Fracture failure analysis of gearbox shaft V.S. Rocha et al.

Fracture failure analysis of gearbox shaft

V.S. ROCHA(1), J.B. DE CAMPOS(1), A.R. PIMENTA(1,2), M.G. DINIZ(1)*, M.A.S. PINHEIRO(3)

(1) Universidade do Estado do Rio de Janeiro, Programa de Pós-Graduação em Engenharia Mecânica, Rio de
Janeiro, 20940-240 Rio de Janeiro, Brasil
(2) Instituto Federal do Rio de Janeiro, Dept. de Mecânica, Campus Paracambi, 26600-000 Rio de Janeiro, Brasil
(3) Universidade do Estado do Rio de Janeiro, Programa de Pós-Graduação em Ciência e Tecnologia dos
Materiais, Nova Friburgo, 97282 Rio de Janeiro, Brasil
*diniz@[Link]

ABSTRACT: The aim of this study was to analyze a fracture failure in a gearbox shaft of a scrap compressing machine and
to suggest methods to prevent the problem. The shaft was made of a quenched and tempered SAE 1045 (ABNT 1045) steel.
Scanning electron microscopy (SEM), chemical analysis by semi-quantitative EDS (Energy Dispersive Spectroscopy) and
optical microscopy (OM) were also used in order to get hardness measurements. The fracture mechanism has been found as
brittle fracture by cleavage and the chemical analysis by EDS results have shown that it was steel with a significant presence
of iron and manganese. Micrographic analyses have also revealed a typical microstructure of tempered martensite. The
microstructure has shown to be in the whole length of the bar. The conclusion was that the failure has occurred because of
quenching and tempering it not in a proper way, together with some overload in the operation.

Keywords: Failure analysis, Fractography, Fracture prevention, Brittle fracture.

RESUMO: O objetivo deste trabalho foi analisar a falha por fratura de um eixo de uma caixa redutora de uma máquina
para compactação de sucata e sugerir métodos de prevenção do problema. O eixo era de um aço SAE 1045 (ABNT 1045)
temperado e revenido. Foram utilizadas técnicas de microscopia eletrônica de varredura (MEV), análise química semi-
quantitativa por EDS (Energy Dispersive Spectroscopy), microscopia ótica (MO), além da obtenção de medidas de dureza.
O mecanismo de fratura encontrado foi o de fratura frágil por clivagem e os resultados da análise química por EDS permitiram
concluir que se tratava de um aço com presença significativa de manganês. A análise micrográfica revelou ser a de um aço
temperado e revenido com microestrutura típica de martensita revenida. A microestrutura apresentou-se bastante homogênea
para toda a extensão da barra. A conclusão foi de que a falha ocorreu por têmpera inapropriada associada a uma possível
sobrecarga na operação.

Palavras chave: Análise de falha, Fractografia, Prevenção de falhas, Fratura frágil.

1. Introduction ness-for-service criteria are continuously evolving, reflecting

advances and new consensus in the technical and scientific
The development of improved materials in Applied Mechan- communities. Although there is a good number of papers on
ics, with optimized properties, is one of the main develop- this topic [4-7], the continuous development of the subject
ment areas in Materials Science, since there is an increase justifies the need for an updated critical survey.
of the technological demands and of the technical character-
istics needed for the most different applications. The analy- Although general safety is the main concern of legislation,
sis of failures aims at avoiding project errors and reducing defective products, which do not meet the expected perform-
ance, are some of the causes of financial loss. This kind of
losses and costs due to sudden production line stops. There
cost includes [5]: direct repair or replacement cost, loss of
are several ways for a material to fail, such as inadequate use
revenue while unavailable, costs of finding replacement serv-
specification, misuse (overload) under operation, errors in
ices/items during unavailability, costs of consequential dam-
the thermal treatment, pre-existing and propagating cracks in
age, consequential costs to avoid failure on similar items, and
the equipment working parts, and even the presence of inclu-
other implications, such as safety, loss of confidence, image
sion impurities that make the material more brittle [1-3]. We
and trade.
are sure that many failures are not related to structural integ-
rity problems, such as fatigue and fracture, however, many This paper intends to prevent new failures in a scrap-cutting
of them were originated by these and this may bring tragic machine, from which a fractured piece (gearbox shaft) was
consequences – as for example, in transportation equipment, collected and analyzed, or, in similar situations, with other
aircraft or railways parts, or in the process industries. Fit- material and/or equipment.

