ISSN 0208-7774 T R I B O L O G I A 1/2020 p.

49–55
DOI: 10.5604/01.3001.0014.0844

Michał Libera*

A SIMPLIFIED WAY OF ESTIMATING a DYNAMIC LOAD rating

for thrust roller BEARINGS

Uproszczony sposób szacowania nośności dynamicznej

łożysk wzdłużnych wałeczkowych

Key words: ISO 281, rolling bearings, dynamic load rating.

Summary: The method of calculating the bearing capacity of rolling bearings is described in the ISO 281 standard. The
calculation procedure for roller thrust bearings presented there, depending on the value of the nominal bearing
angle, requires the selection of one of two formulas. Then, using the table, one reads the value of the factor
depending on the geometry of the bearing components. To facilitate and speed up calculations (and perhaps
also increase their accuracy), this article proposes a formula that is adapted to numerical applications, replaces
linear interpolation with a proper non-linear function and allows calculations to be made for a specific value
of the nominal bearing angle, but not within the range of 15°. The difference between the values calculated
according to the proposed formula and the value calculated according to ISO 281 is, on average, around 3%.

Słowa kluczowe: ISO 281, łożyska toczne, nośność dynamiczna.

Streszczenie: Sposób obliczania nośności łożysk tocznych jest opisany w normie ISO 281. Przedstawiona tam procedura
obliczeniowa dla łożysk wzdłużnych wałeczkowych, w zależności od wartości nominalnego kąta działania
łożyska, wymaga wyboru jednego z dwóch wzorów. Następnie trzeba, korzystając z tabeli, odczytać war-
tość współczynnika zależnego od geometrii części składowych łożyska. By ułatwić i przyspieszyć obliczenia
(a być może także zwiększyć ich dokładność), w niniejszym artykule zaproponowano wzór, który jest przy-
stosowany do aplikacji numerycznych, zastępuje interpolację liniową właściwą funkcją nieliniową i umożli-
wia wykonanie obliczeń dla konkretnej wartości nominalnego kąta działania łożyska, a nie dla zakresu o roz-
piętości 15°. Różnica między wartościami obliczonymi według proponowanego wzoru a wartością obliczoną
według normy ISO 281 średnio wynosi ok. 3%.

Introduction The history of calculating bearing life began over

60 years ago when Gustaf Lundberg from the Chalmers
Most rolling bearings lose their usability due to surface Institute of Technology and Arvid Palmgren from the
fatigue. Surface fatigue is a type of wear in which local SKF AB bearing company applied Weibull's theory of
loss of consistency and associated material losses are probability [L. 3, 4] regarding fatigue of material to
caused by material fatigue as a result of cyclical contact determine the life of rolling bearings. Their fundamental
stress (within Hertz stress limits) in the surface layer of work of 1947 [L. 5, 6] and 1952 [L. 6, 7] regarding
mating elements (rolling or sliding with slipping), with internal stress distribution, equivalent loads, and
lubricated contact (pitting) or dry (spalling) [L. 1, 2]. statistical bearing life distribution shaped the foundations
The element's service life until reaching the limit state of ANSI/ABMA and ISO standards describing bearing
due to fatigue wear was adopted as the term “rolling life, giving rise to catalogues of bearing manufacturers.
contact fatigue” (RCF). First, the simplest method for calculating bearing life

*
ORCID: 0000-0003-2698-9227. Poznan University of Technology, Faculty of Civil and Transport Engineering, Piotrowo 3
Street, 60-965 Poznań, Poland.
50 T R I B O L O G I A 1/2020 ISSN 0208-7774

