an arbitrary value, generally estab- lished by experience and, in some cases, as a matter of opinion, In selecting a value for FS, the consequence of failure should be considered. Ifthe consequence of failure is high, an FS considerably greater than unity should be used. If the consequence of failure is minimal, an FS closer to unity may be used 6 OTHER CONSIDERATIONS The fatigue life of a shaft is not always the limiting factor in its design. The effect of the deflection of shaft ‘on the performance of support bearings and seals, as well as on such other key power transfer components as gears and pulleys, must also be taken into account, Shafts can be strong enough to meet fatigue life requirements, yet not stiff enough to keep the natural frequency of the shaft system above the operating speed range of the ma- chinery being designed. Careful attention to these fac- tors is essential for achieving a reliable design. Although ‘design information for these other factors is beyond the scope of this Standard, it may be found in most machine design texts. — APPENDIX A THEORETICAL BACKGROUND (This Appencix is not part of ANSI/ASME 8108.1§-1985, and is included for information purposes only.) 1 INTRODUCTION ‘The judicious use of power transmission shafting i portant, not only from a machine reliability standpoint, bbut also in light of cost and energy conservation. Al though the prime design consideration is whether the shaft will provide adequate servic life (that is, whether it will resist failure), it is seldom the only design consid- eration. The shaft must also be stiff enough between supports to limit deflections of key power transfer ele- ‘ments and stiff enough to avoid vibrational excitation, However, working knowledge in these other areas is ‘more complete in comparison to the limited knowledge of fatigue behavior of materials in shafting applications. Applying experimental fatigue data to shafting design is certainly not a new approach. However, rarely does the shaft designer have the appropriate fatigue data at hand to match the application. Although running screening tests on prototype parts is the most prudent approach, {few organizations can afford the cost and time of long- tetm endurance testing. Usually, the designer can consult design references containing shafting design formulas that sive acceptable designs for most applications. However, different formulas are sometimes inconsistent. There is often confusion as to which fatigue factors to use and ‘what relative importance to place upon them. ‘A2 FATIGUE FAILURE A ductile machine element subjected to repeated fluc- tuating stresses above its endurance strength but below its yield strength will eventually fail from fatigue. The insidious nature of fatigue is such that it occurs without visual warning at operating stresses below plastic defor- ‘mation, Shafts sized to avoid fatigue will usually be strong enough to avoid elastic failure, unless severe tran- sient or shock overloads occur. Failure from fatigue is statistical in nature, in that only the likelihood of failure based on a large population of specimens can be precisely predicted, rather than the fa- tigue life of any particular specimen [22-25] .Foragroup of specimens or parts made to the same specification, the key fatigue variables are the effective operating stress, the number of stress cycles, and the volume of material under stress Since the effective tresses are usualy highest at points along the surface where discontinuities occur, such as Keyways, splines, and fillets, these points are ‘those from which fatigue eracks are most likely to en nate. However, each volume of material under stress cartes with ita finite probability of failure. The product of these element probabilities (the “weakest ink” crite- rion) gives the likelihood of failure for the entice part of a given number of loading cycles. This is one ofthe under lying reasons why larger shafts generally have shorter fa tigue lives than smaller shafts under identical stress level. ‘At present, there is no unified statistical failure theory to predict shafting fatigue. However, reasonably accurate life estimates can be derived from general design equa- ions coupled with bench-type fatigue data and material static properties. Bending fatigue test data are usually ‘obtained from a flexure of rotating beam tester under ‘the conditions of reversed bending. The data generated from these machines are