Failure Analysis Case Studies II Episode 9 potx

Fracture analysis of outer layer Opening the outer layer fractures and observing by SEM, it was concluded that cracks initiated from the outer surface and propagated towards the inner

(1) The steel wire layer was intact

(2) In all A type hosepipes, there was only one opening in the inner rubber layer Similar results

were obtained for most B type hosepipes except two, where two openings were observed

(3) The opening in the inner rubber layer was situated in the outside of the bend near the metal elbow adapter The distance between the opening and the metal core is about 2.0-12.5 mm (see Fig 2)

(4) The openings in the inner rubber layer are circumferential, as shown in Fig 2

( 5 ) No other defects were observed in the inner layers of the hosepipes

269

A type hosepipe

B type hosepipe

Fig 2 The openings in the inner rubber layer

(6) There were small openings in the outer surface of the inside middle layer in one B type hosepipe (see Fig 3)

4 Fracture analysis of layers

4.1 Fracture analysis of inner layer

Opening the fracture and inspecting by SEM, it was found that the fracture surfaces in both A

and B type hosepipes have the same characteristics, showing flats and radiating ridges (see Figs 4

and 5) Fatigue markings in the later period of propagation could be clearly observed (Fig 6) The

270

Fig 3 The openings in the outer surface of the inside middle layer

Fig 4 The fracture appearance of an A type hospipe

fatigue cracks were possibly initiated at the interface between the cotton thread and the inner layer

Finally, there were no obvious defects in the fatigue initiation zone (Fig 5)

Smaller cracks had similar fatigue characteristics (Fig 7)

4.2 Fracture analysis of outer layer

Opening the outer layer fractures and observing by SEM, it was concluded that cracks initiated

from the outer surface and propagated towards the inner surface Multiple initiation sites and some fatigue striations were observed (Fig 8)

27 1

~ _ _ _ _ _ _ _ Fig 5 The fracture appearance of a B type hosepipe

Fig 6 Fatigue beach marks

5 Analysis of hosepipe bursting failures

Fracture analysis shows that hosepipe bursting is caused by fatigue failure, but the characteristics

of failure are different in the inner and outer layers

In structure, the hosepipe is composed of seven layers (Fig 9) Steel wire is the main load- bearing layer The inside layer was projected from the steel wire by a middle layer and a cotton layer internal to the steel wire layer The inner layer was in direct contact with kerosene and carried

272

Fig 7 The fracture appearance of small cracks

'liP.V- mi-

Fig 8 The fracture appearance of the outer layer

alternating stress caused by pressure variations The outer layer bears stress and bursts after the inner layer bursts When kerosene has penetrated the inner and middle layers as well as the steel wire layer, kerosene gathered between the steel wire and the outer layer Large fatigue cracks initiated in the most highly bent position of the outer layer and formed the multiple fatigue crack sources

There were three openings in the inner layer of one B type hosepipe; two of them were fully penetrating and the other was not Fatigue cracks originated from the interface between the cotton

273

I inner layer 2-cotton 3 mnddlc layer 4-stccl wrc laycr

5-middle layer 6 conon 7-outer layer

Fig 9 Sectional diagram of the bursting hosepipe structure

layer and the inner layer The two penetrating cracks propagated both inwards and outwards The non-penetrating crack outwards

From the above-mentioned, the course of hosepipe bursting failure can be described as follows: fatigue cracks originated first from the interface between the cotton layer and the middle layer, and then propagated into the middle layer Afterwards, further fatigue cracks initiated at the interface between the cotton layer and the inner layer and propagated into the inner layer When the crack had penetrated the inner layer under the steel wire layer, the outer layer could not bear the pressure of the kerosene, resulting in a large stress fatigue failure leading to bursting

Improper design is the main reason for inner layer fatigue failure The interface between the cotton and the inner layer is a weak position Under working conditions, pressured pulses cause a radial bulge in the hosepipe At the same time, there were bending deformations in the hosepipe Insufficient fatigue resistance of the hosepipe is the most important reason for failure However, only static pressure requirements were demanded for the hosepipes

Serious bending deformation in mounting is another important reason for hosepipe bursting Limited by the space, bending deformation could not have been avoided in mounting the hosepipes

It was reported that the minimum bending radius in a high pressure hosepipe is about 6-7 times

the external diameter [I] The pressure capacity of the hosepipe will drop rapidly if the radius is too small When the hosepipe works in normal conditions the working life will drop For example, the service pressure of A style hosepipes is 700 MPa When the bending radius is 70% of the

minimum required in mounting only 58% of the rated working pressure of the hosepipe can be applied in service, i.e., 406 MPa (Fig IO) If the working pressure is kept at 700 MPa, the working life is shortened to 12% (Fig 10)