adopted by ISO in 1962 was presented in ISO 281, and Z– the number of rolling elements;
then it was modified, and the current version dates from bm – rating factor for contemporary, commonly used,
2007 [L. 8, 9, 10]. Research carried out over the last high quality hardened bearing steel in accordance
several years, e.g., [L. 11–24], was not reflected in the with good manufacturing practices the value of
standard. which varies with bearing type and design;
fc – factor which depends on geometry of the bearing
components, the accuracy to which the various
The purpose of the article components are made, and the material;
α– nominal contact angle, in degrees.
In all versions of the standard, to calculate the dynamic Estimating dynamic load rating according to ISO
load capacity, it is necessary to select the appropriate 281 should start with calculating the quotient values:
formula depending on the bearing geometry and read
the corresponding data from the table. This procedure Dwe
facilitates difficult numerical calculations and quick – for α = 90° (1)
D pw
estimation of bearing capacity for various bearing
construction variants.
In addition, the tables are discrete (provide Dwe cosα
– for α ≠90° (2)
selected values), and intermediate values, according to D pw
the comment in ISO 281, should be calculated using
linear interpolation. However, the parameters given in Then one can read the value of the factor fc
the tables are not linearly dependent, which makes this from the table contained in the standard. The first two
interpolation inaccurate. columns of the table relate to bearings in which the angle
The third disadvantage of calculations according α = 90°. The first column presents the values of the
to ISO 281 is that, for example, for an nominal contact quotient (1) from 0.01 to 0.3 every 0.01 – and in the
angle between 45° and 60°, the column “α = 50°” second, the corresponding values of the coefficient fc.
should be used – similarly for other values, in ISO 281 The third column presents the values of the quotient (2),
standard, the angles are approximate and apply to angle and in the next three represent the corresponding values
ranges (15°). of the coefficient fc for the angles a of 50°, 65°, and 80°.
In the context of the listed imperfections of the ISO The table is supplemented by a comment which shows
281 standard, the purposes of the work are the following: the following:
–– Simplifying the calculation procedure and adapting – The values of the fc coefficient for the values of the
it to numerical applications, quotients (1) or (2) other than those given in the table
–– The replacement of linear interpolation with an should be calculated using linear interpolation;
appropriate non-linear function, and – Column “α = 50°” is used for angles from 45° to 60°,
–– Enabling calculations to be made for the exact angle column “α = 65°” for angles from 60° to 75°, and
α, not for a range of 15°. column “α = 80°” for angles 60°–75° and 75°–90°.
Depending on the angle α, the axial dynamic load
capacity, Ca, of single row thrust roller bearings is
Estimation of dynamic load rating expressed by the following formulas [L. 8, 9]:
of cylindrical roller thrust
bearings according to ISO 281
– For α = 90° 7
Ca = bm f c Lwe 9 Z 4 Dwe
3 29
27 (3)
The ISO 281 standard provides methods for calculating
– For α ≠ 90° Ca = bm f c ( Lwe cos α ) tgα Z 4 Dwe
7 3 29
(4)
9 27
the basic dynamic rating of rolling bearings and basic
rating life. The basic dynamic rating is defined as
a constant stationary load which a rolling bearing can Proposed changes in the method
theoretically endure for a basic rating life of one million of calculating dynamic load rating
revolutions. The basic rating life is associated with 90%
reliability for bearing operating under conventional To simplify the calculation procedure and adapt it
operating conditions [L. 8, 9]. to numerical applications, one can replace linear
For the purposes this paper, the symbols given in interpolation with a proper non-linear function and
ISO [L. 8, 9, 25] and the following apply: enable calculations for the exact angle α (and not for the
Ca – basic dynamic axial load rating, in Newtons; range) in accordance with the purpose of the work as
Dwe – roller diameter applicable in the calculation of follows:
load ratings, in millimetres; –– Develop a formula that would not require the use of
Dpw – pitch diameter of bearing, in millimetres; tables,
Lwe – effective roller length applicable in the calculation –– Develop one general formula for all values of the
of load ratings, in millimetres; angle α (also for α = 90°).
ISSN 0208-7774 T R I B O L O G I A 1/2020 51