commonly plotted in the frm of stresslife (SN) diagrams. On these diagrams the bend- ing stress at which the specimens did not fail after at least 10° cycles for steel is commonly refered to as the fatigue limit. Due to test data scatter, the fatigue limit values determined from SIN diagrams usually represent some mean value and must be statistically corrected for higher reliability levels, as will be discussed late. I is customary to consider that design stresses less than the fatigue limit will produce an “infinite” lie design. This can be misleading, since no part can have a 100% prob- ability of surviving n infinite numberof stress cycles. ‘A3 FATIGUE UNDER COMBINED STRESSES For applications where 2 simple fluctuating stress of, the same kind is acting (for example, a steady bending. stress superimposed on a reversed-bending stress), a modi- fied Goodman failure line connecting the fatigue strength with the static strength provides an acceptable design [25]. However, most power transmission shafting is sub- jected to a combination of reversed-bending stress (8 10- tating shaft with constant moment loading) and steady, or nearly steady, torsional stress, Although a large body of test data has been generated for simple stress condi- tions such as pure tensile, flexural, or torsional stress, lit tle information has been published for combined bending and torsion stress conditions [26] . However, some cyclic bending and static-torsional fatigue test data were re- ported by Kececioglu and Lalli [27], and Davies [28]. ‘The endurance limit characteristics of notched UNS G- 43400 steel specimens were determined for theoretical bending stress concentration factors of 1.42 and 234 27). Fatigue testing for 3% nickel and nickel-chromium steel specimens was conducted under the same stress combination in a modified Wohler machine [28]. The results of both these experiments appear in Fig. Al, where the reversed-bending fatigue limit , is shown to decrease with an increase in static shear stress S,. Consid- ering that either fatigue fracture or torsional yielding represents failure, the following elliptical relation rea- soning fits the data [29] (Sp/SPP + GelSyy)? = 1 aay where oS reversed-bending fatigue limit of the test speci ‘men under bending only Syy = torsional yield strength The failure relation of Eq. (A-1) is similar to that ob- served by Gough and Pollard [30] for rotating beam specimens loaded under reversed bending in phase with reversed torsion as shown in Fig. A2. Those data, together with those shown in Fig. Ai, are in reasonable agreement with the distortion energy or von Mises-Hencky failure criterion. This theory predicts scatic elastic failure when the distortional energy under combined stresses equals or exceeds that in simple tension or bending. There isa reat deal of experimental evidence which indicates that Of all the failure theories, the distortion-energy theory predicts most accurately the yielding of ductile materials under static loading. However, itis not clear why the dis- tortion energy theory also seems to hold for some fatigue failures as well. ‘The distortion-energy elliptical failure relation is not the only one to be proposed for combined cyclic-bending and static-torsion loading. The tests performed by Ono [31] and Lea and Bogden {32} suggest that the bending 2 Se Reversed Bending Stet at Fatigue Limit FIG. A1 COMBINED STRESS FATIGUE TEST DATA, FOR REVERSED BENDING IN COMBINATION So Reversed Bending Stet at Fatigue Limit FIG. AZ COMBINED STRESS FATIGUE TEST DATA FOR REVERSED BENDING IN COMBINATION Fatigue Limit in Pare Banding Fatigue Limit in Pure Banding 101 [Ni-Cr-Mo Stel, AISI 4340 {from Fe. (271) © Ke 1.42 (Bending) K_= 2.08 (Bending) NiCr Stel (from Fe. (28)) 3% Ni ste! (from Ret. [28)) ea00 ° (2 ont OnpeetosmemsD) Static Torsional Suess _Ss Torsional Yield Seenath “5, ©) WITH STATIC TORSION [28] © 0.1% carbon Stee 2 3.5% NiCr Stet Fatigue Limitin Pure Torsion Syye ® WITH REVERSED TORSION [30] Safe shaft design regions Combined strese method (Ea, (21) ‘Separate etree method (Ret, (81) Sey cs Strength ee Reveread-encing Stress Sp Allowable Torsional Summon Mean-Torsonal Stross S, FIG.A3_ COMPARISON OF COMBINED STRESS AND SEPARATE STRESS DESIGN METHODS fatigue strength of steel is unaffected by the presence of 4 staticorsional stress, even above the torsional yield strength. On the basis, in part, of this test information, Wellauer [33] recommends that the