274

the highest pressure

Fig 10 Relation between bent radius, pressure capacity and working life of hosepipe

( 3 ) The main reasons for hosepipe failure are as follows:

(a) inadequate structure;

(b) low fatigue resistance;

(c) severe bending deformation in mounting

Reference

[I] Beijing Institute of Aeronautical Materials, Aeronautical Material Science, Shanghai Science and Technology Press,

1985, p 6

Environmental attack

277

FAILURE OF AUSTENITIC STAINLESS STEEL

COMPONENTS USED IN NITROGEN OXIDE PLANT

S ARUMUGHAM and T S LAKSHMANAN*

Material Characterisation Division, Materials and Metallurgy Group, Vikram Sarabhai Space Centre,

Thiruvananthapurarn, 695 022, India

(Received 2 1 March 1997)

Abstract-Austenitic stainless steel components of a nitrogen oxide plant have been found to leak in service The failed components, namely pipe-to-pipe joints and pipe-to-flange joints, have been studied through standard metallographic techniques to analyse the cause of the failure In case study I, involving the failure of

pipe-to-flange joints, cracking was observed in the pipe wall next to the pipe-to-flange weld In case study 11, involving the failure of a sight port flange, cracking was observed in the flange adjacent to the pipe-to-flange weld In both cases, cracking was by an intergranular mechanism, and carbon contents were much higher than permitted for "L" grades of austenitic stainless steel Q 1997 Elsevier Science Ltd

Keywords: Metallography, residual stress, fractography

Liquid propellants, with their higher specific impulse than solid propellants, have emerged as efficient fuels for satellite hunch vehicles For combustion during flight, various types of oxidizers have been used One such oxidizer is dinitrogen tetroxide (N,O,) In view of the nonusability of dinitrogen tetroxide with some of the storage materials, mixed oxide of nitrogen (MON) has emerged as a better choice Its lower freezing point (- 14°C) makes this one of the best oxidizers MON mainly consists of a mixture of N204, NO and NO, Further classification of MON has been done as MON-

3 (3% NO) and MON-10 (10% NO), depending on the content of NO The plant, which produces mixed oxides of nitrogen (MON-3), consists of reactors, absorption columns, collection tanks,

storage tanks, interconnecting piping, flanges and valves One such plant, commissioned 7 years

back, is in operation However, for the last 2 years, leaks were observed in some of the components,

mainly in pipe-to-pipe joints and pipe-to-flange joints While attempting to rectify such a leak by tightening the flange, the flange broke into pieces Various components, which have cracked and leaked at different locations of the process plant, have been grouped into two categories In the first, the failure is in the pipe portion, and, in the second, the failure is in flange portion As such, two failed components, (i) a pipe-to-flange joint and (ii) a sight port flange, were selected for detailed studies

The joint consists of a tube (25mm ID and 32mm OD) made of AIS1 304 L grade stainless steel welded to a flange made of the same steel (Fig 1) The joint is used to transfer HNO, from a storage tank to an NO reactor A leak was noticed in the pipe portion of the joint

*Author to whom correspondence should be addressed

Reprinted from Engineering Fuihre Analysis 4 (3), 17 1-1 78 (1 997)

shows the chemical analysis of the material of the components

The chemical analysis of the material from the flange portion conforms to AISI 304L grade,

whereas that from the pipe portion conforms to AISI 304 grade only The whole component is

exposed to 90% HNOS at room temperature The component had a crack 5mm long in the pipe portion, about 4mm away from the weld between the flange and the pipe The component was cut open to reveal the inside portion of the pipe It was noticed that the inside surface of the pipe had

a corroded band about 7 mm in width throughout the inner circumference of the pipe, and the crack

l

I

Fig 2 Stereomicrograph of the crack on a pipe-to-flange joint x 16

Table 1 Chemical analysis (wtX) of the failed pipe-to-flange joint

~~ ~~~

Flange 0 034 0 010 19 3 9 8 1 3 0 3 Balance Pipe 0 093 0011 19 7 9 4 1 2 0 2 Balance

279

( d ) Fig 3 Optical micrographs of the failed pipe-to-flange joint (a) Microstructure near the crack, showing

equiaxed austenite grains with slip bands, and the absence of annealing twins (circumferential side) (b) and