Ciência & Tecnologia dos Materiais, Vol. 24, n.º 1/2, 2012 19
V.S. Rocha et al. Fracture failure analysis of gearbox shaft

Failures in shafts are common and many of them have al-

ready been studied [5], but the problems with shafts that
fail by fracture are still usual, which reinforces the need for
dissemination of failure analysis studies. Despite the exist-
ence of a strict quality control for metallic component parts
in equipment and for more specific and efficient mechanical
and chemical analysis, it is not guaranteed that some material
will not fail.

Failure prevention is very important because the failure oc-

currence is not limited to one specific piece of equipment or
material system. For such investigation, some methods and
stages can be used [8,9], like the analysis of the history and
context of the failing piece, the selection and the cleaning of
the parts that will be analyzed, a non-destructive test, a de-
structive test, some observation with microscopy techniques,
an analysis of the results, as well as conclusions and recom- Fig. 1. Image of the fractured surface of the shaft obtained with help of a
mendations. magnifying glass. The scale of magnification was not measured. The marks
displayed on the fracture surface indicate that the onset of fracture occurred
in the region highlighted. The regions 4, 5, 6 and 7 of which were obtained
fractographies are exemplified by the images of Figures 4, 5, 6 and 7, re-
2. Materials and methods spectively.

In this experiment several techniques for mechanical and met- The feeding/supply of the material to be chopped is manual,
allurgic characterization of materials were used in order to and after being deposited into the feeding chute, it is pulled by
evaluate the failure mechanisms. The methodology has used rollers that direct it to the blades (so-called knife and counter
the following steps: Equipment analysis, visual inspection knife), that are mounted on an axis with a 90º radial spacing.
of the fractured part with macro and micrographic images, The material enters between the spinning knives and passes
hardness measurements and Scanning Electron Microscopy between the counter knives in a pre-settled time interval to
(SEM), together with Energy Dispersive Spectroscopy (EDS) assure the 50mm standard length. At this moment a torque
measurements. Figure 1 shows the fractured surface of the of the drive assembly is applied in the same area where the
shaft taken with a digital camera through a magnifying glass. fracture of the shaft has occurred. Figure 2 shows the shaft in
position before the failure.
The machine was used to chop steel cuttings into standard
sizes (SAE 1020 steel-plate edges with thickness ranging be-
tween 2 and 6,35 mm, and width between 10 and 30 mm).

Fig. 2. Shaft before the failure. Assembled on the roller bearing and the gear, and disassembled, respectively.

Table 1 shows the manufacturer scrap compacting machine information. The fractured shaft is a part of a prototype built ma-
chine and it’s the only equipment available in use, developed by this manufacturer.

20 Ciência & Tecnologia dos Materiais, Vol. 24, n.º 1/2, 2012
Fracture failure analysis of gearbox shaft V.S. Rocha et al.

Table 1. The scrap compacting machine’s information: Paral- were carefully preserved and kept isolated from each other for
lel-axis helical reducer (2 stages). the achievement of some important fractographic pictures.