allowable bending fatigue strength and the allowable static-torsional stress for gear drive shafts be calculated separately. A compari- son between separate stress and combined stress shaft methods is illustrated in Fig. A3. From a reliability stand- point, the combined stress relation of Eq. (A-1) produces a slightly more conservative design. However, the differ- fences are not great. For most designs, the difference in shaft diameters will be less than 15% when equal allow- able stresses are used. ‘Ad DEVELOPMENT OF FORMULAS. ‘The following shaft design formulas are applicable to rotating solid shafts under the most common variety of loading conditions, namely fully reversed bending in combination with static torsion, les than torsional yield, ‘with nogligible axial loading, For design purposes, a factor of safety FS can be in- corporated into the failure rolation of Eq. (A-1). The fol- 13 owing equation can be written (SolS pa)? + (SulSey. (a2) allowable shaft endurance limit = Sy/FS allowable shaft torsional yield strength IPS Therefore, (SpPSISP? +(SYFSISzy = 1 (AB) For most steels, the torsional yield strength is related | to the tensile yield strength by Say = SIV (aay ) ‘and the reversed-bending stress equals i Sp = 32M fea as 7 and the steady-tosional stress can be found from A) S,= 16 7/xa° aol U Substituting Eqs. (A-4), (A'S), and (A-6) into (A-3) ee) G7 eye eA VE @]" AS LIMITED LIFE AND VARIABLE AMPLITUDE LOADING DESIGNS 2) VOR ‘Traditional shaft analysis generally considers that the nominal loads acting on the shaft are essentially of con: stant amplitude and that the shaft life is to exceed 10° or 10” cycles. Sometimes shock or overload factors are applied, However, most shafts in service are generally exposed to a spectrum of service loads. Occasionally, shafts are designed for lives that ae less than 10° cycles for purposes of economy. Both of these requirements complicate the method of analysis and increase the uncet- tainty of the prediction. Under these conditions, proto- type component fatigue testing under simulated loading ‘becomes even more important. Short Life Design. Local yielding of notches, fillets, ‘and other points of stress concentration are to be ex- pected for shafts designed for short service lives, les than about 1000 cycles. Since fatigue cracks inevitably orig inate at these discontinuities, the plastic fatigue behavior of the material dictates its service life. Most materials have been observed to either cyclically harden or soften, depending upon their initial states, when subjected to cyclic plastic strain. Therefore, the cyclic fatigue prop- erties of the material, which can be significantly different from its static or monotonic strength properties, need to bbe considered in the analysis. For short, low cycle life designs, the plastic notch strain analysis, discussed in detail in (17, 34, 35] is considered to be the most accu- rate design approach. This method, used widely in the automotive industry, predicts the time at which crack formation occurs based on an experimentally determined relationship between local plastic and elastic strain and ‘the number of reversals to failure. Intermediate and Long Life Designs. For intermediate and long life designs both total strainsife and nominal stressdife (SN curve) methods have been successfully applied, [16, 17, 35]. Although both methods provide reasonable fatigue life predictions, the nominal stress- life method is more commonly applied for shaft design. ‘The reversed-bending fatigue limit S, of the shaft to ‘be designed is different from the endurance limit of ro- tating beam specimens Sf normally listed in design ta- bles. A number of factors which can affect the endurance ‘of the material in actual service have been identified. Modifying factors should be applied to the uncorrected “4 endurance limit of the test specimens to determine Sy [19]. These are: Sy=kakpkokakekphy Sf ‘Values for these modifying factors are discussed in Section 4. ‘The key to accurate fatigue life prediction is obtaining good definition of stress, S-N, characteristics of the shaft material. Mean bending andjor torsional stress ef- fects should be taken into account if present. Further- more, a good definition of the loading history is also required. Even when these requirements are met, the ac- curacy of the prediction is approximate with today'sstate of knowledge. As an example, an extensive cumulative fatigue damage test