34th INTERNATIONAL CONFERENCE ON
PRODUCTION ENGINEERING
28. - 30. September 2011, Niš, Serbia
University of Niš, Faculty of Mechanical Engineering
INFLUENCE OF PRODUCTION PROCESS ON FATIGUE PROPERTIES OF HEAVY
CASTINGS - A CASE STUDY
1)
Radivoje MITROVIĆ1), Dejan MOMČILOVIĆ2), Olivera ERIĆ3), Ivana ATANASOVSKA3)
Faculty of Mechanical Engineering, University of Belgrade, Kraljice Marije 16, Belgrade, SERBIA
2)
Institute for Testing of Materials, IMS, Bulevar vojvode Mišića 43, Belgrade, SERBIA
3)
Institute Kirilo Savic, Bulevar vojvode Stepe 51, 11000 Belgrade, Serbia,
[email protected], [email protected], olivera.erić@iks.rs, [email protected]
Abstract: This paper describes the analysis of production process on failure of heavy flange casting of hydro
turbine shaft. Special emphasize on metallurgical failure analysis of in-service crack initiation is presented
in this paper. The presence of dendrite structure points to incomplete or irregular flange heat treatment.
Testing the mechanical properties of the material, revealed slight discrepancies between the experimental
results and the requirement of the specification for the production of the flange material.
The conclusion of this research shows that production process resulted of inhomogeneous microstructure
and subsequent data scatter. This indicated that the heat treatment regime, as a part of production process,
had some influence in the fatigue fracture of flange casting.
Key words: casting, microstructure, fatigue, fracture
1. INTRODUCTION
This paper describes the partial investigation of turbine
shaft Kaplan’s 28 MW bulb turbine failure. The bulb
turbine generator’s horizontal shaft, Fig. 1, is made by
joining the forged and cast parts by slag welding, Fig. 2.
The shaft is manufactured as hollow, housing a
servomotor inside it, for shifting the runner blades. The
flange, on which the crack occurred, is made of steel
casting of 20GSL designation, according to GOST 97788, [1]. The operating speed of turbine shaft was 62.5
rpm. The fact that cracks were found on heavy casting
(approx weight 20 tons) raised question of origin of
failure in casting and subsequent heat treatment process.
Fig. 2. General appearance of the cracked zone
2. EXPERIMENTAL RESULTS
2.1. Chemical Composition and Mechanical test
Table 1 shows the chemical composition of flange
material in accordance with the requirements of the
reference standard, [1]. The testing determined that the
chemical composition meets standard’s requirements.
Fig 1. Position of crack on general assembly of hydraulic
turbine shaft
General appearance of the cracked zone of 20GSL heavy
casting is given at Fig 2. The close view of zone with
cracks is given at the Fig 3.
Table 1. Chemical composition of flange
%C
%Si
%Mn
%S
Min
0,16
0,60
1,00
Max
0,22
0,80
1,30
0,030
%P
0,030
Testing of the fatigue strength was completed according
to the requirements of standard, [1], in open-air
conditions, at the +20oC, at load ratio R = -1. Fatigue
limit SFL of the shaft flange material was tested on
ZWICK ROELL HB 250 and it is 168.0 MPa. The results
of fatigue tests are shown on table 3, and the S-N diagram
on Figure 5.
Table 3. Average and reference fatigue values of the
20GSL flange steel casting, [1].
Average
Reference values
values
Fatigue limit
on air
in water
168
SFL (MPa)
225
140
2
Table 2. Average and required mechanical properties at
room temperature of the 20GSL steel casting, [1].
Required
Average
(GOST 977)
values
- min. values
Yield strength (MPa)
310
294
Tensile strength (MPa)
509
540
Elongation, in 2 in. (%)
17.6
18
Reduction of area (%)
35.2
30
Brinell hardness HB
153
Charpy-V notch, +20oC (J)
74.4
23.4
2.2. Fatigue properties
Stress, S (N/mm )
During cutting of samples for the chemical, mechanical
and metallographic tests, numerous gas holes and pores
were detected that resulted from the casting process, Fig.
4. Gas pores and holes were not perceived around the
crack initiation location, i.e. around the outer shaft
surface, but exclusively in the casting volume. The testing
of casting meets standard’s requirements, [1]. The zone of
cutting out of all samples for the mechanical and
metallographic tests was 20 mm below to the external
shaft surface. The results of mechanical test are given at
the Table 2.
300
280
260
240
220
200
180
160
140
120
100
80
60
40
20
0
2
din = 168 N/mm
ZWICK ROELL HB 250
4
10
5
6
10
10
Number of cycles, N
a)
Fig. 3. Close view of the transient flange radius after
detecting the crack
a) Main crack, b) anti-corrosive protection layer, c)
oil residue on corrosion pits
b)
Fig. 4. Gas holes and shrinkages on casting, zone near
flange radius R80
Figure 5. S-N diagram for 20GSL Casting
a) S-N diagram
b) Position of samples on testing machine
7
10
2.3. Metallographic inspection
By reviewing the original documentation from the time of
delivering and installing of the shaft, it was discovered
that heat treatment of the shaft was done by applying the
regime shown of Fig. 6:
b)
Fig 6. Applied heat treatment regime
Such a heat treatment regime indicates that complete
austenization was done, with complete breaking of the
microstructure that had remained from the casting
process.