Output Torsional Moment 1.646 N.m / 168 kgf.m The material of the shaft was originally described as SAE
Nominal Reduction 22 1045 steel, tempered and quenched, which is largely used in
Model REDUTORK MOD. TRH04 the manufacturing of motor axes. This choice was based on
the fact, that it has good mechanical resistance, high hardena-
Input Rotation 1750 rpm
bility, high machinability, high forgeability and medium weld-
Capacity (Power) 14 kW / 19,0 cv ability [10,11]. Information about details of thermal treatment
quenched and tempered SAE 1045 used on the material was not available, but these steel classes,
Shaft material
steel which combine high mechanical resistance and hardness, as
Siemens Three-phase Cage for example, tempered and quenched, is not indicated for the
Induction inner parts inside the shafts [12,13]. The commercial hard-
Specification of the engine Rotation: 1755 rpm ness for SAE 1045 quenched by induction can reach values
Power: 7,5 kW / 10 cv that range from 50 to 55HRc (513 and 595 HV, respectively)
Frequency: 60 Hz [14]. There isn’t any former knowledge of previous failures
in this prototype equipment. Mathematical calculations for
The assemblage of the scrap compacting machine was em- life cycle evaluations have not been performed either.
piric. For this reason, adjustments were made, according to
The chemical analysis by EDS – Energy Dispersive Spectros-
the operation needs. Further investigations carried out in situ
copy (Fig. 3) shows that there was a high percentage of iron
have revealed that the power load of the machine (manual and some manganese and also a low percentage of silicon
feed) has frequently exceeded a value of 15% of the amount and chrome, which is expected in the chemical composition
recommended by the original design specification. of SAE 1045. The elements identified by the EDS technique
After the failure, the parts containing the fractured surfaces are in accordance with the chemical composition given by lit-
erature for SAE 1045 steels (Table 2) [15]. Thus, no chemical
were isolated with bubble wrap to avoid contamination. The
modifications were found in the material composition. Figure
first step was the analysis of the history of the piece, the op-
3 exemplifies one of the several spectra obtained from the
eration of the gearbox and visual inspection with macro and
fractured shaft.
micrographic images. In a second step, a sample of the frac-
tured piece (surface of fracture) was observed and analyzed
with scanning electron microscopy analysis (SEM ZEISS™
operating in 20kV and 80μA) and with chemical EDS analy-
sis. Measurements of near and distant points from the frac-
ture were also taken for metallographic procedures and were
exposed to chemical attack of Nital 5% for 5 seconds. This
sample has been observed using a Tabletop Microscope TM
– 1000 Hitashi, with 15 kV acceleration and Back Scattered
Electron (BSE) mode for micrograph images. The microstruc-
ture of the fractured shaft pictures were taken by a ZEISS™
Axiolmager M1m microscope (with a 500x magnification).
Fig. 3. EDS result, significant peaks of Fe and Mn.
Vickers hardness tests were performed using a 98,07N (10kgf)
load and a 15-second-loading time, according to ABNT NBR Table 2. Chemical composition given by literature for SAE
NM ISO 6507-1. Five hardness test measures were taken in 1045 steels (ABNT 1045) [15].
two distinct areas, one near the fracture occurrence, and the
Element Quantity [%]
other in the radial direction of the fractured area.
C 0,43 / 0,50
Mn 0,60 / 0,90
3. Results and discussion P 0,012
S 0,008
A general approach must be taken: it is important to look for
Si 0,2
a general context of the failure and use extra-technical mate-
rial information for further assumptions. Sample evaluation Al 0,02
should be carried out shortly after the failure. Later, at the Cr 0,01
laboratory examination step, SEM observations should be Mo 0,02
carried out. Sometimes, a simple metallurgical analysis might
lead to the conclusion of the problem. In figures 4, 5, 6 and 7 the fracture surface of the material
(fractographies) was observed. Figure 1 shows the pointed
Fracture surfaces are extremely delicate and contain relevant regions 4, 5, 6 and 7 where the figures 4, 5, 6 and 7 have been
information for the failure analysis. In this experiment, they obtained. Images from the right, left, superior and inferior

Ciência & Tecnologia dos Materiais, Vol. 24, n.º 1/2, 2012 21
V.S. Rocha et al. Fracture failure analysis of gearbox shaft

extremities of the fracture of the shaft have been taken, as

well as images from the central area. The fractographies were
very similar along the whole surface of the rupture. Their as-
pects have revealed that there is a brittle cleavage fracture,
which means they have revealed common failures that occur
through planes with low crystallographic index. The surface
of the fracture by cleavage is typically found in brittle materi-
als (or fractures by impact) with low energy absorption ca-
pacity. Its surface morphology is formed by a flat and smooth
region. The brittle fracture by cleavage occurs when the plas-
tic deformation is also hindered (for example, at temperatures
below the DBTT – ductile-brittle transition temperature). It
occurs in a transgranular way, which means, through planes
with lower atomic density, since they have less breaking
bonds and greater interplanar space [3,10,16].