program was conducted by the SAE to assess the validity of various fatigue life prediction ‘methods [17]. Numerous simple geometry notched steel plate specimens were fatigue tested in uniaxial tension, ‘Tests were conducted under constant amplitude loading and also under a variable amplitude loading that closely simulated the service loading history. The test specimens” ‘material fatigue properties and the actual force-time his tory were very well defined. Under these well-controlled conditions, predicted mean life from the best available method was within a factor of 3 (¥/s to 3 times) of the true experimental value for about 80% of the test speci- ‘mens while some of the other methods were considerably less accurate [17]. Under less ideal conditions, such as when the loading history and material properties are not as well known or when a multiaxial stress state is im- posed, a predictive accuracy within a factor of 10 of the ‘rue fatigue life would not be unacceptable with today’s state of knowledge In view of these considerations, variable amplitude and limited life designs have not been formally addressed in this Standard. Howover, references (2,5, 13,14, 16,17, 34, 35] can be consulted for further guidance. APPENDIX B MATERIALS AND PROPERTIES (This Appendix is not part of ANSI/ASME B106.1M-1985, and is included for information purposes only.) 81 INTRODUCTION To use the shaft design formulas in Section 3, itis nec- essary to have accurate values for the reversed-bend- ing fatigue limit $7 and yield strength Sy of the shaft material. If the reversed-bending fatigue limit SF can be ob- tained by testing for @ particular stel, that value should bbe used. If laboratory values or published data are not available, the approximation, Sf = 0.5 (tabulated ultimate tensile strength), should. provide reasonable accuracy. For tensile strength over 200,000 psi (1380 MPa), use 5} = 100,000 psi (690 MPa) The tensile and yield strength properties of the shaft material may be obtained through testing or from the steel supplier If data are not available from these sources, 1s ‘Table BI may provide some guidance. It contains repre- sentative data from several sources on mechanical prop- erties of steels in various conditions of processing and thermal treatment. The table values shown are not spect- fications and, therefore, should not be interpreted as guar- anteed, absolute, or implied values. They are intended only as a guide to the selection of a material. Note that the table is accompanied by descriptive information re- garding the condition of the steel to which the data apply [7, 36-39]. Many variables affect the properties of steel; ‘therefore, a given material should not be specified with- ‘out frst establishing with the steel producer the certainty of his furnishing the desired mechanical property require- ‘ments for the grade of steel under consideration. When ‘minimum mechanical properties are required, they should be specified when the steel is ordered. Ordering a specific ‘rade and/or thermal treatment does not guarantee spe- cific mechanical properties. oe {sonunv09 age) © ‘}qe4, 30 PUB # a10N pue SUEWIEPEHMOLYEY 805, ee el ov 19 3 ts ao bu 29 soe ees eu wy ier ew ese 9 e08 “oN ere ut 63 os ee sz 399s eo uH o0z01-9 ez tot ov st os ov us-LH-a9 zmb ube ov ot ss sy S-LH-a9 yim tb 9° oz 09 sy US-LH-09 wm tet se oz so Sy uS-L#-09 eee 0 se st ss sy ao zur zt se st 09 os aa wim Leb or ot so 3s ao ume ee ov al on 09 ao oz ov as v9 ¥s ao et os sz 8s we UH 2101-9 ue ov aL ose uw a9 mt 69 “e as ew uy wt 969 a sie & “won, ozt to 19 os ee a 13s os say saz UH osioV-0 sor ov er +9 sv 09 86 05 9 ey soz aH oz101-9 sor ov oz es vy 99 36 os a we 9 UH 0101-9 86 sv oz ey we 09 98 ss of ey ve aH 9001-9 “uy LbeL WSV vier Lie WSY PLPr Lie SY PiOr hel oWSY Plor Li-@L SY bier _uoRpuog pu wera IsNy aS ISNY avs ISNY 3vS ISN avs ISNY avs) ONT c sequin SNA NH ‘ssoupi0H, eeiy 30 “uz Tt AnBuong onsen, rsnduens prea woponpey iusaieg "wopeBu0i3 w99te4 STHALS ONUIVHS 40 S3LLN3d0Ud TVOINVHOSIN BALLWLNASIUGIY La TeV. 16 fsonuyuos ayges) ‘\g2L.J0 pus 28 ION puE SwUOLIBpoMoUYDY 90s ze 80 ez 26 £0 LBD 20021 ziz 5 re 92 vo te 4180 0001 tee vs Wz 01 08 180 008, ssc : es oe e 98 180 2009 a. ee 6 e 98 10 008, se . oe 3s s-1H-09 ov se 08 usaH-a9 sy 08 so US-LH-a9 ap se 99 uS-AH-09 oF og ox S109 oe se 4 ae us-17-09 se 06, 08 uss-09 se 36 se us-Lrc9 08 se $8 0 oe 08 ox ey se se se a9 se 08 o8 09 ou se za se v4

You might also like