However, the microstructure of the material is the cast
ferrite-pearlite one, with oxide type non-metallic
inclusions and with minimal participation of dendrite
structure, Fig. 7a). Dendrite structure presence points to
the possibility of incomplete or irregularly executed shaft
flange heat treatment, which explains the observed scatter
of results in all the mechanical tests. It was also
determined that there were large non-metallic oxide
inclusions in the structure, which occurred in sequences.
The fine micro cracks were found in microstructure, Fig 7
b). The results of quantitative metallography are given at
the Table 4.
Table 4. Results of quantitative metallography
Grain size Number of
Grain
Standard
scale
grains/mm2 diameter
ISO 643:2003
7
1204
32µm
GOST 5639-82
8
2048
22µm
a)
Fig. 7. Microstructure of flange material
a) Cast, ferrite-pearlite microstructure, x 200, 5% nital
b) Micro cracks emanating from shrinkage, x 500,
5% nital
3. DISCUSSION
The chemical composition of the tested shaft flange
material corresponds to the requirements of the GOST
977-88 standard, [1]. The fact that some of the obtained
elongation values were lower than the minimal required,
shows that there are locations in the microstructure that
have reduced ability for plastic deformation. So, in these
locations easier occurrence of initial cracks can be
expected.
The established tensile properties, specifically tensile
strength and elongation values, are lower than required
ones, i.e. the tested steel casting sample does not fully
meet the tensile characteristics requirements stated in the
GOST 977-88 standard, [1].
From the point of view of toughness, the presented and
analyzed results of testing the notch impact strength
shows that the material meet the requirements of the
reference standard, [1].
The results of micro-structural analysis in light
microscope show that the microstructure was cast fine
grained, ferrite-pearlite microstructure with non-metallic
oxide-type inclusions and with sporadic dendrite-structure
participation. The presence of micro cracks is also
noticed.
According to the heat treatment regime stated by the shaft
manufacturer, it could be expected that complete
austenization was done, with breaking the dendrite
microstructure that had remained from the casting
process. However, the obtained microstructure test
results, shown on Fig 7, demonstrate that some portion of
residual dendrite microstructure has remained. The
presence of dendrite structure points to incomplete or
irregular flange heat treatment. Moreover, microstructure
and subsequent data scatter indicated that the heat
treatment regime had some influence in the fracture
process.
Such non-homogeneous microstructure, together with the
observed porosity and micro cracks, has contributed to
faster propagation of the fatigue crack, compared with the
case of the flange material that completely fulfill
requirements, [1]. The fact that was spotted significant
data scatter, shown on Fig 5a), also indicates the influence
of casting defects on fatigue properties of casting. On Fig
8, it is shown one of fatigue test specimens with cracks,
prematurely cracked during fatigue test. This figure
clearly demonstrates that the casting defects was the weak
links during fatigue crack growth on flange. The result of
few specimens like the specimen shown on Fig 8, was not
incorporated into diagram on Fig 5a).
-
-
Periodical renewal of the anticorrosive protection on
the shaft flange, especially in the shaft-flange
transition zone,
Keeping the start/stop cycles at minimal values as it
is possible,
The key recommendation is the improvement of
general technical conditions of delivery for flange
material upon a new commissioning [3, 4].
Acknowledgements
This research was supported by the Serbian Ministry of
Education and Science, Project TR 35029 - Development
of Methodology for Improvement of Operational
Performance, Reliability and Energy Efficiency of
Machine Systems used in the Resource Industry
REFERENCES
Fig. 8. Premature failure of test specimens due to casting
defect - tested by liquid penetrants
According to [1], fatigue limit at R=-1 in open air is 225
MPa, and 140 MPa in water. However, by testing the
shaft material samples, the value of 168 MPa of fatigue
limit in open air was obtained, which is ≈25% less than
the reference value of 225 MPa.
4.
CONCLUSIONS
The chemical composition of flange casting met
requirements of the relevant standard.
The fact that 20GSL flange casting mechanical properties
does has not completely fulfilled the requirements [1,2],
indicate that there is a gap in production process, casting
and heat treatment that led to lower properties compared
to required ones.
Fine grained microstructure indicate that heat treatment
was done properly, which points to casting process as a
major point which induce lower mechanical properties.
However, the reasons for the premature failure can’t be
completely attributed to the material quality and the heat
treatment.
The shaft failure occurred due to the combination of
several factors:
- inappropriate corrosion protection in the zone of
critical radius and lack of procedures of renewing
corrosion protection of turbine shaft.
- high stresses during start/stop cycles and during
regular operating regime in the zone of R80 radius
for “wet” environment.
The following suggestions emerged from this production
process based failure analysis:
- Redefining of procedures for periodical nondestructive inspection of the shaft-flange transition
zone status, with increased frequency,
[1] GOST 977-88, Steel Catings. General Specification
[2] ТРОЩЕНКО, В.Т. (1987) Сопротивление
усталости металлов и сплавов, Наукова думка,
pp 668-670 (in Russian)
[3] MARICIC, T., HABER, D., PEJOVIC, S. (2007)
Standardization as Prevention of Fatigue Cracking
of Hydraulic Turbine-Generator Shaft, IEEE
Electrical Power Conference, Montreal, CD
[4] ATANASOVSKA, I., MITROVIĆ, R.,
MOMČILOVIĆ, D. (2011) FEM model for
calculation of Hydro turbine shaft, Proceedings – the
Sixth International Symposium KOD 2010, 2930.09.2010., Palić, Serbia, Published by Faculty of
Technical Science – Novi Sad, Serbia, pp.183-188