To achieve this, replace the expression Deriving the formula replacing the tables with
coefficient values fc
f c ( cos α ) 9 tgα
7
(5) In order not to need a table to determine the value of
the fc coefficient, it must be enclosed in the form of
from Formula (4) such that a formula which, as described in the previous section,
–– It will allow the calculation of fc (without tables). depends on the angle and the ratio of the roller diameter
–– For angle α = 90°, it will reduce to fc. to the pitch diameter (1), (2). The following chart
For this purpose, the parameter Mα has been (Fig. 1) shows the fc values read from the ISO 281
introduced, which, for the angle α = 90°, will be 1, and standard for individual angles α (201 cases).
for the other values of the angle α, it will be calculated
from the following trigonometric expression:

 for α = 90° 1
Mα = 
 (6)
M
 for α ≠ 90° ( cos α )
7


9
tgα

So Formulas (3) and (4) can be reduced to one

equation of the following form:
7 3
Ca = bm Lwe 9 Z 4 Dwe
29
27
fc Mα (7)

Then the task of simplifying the method of Fig. 1. The values of the fc factor according to the table in
estimating the dynamic load rating of thrust roller the ISO 281 standard
bearings comes down to developing formulas that allow Rys. 1. Wykres wartości współczynnika fc opracowany na
calculating the product fcMα for various angles α and podstawie tabeli w normie ISO 281
diameter ratios, without the need to use tables.

For each of the angles α separately, the relationship be described by a logarithmic function, obtaining high
between the Dwe/Dpw quotient and the fc coefficient can fit values (fc' means approximation fc):

D 
– For α = 50° f c' = 22.338 ln  we  + 210.82 R² = 0.9158 (8)
 D pw 

D 
– For α = 65° f c' = 25.335 ln  we  + 204.65 R² = 0.9791 (9)
 D pw 

D 
– For α = 80° f c' = 27.347 ln  we  + 183.11 R² = 0.9968 (10)
 D pw 

D 
– For α = 90° f c' = 36.839 ln  we  + 264.09 R² = 0.9884 (11)
 D pw 

As measures of the quality of the approximation

function (fc’) fit to the data from ISO 281 (fc), the
following were adopted:
2
2
 cov( X , Y ) 
– The coefficient of determination [L. 26] R =  (12)
 σσ
 x y 
52 T R I B O L O G I A 1/2020 ISSN 0208-7774

therefore, a separate formula for α = 90° is given in the

f ci − f ci'
– The maximum relative difference max (13) ISO 281.
f ci In addition, in Formula (4), there is tg α, which is
not in Formula (3). From the properties of trigonometric
functions, the following is known:
201

– The average relative difference

∑ f ci − fci'
fci cos α ⋅ tgα = sin α (15)
i =1 (14)
201
Therefore, for α = 90°, this product would be 1 and
The maximum relative difference between the Formulas (3) and (4) would not be different, but one
value of the fc coefficient selected from the ISO 281 cannot forget the exponent, and, in fact, in Formula (4),
standard and calculated from Formulas (8) to (11) is the expression related to the angle α has the following
10%, and the average relative difference is 1.4% (201 form:
( cos α )
7
cases were analysed, i.e. all resulting from the data in 9
tgα (16)
the table, which is in ISO 281; fci stands for the ith case).
which is not equal to sin α.
Derivation of dependence replacing trigonometric
functions In the tables of ISO 281, four angle values α are
If, in Formula (4), instead of the cosine function, there considered: 50°, 65°, 80°, and 90°. Expression (16)
was a sine function, then for the angle α = 90°, the assumes for these angles the following values: 0.85,
expression with the exponent 7/9 would simplify to 1.10, 1.45, and 1 (for 90° angle).
Lwe (sin 90° = 1). Then, in this part, Formula (3) would Expression (5) can be replaced by products of
be a special case of formula (4) for α = 90°. However, Formulas (8) to (10) and Expression (16):
Formula (4) has the cosine function (a cos90° = 0);

D 
f c' ( cos αa ) 9 tgaα = 18.878 ln  we  + 178.16
7
– For α = 50° R² = 0.9158 (17)
 D pw 

D 
f c' ( cos αa ) 9 tgaα = 27.805 ln  we  + 224.6
7
– For α = 65° R² = 0.9791 (18)
 D pw 

D 
f c' ( cos αa ) 9 tgaα = 39.74 ln  we  + 266.9
7

– For α = 80° R² = 0.9968 (19)

 D pw 

– For α = 90° Formula (11) in its unchanged form.

