Engineering Failure Analysis 17 (2010) 802–809

Contents lists available at ScienceDirect

Engineering Failure Analysis

journal homepage: www.elsevier.com/locate/engfailanal

Metallurgical failure analysis of the 560 mm dia 0.15C–1.25Cr–1Mo–0.25V

steel pressure vessel
Abhay K. Jha a,*, K. Sreekumar a, P.P. Sinha b
a
Material Characterisation Division, Materials and Metallurgy Group, Vikram Sarabhai Space Centre, Indian Space Research Organisation,
Trivandrum 695 022, India
b
Materials and Mechanical Entity, Vikram Sarabhai Space Centre, Indian Space Research Organisation, Trivandrum 695 022, India

a r t i c l e i n f o a b s t r a c t

Article history: 0.15C–1.25Cr–1Mo–0.25V steel, a high strength low alloy (HSLA) steel, is being extensively
Received 18 June 2009 used in space programme. This alloy being a versatile material and its ease of fabrication
Accepted 11 October 2009 and welding made it the most suitable for the fabrication of 560 mm dia pressure vessel
Available online 31 October 2009
for smaller rockets. During proof pressure test for 73 KSC, one of the pressure vessels burst
opened at 71 KSC. Metallurgical analysis of the failed motor case was carried out to under-
Keywords: stand the causes of its failure. Observations indicated that a deviation in normal heat treat-
0.15C–1.25Cr–1Mo–0.25V steel
ment resulted in low strength of material, which caused yielding and premature rupture at
Burst failure
lower pressure than expected. This paper brings out the salient features of this analysis.
Ó 2009 Published by Elsevier Ltd.

1. Introduction

High strength low alloy (HSLA) steel with nominal composition of 0.15C–1.25Cr–1Mo–0.25V is being used in space
programme extensively. This alloy is a versatile material and its ease of fabrication and welding made it the most suitable
for the fabrication of 560 mm dia motor case for smaller rockets. During proof pressure test designed to withstand 73 KSC,
one of the pressure vessels, fabricated using this material burst opened at 71 KSC. Metallurgical analysis of the failed com-
ponent was carried out to understand the causes of its failure. This paper brings out the salient features of this analysis.

2. Material

The 0.56 m dia pressure vessels are fabricated using this HSLA steel sheets of dimension 2000  1000  2.60 mm,
procured in annealed condition. All the sheets were produced from four ingots of the same heat in different batches. These
sheets satisfied the chemistry of the material and met the specified properties in heat-treated condition. The specified chem-
istry and mechanical properties in hardened and tempered condition are given in Tables 1 and 2 respectively.
The cold rolled and annealed sheets were converted into 0.56 m dia cylindrical segments by roll bending and joining the
edges by TIG welding. Four such segments were joined end to end to form a pressure vessel of 2.855 m length by circum-
ferential TIG welding after hardening and tempering to the specified properties. Three rings were used, one at both ends
(dome side and nozzle side) and the remaining in the centre to provide necessary rigidity as per design requirement. The
sheets from two different lots of the same heat, say batches ‘A’ and ‘B’ were used for fabrication of pressure vessel. The sheet
used for the segment which failed belonged to batch ‘A’ while the another segment of the same pressure vessel, made out of
sheet of batch ‘B’ did not fail during the proof pressure test. The schematic of the motor case is shown in Fig. 1.

* Corresponding author. Tel.: +91 471 2563628; fax: +91 471 2705048.
E-mail address: [email protected] (A.K. Jha).

1350-6307/$ - see front matter Ó 2009 Published by Elsevier Ltd.

doi:10.1016/j.engfailanal.2009.10.012
 A.K. Jha et al. / Engineering Failure Analysis 17 (2010) 802–809 803

Table 1
Chemical composition of material, as specified.

Elements %C %Cr %Mo %V %Fe

wt.% 0.12–0.18 1.25–1.50 0.80–1.0 0.2–0.3 Rest

Table 2
Mechanical properties of material in hardened and tempered condition.

Mechanical properties UTS (MPa) min. 0.2% Y.S (MPa) min. % E, min. (50 mm GL) Hardness (BHN), min.
Austenising at 925°C – soak for 20 mts – AC and 980 835 9 300
tempering at 640°C – soak for 30 mts – AC

Fig. 1. Schematic sketch of the motor case.