Fig. 6. Surface of fracture of the shaft, 1200X magnification, MEV, SE

mode.

Fig. 4- Surface of fracture of the shaft, 1200X magnification, MEV, second-

ary electrons mode (SE).

Fig. 7. Surface of fracture of the shaft, 1200X magnification, MEV, SE

mode.

A careful investigation of the fractured surface by SEM tech-

nique (Fig. 4, 5, 6 and 7) has shown no evidence of the pres-
ence of the typical marks of fatigue (typical grooves formed
by the advancing crack) [3]. The marks displayed on the frac-
tured surface with a magnifying glass indicate that the onset
of the fracture has occurred in the region highlighted in the
image of Figure 1, but no clue on factors, such as microcraks
on the surface of the material, inclusions or corrosion pits, has
been found in this region with the techniques used. However,
fracture has occurred at the lower cross section of the piece,
reinforcing the concentration of tensions in the region.

Fig. 5. Surface of fracture of the shaft, 1200X magnification, MEV, SE The image on figure 8 exemplifies one of the several images
mode. obtained from the microstructure of the material of the shaft,

22 Ciência & Tecnologia dos Materiais, Vol. 24, n.º 1/2, 2012
Fracture failure analysis of gearbox shaft V.S. Rocha et al.

usually quenched and tempered steel, with a tempered marten- after reaching that temperature, and tempering at 600ºC for
site microstructural aspect. All samples had the same aspects. 100s. These procedures give the SAE 1045 yield strength of
1028 MPa [15]. These would be important parameters to be
considered for control of the quenching and tempering of the
material, however, in this case, the use of a piece with superfi-
cial quench and temper is recommended, instead of this treat-
ment in the whole piece. Such application demands tempered
martensite only in the surface of the shaft.

The analysis of the microstructure revealed that it was

quenched and tempered steel with ferrite and cementite phas-
es, homogeneously distributed, which corroborates the speci-
fication of the material of the shaft from the scrap compact-
ing machine. The use of a piece with superficial quench and
temper is recommended, instead of this treatment in the whole
piece, because an improperly quenched and tempered material
can get more brittle and should never be used in a gearbox.
Such application demands a surface in tempered martensite,
hard, and with a nucleus that maintains its ability to make the
Fig. 8. Microstructure of the fratured shaft – 500X magnification. Every transmission of the torque, when it must deform and act like
analyzed regions had the same microstructural aspects, exemplified by this a torsion spring. The data provided by the manufacturer (Ta-
image. 5% Nital attack.
ble 01) has not revealed details about the thermal treatment,
The hardness measurements in Vickers scale were apparently which are relevant to an analysis of the mechanical behavior of
compatible with the microstructural homogeneity obtained in the shaft. The probable overload situation (due to the manual
the material, because the value HV 10 = 335,72 was frequent. feeding of the machine) and the thermal treatment for quench
(that should be only superficial and was wrongly specified)
The hardness HV 10 values found for the five measured points
have, presumably, restrained the transversal area necessary to
were 335,2; 308,87; 350,47; 335,72 and 335,72. The image on
do mechanical work. The power load of the machine (manual
figure 9 exemplifies one of the several images obtained from
feed) has frequently exceeded a value of 15% of the amount
the microstructure of the material in BSE mode. The BSE im-
recommended by the original design specification.
ages have revealed a homogeneous microstructure behavior,
confirming the results obtained by optical microscopy.
4. Conclusion

Corrosion evidences have not be found, which, at first, dis-

cards the possibility of a failure caused by embrittlement.
However, the brittle characteristic can be associated with an
empiric and uncontrolled quench, since the pieces are usually
only superficially quenched.