The use of these formulas frees one from the need where Mα (6) for the angle α = 90° is 1, and for the other
to use a table, but it introduces another four formulas that values of the angle, α is calculated from Expression
should be used depending on the angle, which is difficult (16), while a and b are coefficients dependent only on
to call a simplification. However, these formulas can be the angle α, which can be approximated by the following
saved in general form as formulas:

'
 Dwe  a = 0.5045α − 5.1308 R² = 0.8742 (21)
f c M α = a ln  D  + b, (20)
 pw  b = 2.2752α + 71.128 R² = 0.9254 (22)

Inserting (21) and (22) into (20) produces:

D 
f c' M α = ( 0.5045α − 5.1308 ) ln  we  + 2.2752α + 71.128 (23)
D 
 pw 
ISSN 0208-7774 T R I B O L O G I A 1/2020 53

The directional coefficient of the relationship the free word equal to 0); therefore, Formula (23) was
between the expression fcMα calculated according to corrected by dividing it by this value and obtained the
ISO 281 and according to Formula (23) is 0.9794 (with following:

D 
f c' M α = ( 0.5151α − 5.2387 ) ln  we  + 2.3229α + 72.6241 (24)
D 
 pw 

For all 201 cases that can be calculated on the difference is 13%, but the average difference is about
basis of data from the ISO 281, a calculation was made 3%. The points in the graph are arranged in several
according to the proposed Formula (24), and the effects series, which is the effect of the angle ranges and two
of the comparison are presented in Fig. 2. formulas adopted by the ISO 281 standard. However,
in reality, the phenomenon is continuous (which further
justifies the use of the proposed formula).

Modified formula for dynamic load rating

After substitution (24) to (7), a simplified formula was
obtained for the dynamic load capacity of thrust roller
bearings:

 D
Ca = bm Lwe 9 ⋅ Z 4 ⋅ Dwe 27 ⋅ ( 0.515α − 5.24 ) ln we +
7 3 29

 D pw
 Dwe 
Ca = bm Lwe ⋅ Z ⋅ Dwe ⋅ ( 0.515α − 5.24 ) ln
7 3 29
9 4 27
+ 2.32α + 72.6
 D pw 
(25)

The only source of differences between the dynamic

Fig. 2. Relationship between the fcMα calculated according load rating values calculated according to ISO 281 and
to ISO 281 and the proposed formula
according to the proposed Formula (25) is the expression
Rys. 2. Zależność między wyrażeniem fcMα obliczonym wg
normy ISO 281 i wg proponowanego wzoru fcMα. Therefore, these differences amount to an average
of about 3%, which is described in the commentary to
Using the measures of fit quality described by Formula (24) and Fig. 2. Table 1 presents examples of
Formulas (12), (13), and (14), it can be concluded that calculations made on the basis of ISO 281 and according
Formula (24) gives a high approximation accuracy. to the proposed formula (25).
The coefficient of determination is 0.97, the maximum

Table 1. Dynamic load rating of the example bearings calculated on the basis of ISO 281 and the proposed formula (25)
Table 1. Nośność dynamiczna przykładowych łożysk obliczona na podstawie ISO 281 oraz proponowanego wzoru (25)