3. Observations

3.1. Visual observations at failure site

During Proof Pressure test, the motor case failed at 71 KSC as against the proof pressure of 73 KSC. Preliminary observa-
tion, made at the test site revealed that the material of one out of total four segments of the motor case, yielded near the long
seam weld (15–30 mm away from the weld fusion line) and failed in overload mode (Fig. 2). The opening of the crack of
length 790 mm was the maximum (80 mm) at its centre. The propagation of the crack stopped almost 27 mm away from
the cirseam weld fusion line on either side as shown in Fig. 3. Four dents, two on either side of the crack opening and almost
equidistance (approx. 200 mm) from the fracture path (Fig. 4) were also noticed. These dents were due to inward flow of
material that resulted from outward bulge of material at fracture edge on yielding during proof pressure test.
The circumferences of the failed segment were measured at three locations, i.e. at the centre of fracture path and at both the
cirseam welds on either side of the crack opening. The circumference measured was 1900 mm at the centre of fracture while it
was 1765 mm at both the cirseam welds. The difference of 135 mm was a measure of bulging during yielding of material.

Fig. 2. The failed motor case.

804 A.K. Jha et al. / Engineering Failure Analysis 17 (2010) 802–809

Fig. 3. Crack opening.

Fig. 4. Dents on the failed hardware.

The thickness measurements were carried out at locations with an interval of approx. 10 mm all along the fracture edge.
Thickness was also measured at a distance of approximately 50 mm away from the fracture edge at certain other locations,
on either side of the fracture edge. A pictorial depiction of locations and measured thickness is given in Table 3. Thinning of
the material was observed all along the fracture edge from an original reported thickness of 2.52/2.60 mm. The measured
minimum thickness of the shell at the centre of the fracture edge was 2.00 mm. However, the thickness of the shell at about
50 mm away from the fracture edge was as low as 2.30 mm. This confirmed that material underwent yielding before final
fracture, though, thinning to this extent was not expected for this material in hardened and tempered condition. Taking into
consideration this large reduction in thickness of the shell, hardness measurement was conducted on pieces sliced off from
the failed segment near to the fractured edge. The hardness values observed were in the range of 164–171 BHN.
These values of hardness were lower than the expected minimum hardness of 300 BHN for this steel in hardened and
tempered condition. Hence a detailed investigation to the metallurgical nature of the material of the segments of the failed
motor case was found necessary.

3.2. Specimens selection

Specimens, details of which are given below, were sliced off from the failed hardware at the manufacturer site for this
purpose.

1. A cut piece of size 160  130 mm taken out from the failed segment and at location opposite to the opened surface sheets
belonged to batch ‘A’).
2. A cut piece of size 320  350 from the segment, which was part of the motor case but did not fail (sheets belonged to
batch ‘B’).
3. A cut piece of size 70  120 mm from the cirseam weld with the weld bead at the centre and material from the failed
segment (batch ‘A’) and from good segment (batch ‘B’) on either side.
 A.K. Jha et al. / Engineering Failure Analysis 17 (2010) 802–809 805

Table 3
Pictorial depiction of locations for thickness measurement and values.

Thickness, 50 mm away Thickness of fracture edge at Pictorial view of fracture Thickness of fracture edge at Thickness, 50 mm away
from fracture edge interval of 10 mm edge interval of 10 mm from fracture edge
2.20 2.20
2.21 2.20
2.20 2.20
2.31 2.20 2.21 2.52
2.20 2.24
2.20 2.21 2.42
2.31 2.24 2.18
2.20 2.24 2.45
2.47 2.20 2.20
2.20 2.19 2.38
2.48 2.20 2.11
2.20 2.17 2.47
2.21 2.21 2.45
2.42 2.20 2.20
2.20 2.21
2.41 2.20 2.20 2.40
2.20 2.19
2.19 2.12
2.35 2.18 2.08
2.12 2.05
2.30 2.05 2.00 2.30
2.02 2.08
2.34 2.00 2.05
2.01 2.08 2.36
2.05 2.12
2.32 2.08 2.13
2.12 2.18
2.14 2.20 2.32
2.39 2.15 2.27
2.14 2.05 2.33
2.38 2.15 2.13
2.15 2.15 2.31
2.38 2.15 2.14
2.14 2.06
2.16 2.10 2.30
2.36 2.17 2.11
2.17 2.12
2.36 2.18 2.12
2.20 2.22 2.33
2.19 2.20
2.19 2.11
2.32 2.18 2.12
2.20 2.20

The chemical assay, mechanical properties and microstructural studies were carried out on these pieces. Apart from this,
different heat treatment cycles were carried out on the material from the failed segment to evaluate its response. Hardness
traverse on the cut piece consisting of sheets from batch ‘A’ and that from batch ‘B’ with cirseam weld in between was also
carried out and the values was reported in Table 4.