As it was supposedly a SAE 1045 steel, the chemical analysis

have assured that the existing chemical elements are equiva-
lent to the predetermined commercial values for this type of
steel. This material had no chemical element that could have
caused brittleness, since the quantity of silicon and chrome
found was not significant. These results have been confirmed
by the constant hardness values obtained through the sample
cross-section and the BSE images.

All observed factors and the probable operation in overload,

Fig. 9. Microestruture of the fractured shaft, 5000X magnification, Tabletop
due to manual scrap feeding, lead to the conclusion that it
MEV, BSE mode. Every analyzed regions had the same microstructural as- was a failure caused by misuse (overload), associated with
pects, exemplified by this image. a microstructure that should be only superficial, instead of
along the axis. The analysis showed that the dimensioned ma-
It is known that the values of mechanical properties, such terial was only considered to operate in normal conditions,
as hardness, tensile strength and yield strength may vary ac- excluding critical overload conditions. However, the material
cording to the conditions in which thermal treatments are per- has been considered not suitable for mechanical applications
formed, such as quenching and tempering. The austenitizing with normal operational conditions, since the fractured part
temperature and tempering time are important parameters to had been submitted to an inadequate quenching. As a final
consider. The optimization of the mechanical properties of the suggestion, the scrap feeding into the equipment should be
SAE 1045 steel, with high strength and ductility, is obtained at performed automatically, in order to minimize the mechani-
an austenitizing temperature of 900ºC, immediately quenched cal stress.

Ciência & Tecnologia dos Materiais, Vol. 24, n.º 1/2, 2012 23
V.S. Rocha et al. Fracture failure analysis of gearbox shaft

REFERENCES [9] Y.J. Chen, M.A. Meyers, V.F. Nesterenko, Materials

Science and Engineering A, 268 (1999) 70-82.
[1] C. Gandhi, M.F. Ashby. Acta Metallurgica, 27 (1979)
669-729. [10] D.R. Askeland, “The science and engineering of materi-
[2] W.A. Mannheimer, “Microscopia dos materiais uma in- als”, ed. PWS Publishing Company, 1994.
trodução”, ed. E-papers, Rio de Janeiro, 2002.
[11] S.A. Souza “Composição química dos aços”, ed. Edgard
[3] American Society for Metals, “Metals handbook vol- Blücher, São Paulo, 1989.
ume 12 fractography”, ed. American Society for Met-
[12] W.D. Callister Jr, “Materials science and engineering:
als, USA, 1987.
an introduction”, ed. John Wiley & Sons, New York,
[4] P.M.S.T. de Castro, F.M.F. Oliveira, A.A. Fernandes, 1991.
Technology Law and Insurance, 1 (1996) 143-150.
[13] D.R. Askeland, P.P. Phulé, “Ciência e engenharia dos
[5] P.M.S.T. de Castro, A.A. Fernandes, Materials and De- materiais”, ed. Cengage Learning, São Paulo, 2008.
sign, 25 (2004) 117-123.
[14] V. Chiaverini, “Tratamento térmico das ligas metáli-
[6] E. Poursaeidi, M. Aieneravaie, M.R. Mohammadi, En- cas”, ed. Associação Brasileira de Metalurgia e Mate-
gineering Failure Analysis, 15 (2008) 1111-1229. riais, São Paulo, 2003.

[7] M. Krishnadev, M. Larouche, V.I. Lakshmanan, R. [15] C.R. Stein, T.F. Hermenegildo, A.B. Cota, REM-Revis-
Sridhar, Engineering Failure Analysis, 15 (2008)1009- ta Escola de Minas, 58 (2005) 51-56.
1018.
[16] American Society for Metals, “Metals handbook vol-
[8] G. Atxaga, A.M. Irisarri, Engineering Failure Analysis, ume 9 metallography and microstructure”, ed. Ameri-
17 (2010) 714-721. can Society for Metals, USA, 1989.

24 Ciência & Tecnologia dos Materiais, Vol. 24, n.º 1/2, 2012