bm Dwe Dpw Lwe α Z fc ISO281 Ca ISO281 Ca (25) Relative

No.
– [mm] [mm] [mm] [°] – – [kN] [kN] difference
1 1 3 30 5 90 15 175.7 15.2 16.2 6.3%
2 1 20 500 40 90 13 143.4 432.0 449.1 4.0%
3 1 10 40 30 90 11 215.4 217.4 226.5 4.2%
4 1.1 3 30 5 50 15 160.9 13.0 13.5 4.0%
5 1.1 20 500 40 50 13 139.5 390.6 406.2 4.0%
6 1.1 10 40 30 50 11 183.7 172.3 177.8 3.2%
7 1.1 3 30 5 65 15 144.7 15.2 15.1 0.3%
8 1.1 20 500 40 65 13 124.7 453.5 439.1 3.2%
9 1.1 10 40 30 65 11 173.6 211.5 204.6 3.3%
10 1.1 3 30 5 80 15 123.0 17.1 16.7 1.9%
11 1.1 20 500 40 80 13 105.6 508.5 472.0 7.2%
12 1.1 10 40 30 80 11 142.8 230.4 231.3 0.4%
13 1.15 30 300 40 50 15 160.9 810.6 842.8 4.0%
14 1.15 30 300 40 65 13 144.7 850.4 848.1 0.3%
15 1.15 30 300 40 80 11 123.0 844.4 828.6 1.9%
54 T R I B O L O G I A 1/2020 ISSN 0208-7774

Summary the linear interpolation of data from the table

with a proper non-linear function and enabling
The dynamic bearing capacity provided by manufacturers calculations for a specific angle α, not for a range of
is experimentally verified in durability tests [L. 27, 28], 15°).
due to the large spread, covering at least 20 bearings, The average difference between the dynamic load
which makes them costly and time consuming. In unit capacity calculated according to ISO 281 and calculated
applications for which special rolling nodes are designed, according to the proposed formula (25) is about 3%, and
e.g., large-size coronary bearings, it is not possible to in no case exceeds 13%.
carry out such tests, so it is worth improving the tools Therefore, the differences are small, and they do
for calculating the load capacity. One of such tools is the not necessarily mean a worse estimation of the dynamic
ISO 281 standard, which, however, due to its specificity, load rating based on the proposed formula (25). In
described in chapter two, makes it impossible to quickly calculations according to ISO 281, the values ​​of the
assess the load capacity for many variants of the rolling fc coefficient for the quotient values ​​other than those
node structure. given in the table are obtained by linear interpolation,
Thus, in accordance with the purpose of the work, which is less precise than the proposed method (and
a method of calculating the dynamic load rating of requires additional calculations). An even greater source
cylindrical roller thrust bearings (Formula 25) was of imprecise estimation of dynamic bearing capacity
proposed, which accomplishes the following: based on ISO 281 is the fact that, for angles α from 45°
– Simplifies and speeds up calculations (by deriving to 60°, the values ​​of fc from the column “α = 50°” are
one general formula for all values of the angle α, also used, while the next two columns of the table from the
for α = 90°); standard use, respectively, within the ranges of 60°–75°
– Facilitates the use of numerical methods (the and 75°–90°. Therefore, in the ISO 281 standard, the
developed formula does not require the use of angles are approximate and relate to the ranges of angles
tables); and, (every 15°), similarly to the ranges of diameters, and
– Makes the calculation result more dependent on the developed Formula (25) allow calculations for any
bearing parameters than ISO 281 (by replacing values ​​of these parameters.