3.3. Chemical analysis

Chemistry of the material of the failed motor case was evaluated at various locations. The details of the chemical analysis
are furnished in Table 5.
The Specimen 1, from the fractured segment (sheet of batch ‘A’) had a carbon content of 0.11 wt.%. The chemical con-
stituents of the segment (sheet of batch ‘B’), which did not fail, were within specified limits (Specimen 2). The low carbon
content of failed segment was thought to be due to de-carburised layer as the chemistry with respect to the other alloying
additions were within the specified values. Presence of de-carburised layer was subsequently confirmed by optical micros-
806 A.K. Jha et al. / Engineering Failure Analysis 17 (2010) 802–809

Table 4
Hardness values evaluated on both side of cirseam (CS) weld.

Specimen identification Hardness (BHN)

Failed segment (sheet of batch ‘A’) 161,163,164
Sheet (batch ‘A’) on one side of CS weld 162,165,166
Sheet (batch ‘B’) on other side of CS weld 303,304,305

Table 5
Chemical analysis result.

%C %S %Mn %V %Cr %Mo %Fe

Spec 0.12–0.18 0.015max 0.80–1.0 0.2–0.3 1.25–1.50 0.80–1.0 Bal
Specimen 1 0.11 0.001 0.72 0.22 1.30 0.80 Bal
Specimen 2 0.17 0.002 0.70 0.22 1.30 0.81 Bal
Specimen 3 0.15 – – – – – –

copy. Hence, carbon content was evaluated on chips drilled out from another specimen from the same location, after about
0.3 mm material was removed from its both surfaces. After removing materials from both the surfaces, Specimen 3 was
taken out and evaluated for carbon content. The carbon was found to be 0.15 wt.%, well within the specification (0.12–
0.18%C).

3.4. Metallurgical investigations

Microstructural analysis of specimens from locations near to and away the fractured edge of the failed segment and also
from the segment which did not fail was done by conventional metallography with 5% Nital as the etchant to reveal the
microconstituents under an inverted optical metallograph.
The microstructures of specimen adjacent to the fracture edge revealed predominantly polygonal ferrite with fine islands
of peralite. Plate type M7C3 carbides along the grain boundaries of ferrite and disc shaped M6C type of carbide within ferritic
grains were present. These features were similar to that of annealed HSLA steel.
Specimen from the segment of the failed hardware, which did not fail, revealed presence of lower banite as a major con-
stituent with some islands of martensite and massive ferrite, typical of hardened and tempered steel. The typical microstruc-
tures at various locations of motor case and their corresponding hardness are shown in Fig. 5.
The specimens which consisted of both the material, i.e. material from the failed segment (sheet of batch ‘A’) and
from the good segment (sheet of batch ‘B’) with a portion of cirseam weld in between them, was polished across the
thickness and analysed for its microstructural inhomogeneity. Presence of de-carburised layer was seen on either side
of the material from the failed segment (Fig. 6). The depth of this de-carburised layer was estimated to be 170 lm
and 260 lm on either surface of the specimen. The material from the good segment, i.e., the one on the other side of
the weld pool, did not have any de-carburised layer. It had the microconstituents typical of HSLA steel in hardened
and tempered condition. The schematic of micro hardness traverse across the thickness of the specimen is given in
Fig. 7. The measured microhardness values are furnished in Table 6. The microstructure and hardness traverse across
the thickness confirmed the presence of de-carburised layer to a depth of 0.3 mm (max) on both ID and OD of the
failed segment. The marginally low carbon content of specimens initially taken from this segment was due to the pres-
ence of this de-carburised layer.
Hardness measurement was carried out on the lateral surface of the above specimen, which were in agreement with
earlier values. Specimens for mechanical properties evaluation were fabricated from cut pieces taken out from the failed
segment and the segment, which was part of motor case but did not fail. Mechanical properties were evaluated using
Instron and values obtained are furnished in Table 7. The failed segment showed yield strength as low as 580 MPa as
against required value of minimum 835 MPa, whereas another segment showed values more than that of specified
values.
From above it was inferred that the segment which failed was not properly heat treated. This caused lower strength of
material, and resulted in yielding and premature rupturing during proof pressure test of motor case.
To simulate the possible heat treatment deviation, various heat treatment trails were carried out on this material. The
heat treatment, which yielded very similar observations of the failed one was heating to 975 °C holding at 20 min and then
slowly furnace cooled to 650 °C before they were taken out from the furnace. Simulated study of this heat treatment had
shown hardness value in the range of 182–185 BHN and microstructure consisted of ferritic grains with pearlitic islands.
De-carburised layer of depth in the range of 140–170 lm was also measured while viewing of cross section of the specimen
 A.K. Jha et al. / Engineering Failure Analysis 17 (2010) 802–809 807

Fig. 5. Typical microstructures at various locations and typical hardness values.