References

1. Hebda M., Wachal A.: Trybologia. WNT, Warszawa 1980.

2. Hebda M.: Procesy tarcia, smarowania i zużywania maszyn. Wydawnictwo Instytutu Technologii Eksploatacji –
PIB, Warszawa – Radom 2007.
3. Weibull W.: A statistical distribution function of wide applicability. Journal of Applied Mechanics, 1951, No 18.
4. Weibull W.: A statistical theory of the strength of materials. Ingeniörsvetens-kapsakademiens, Handlingar Nr 151,
Generalstabens Litografiska Anstalts För-lag, Stockholm 1939.
5. Lundberg G., Palmgren A.: Dynamic Capacity of Rolling Bearings, Acta Polytech, Mechanical Engineering
Series, Vol. 1, No. 3, 7, R.S.A.E.E., Stockholm, Sweden, 1947.
6. Palmgren A.: Łożyska toczne. PWT, Warszawa 1951.
7. Lundberg G., Palmgren A.: Dynamic Capacity of Rolling Bearings. Acta Polytechnica Mechanical Engineering
Series, Vol. 2, No. 4, Stockholm, Sweden, 1952.
8. Polski Komitet Normalizacyjny: Łożyska toczne. Nośność dynamiczna i trwałość PN-ISO 281. Wydawnictwa
Normalizacyjne 1998.
9. International organization for Standardization, “Rolling Bearings-Dynamic Load Ratings and Rating Life”,
ISO281, Geneva, 2007
10. International Organisation for Standardization 2010 DIN ISO 281:2010-10 – Dynamic load ratings and rating life
10.
11. American National Standard Institute 2015 ANSI/ABMA 9: Load Ratings and Fatigue Life for Ball Bearings.
12. Ioannides E., Bergling G., Gabelli A.: An Analytical Formulation for the Life of Rolling Bearings, Acta
Polytechnica Scandinavica, Mechanical Engineering Series No. 137, The Finnish Academy of Technology, 1999
13. Harris T.A.: Rolling Bearing Analysis, 4th Edition, John Wilsey & Sons Inc., 2001.
14. Gupta P., Oswald F., Zaretsky E.: Comparison of Models for Ball Bearing Dynamic Capacity and Life. Tribology
Transactions, 58: pp. 1039–1053, 2015.
ISSN 0208-7774 T R I B O L O G I A 1/2020 55

15. Harris T.A., Kotzalas M.N 2007 Rolling bearing analysis 5th Edition (London: Taylor & Francis Group, CRC
Press)
16. Libera M.: Prognozowanie trwałości łożysk tocznych. Wydawnictwo Politechniki Poznańskiej, Poznań 2013.
17. Bhadeshia H.K.: Steels for bearings. Progress in material science, 2012, Vol. 57, pp. 268–435.
18. Desvaux S., Duquennoy M., Gualandri J., Ourak M.: The evaluation of surface residual stress in aeronautic
bearings using the Barkhausen noise effect. NDT&E International 37, Elsevier 2004.
19. Morales-Espejel G.E., Lugt P.M., Van Kuilenburg J., Tripp J.H.: Effects of Surface Stresses of Pure Rolling EHL
Contacts. STLE Tribology Transactions, 2003, Vol. 46, pp. 260–272.
20. NASA: RCM Guide Reliability-Centered Maintenance Guide. For Facilities and Collateral Equipment. National
Aeronautics and Space Administration, September 2008.
21. Yang G.: Life cycle reliability engineering. John Wiley & Sons, Inc., Hoboken, NJ, USA, 2007.
22. Zaretsky E., Branzai E.: Effect of Rolling Bearing Refurbishment and Restora-tion on Bearing Life and Reliability.
NASA Technical Memorandum 212966, 2005.
23. Zhang R., Mahadevan S.: Reliability-based reassessment of corrosion fatigue life. Structural Safety, 2001, Vol.
23, pp. 77–91.
24. Hong-Zhong H., Kang Y., Tudi H., He L., Hua-Ming Q.: Reliability estimation for momentum wheel bearings
considering frictional heat.s. Eksploatacja i Niezawodnosc – Maintenance and Reliability, Vol. 22, No. 1, 2020,
pp. 6–14.
25. International organization for Standardization, Rolling bearings – Symbols for physical quantities ISO15241,
Geneva 2012.
26. Stanisz A.: Przystępny kurs statystyki w oparciu o program STATISTICA PL na przykładach z medycyny. StatSoft
Polska, Kraków 1998.
27. Bak M., Libera M.: Usability of stand for rolling contact fatigue life testing of roller bearings elements. Journal of
Research and Applications in Agricultural Engineering, 2010, Vol. 55, pp. 6–10.
28. Jurecki R.: Design of a test rig for the examination of mechanical properties of rolling bearings. Management
Systems in Production Engineering. 2017, No 1 (25), pp. 22–28.