(Fig. 8). The grains within de-carburised region were very fine compared to that away from the surface. This is attributed to
the working imparted during final reduction of sheet making.
The above exercises showed that the fractured segment had deviation during its heat treatment, which resulted in low
strength nearly equal to that of annealed material. Low strength of the sheet caused premature overload failure during proof
pressure test.
808 A.K. Jha et al. / Engineering Failure Analysis 17 (2010) 802–809

Fig. 6. Decarburised layer across the thickness of failed segment.

Fig. 7. Schematic sketch of specimen from the cirseam weld region and hardness (VPN) values across its thickness.

Table 6
Microhardness values across the thickness.

Sheet of batch ‘A’ (failed segment) Sheet of ‘B’

Hardness in VPN along X–X
154 (ID) 265 (ID)
157 317
229 362
236 380
262 380
265 362
202 262
177 (OD) 328 (OD)

Table 7
Mechanical properties.

Specimens taken from UTS (MPa) 0.2%YS (MPa) %E Hardness BHN

Specification 980 min. 835 min. 9 min.
Failed segment 777 610 23.9 168,169,173
766 594 19.4 169,170,179
767 580 19.4 172,173,168
Segment, which did not fail 1128 1002 14.5 336,338,329
1132 993 14.8 337,340,337
1124 994 13.0 329,329,329
 A.K. Jha et al. / Engineering Failure Analysis 17 (2010) 802–809 809

Fig. 8. Ferritic and pearlitic grains and de-carburised layer.

4. Discussion

The primary objective of a material failure analysis is to determine the root cause of the failure, which can be normally
assigned to various categories e.g. design, manufacturing, material or service environment.
In the present failure case study, the burst opening of the motor case was confined to a segment, while another three seg-
ments of the motor case were intact. The opening was maximum at the centre and narrowed down while traversing towards
both the ends. Presence of dents, two on either side of fracture edge and almost equidistant from the fracture edge were due
to inward flow of materials resulted from outward bulge of material at fracture edge on yielding during proof pressure test.
Thickness measurement indicated thinning from 2.52/2.60 mm to 2.00 mm at the middle of fracture edge, which was
much more than the expected thinning for hardened and tempered steel. The measured low hardness indicated that the
material was in annealed condition. Further, the de-carburised layer too had contributed to low hardness, which was con-
firmed by chemical analysis and microstructure. This lower value of hardness was within the specification (<197BHN) for
15CDV6 steel in annealed condition. Mechanical properties and microstructure of this segment further confirmed this.
The high hardness, as observed in sheet of another segment of the same motor case was result of proper hardening and tem-
pering treatment.
The deviation in heat treatment was established by simulated heat treatment.
The calculation (Appendix A) shows that the material with deviation in heat treatment as inferred above and which had
yield strength of 580 MPa could yield at 71 KSC. The calculation also supports why Segments fabricated with sheets (batch
‘B’) having YS more than 990 MPa and UTS more than 1120 MPa did not fail.

5. Conclusions

1. The segment fractured in pressure testing had deviation in heat treatment while the adjoining good segment was in the
specified hardened and tempered condition.
2. Deviation in heat treatment resulted in low strength of material, which yielded and ruptured at lower pressure than
expected.

Acknowledgements

The authors are indebted to Dr. K. Radakrishnanan, Director, VSSC for permission to publish this work.

Appendix A

Arriving conclusion becomes simpler as per following calculation:

The calculation shows that the material with deviation in heat treatment as inferred above and which had yield strength
of 580 MPa could yield at 71 KSC.
The evaluated YS of material from the failed segment ry = 580 MPa
The segment failed at 71 KSC, i.e. P = Pr = 71 KSC
Diameter of the segment = 560 mm = 56 cm
Thickness of the segment = 2.50 mm = 0.25 cm

rmax ¼ PD=2t ¼ ð71  56Þ=ð2  0:25Þ ¼ 7952 KSC ¼ 79:52 Kgf=mm2 ¼ 779 MPa
where rmax is the maximum stress experienced by material during pressure test at 71 KSC.
rmax = 779 MPa, which was more than the YS (580 MPa) and UTS (766 MPa) of the material and hence failure by rupture.
Segments fabricated with sheets (batch ‘B’) having YS more than 990 MPa and UTS more than 1120 MPa did not fail.