(c) Microstructure of through-thickness section, showing intergranular fracture (d) Microstructure away from

the cracked region, showing equiaxed austenite grains with slip bands (note the absence of twins) All x 200

was seen to be located in the middle of this corroded band However the longitudinal weld in the pipe (produced during pipe manufacture) did not reveal any such corroded band The cracked portion of the piece was cut open The outer circumferential portion, as well as the through-thickness section, was polished to metallographic finish, and electrolytically etched using 10% oxalic acid solution for observation under an optical microscope Figure 3(ak(c) shows the microstructure in the cracked portion in both the circumferential portion and the through-thickness section The microstructure showed equiaxed austenite grains with slip bands, and the absence of annealing twins The microstructure slightly away from but adjacent to the crack also revealed the presence

of slip bands in the austenite grains [Fig 3(d)] The fracture surface of the sectioned other piece was observed under the scanning electron microscope (SEM) for fractographic features which revealed

the intergranular nature of the cracks

2.2 Discussion

Welding of dissimilar materials (with respect to chemical composition) will result in a difference

in electrode potential, and, hence, preferential attack may take place near the weld joint Welded tubing of austenitic stainless steel is recommended for use only after full-finishing, wherein the 6- ferrite networks within the weld metal structure are altered by cold work followed by a recrys- tallization anneal or solution anneal at 1065 "C to dissolve most of the ferrite, and to change the

cold worked structure to equiaxed recrystallized grains [ 13 The microstructure revealed the absence

of annealing twins, and, at the same time, the presence of slip bands, indicating that cold rolled sheets were used for tube making Subsequently welding the pipe to the flange might have introduced thermal stresses in the component near the weld However, at and near the longitudinal weld, there was no dissolution of metal (corroded band) It is observed that severely deformed areas will be chemically more active, and will rust before areas on the same part that have been subjected to little

or no deformation Residual stresses due to cold bending of a pipe, together with additional thermal stresses due to welding, have apparently accelerated the corrosion attack, with the subsequent

280

formation of a corroded band and crack initiation within the band during 7 years of service Intergranular cracks, typical of stress corrosion attack, as noticed on the fracture surface, are the evidence for such stress-induced preferential attack along the grain boundaries This type of

intergranular fracture is typical of the H N 0 3 stainless steel system [2] The pipe-to-flange weld has

a longer thermal mass than the pipe weld, so presumably required a comparatively large input Because of the high carbon content of the pipe, sensitization may have occurred in the pipe-to- flange weld as a result of the weld thermal cycle

2.3 Conclusions

Cold working and welding processes used for component fabrication introduced stresses Welding

of dissimilar metal compositions introduced electrode potential differences This combination created conditions conducive to preferential corrosion attack Sensitization may also have been a factor owing to the high carbon content of the pipe, and the high heat input which may have been required to make the pipe-to-flange weld

2.4 Recommendations

(1) Components should be given an annealing treatment after fabrication processes like bending,

(2) Joining by welding of dissimilar materials should be avoided

(3) Possible sensitization can be avoided by using the L grade of stainless steel

welding etc

3 CASE STUDY 11: FAILURE OF SIGHT PORT FLANGE

The second case study refers to the failure of the sight port flange, which was located between flanges in the NO absorption column Austenitic stainless steel AIS1 304L pipe of 80mm OD and

75 mm ID was welded to a flange which was also said to be made of similar steel A Teflon gasket was used between the flanges to provide a leak-proof system

3.1 Observations

Figure 4 shows a photograph of the failed component The failure was in the flange portion 7 mm away from the pipe-to-flange weld zone During the operation of the plant, there was a leak in one portion of the flange, and, while tightening, it crumbled The material in this component is exposed

28 1

Table 2 Chemical analysis (wt%) of the failed sight port flange

Flange 0.34 0.01 18.9 11.4 1.30 0.50 Balance Pipe 0.02 0.01 18.5 9.5 1 .so 0.40 Balance

to 70% HNO, (average) at temperatures between 10 and 35 "C The chemical compositions (Table

2 ) show high carbon (0.34%) in the flange, but conformity with AISI 304L in the pipe

Figure 5 shows a stereomicrograph of the cracked region of the component Figure 6(a) and (b) shows the microstructure of the component in the flange and pipe portion, respectively The microstructure of the flange portion does not correspond to AISI 304 L type steel, whereas the microstructure of the pipe portion is typical of AISI 304 L type steel, which agrees with the chemical analysis results Figure 7(a) shows the microstructure in the interface region of the pipe and flange, revealing the corrosion attack on the interface, and its propagation intergranularly into the flange portion with grain dissolution [Fig 7(b)] However, the pipe portion and the weld region do not show any such corrosion attack [Fig 7(c)] The microstructure in the flange portion shows grain

boundary thickening at and near the interface and intergranular cracking Figure 8(a)-(d) is SEM

fractographs of the failed flange, which show the completely intergranular nature of cracking and grain dissolution The preferential attack of the corrosive medium along slip bands is shown in Fig

8(b) and (c) Figure 8(d) shows the corrosion product (tar-colored powder) observed on the fracture

surface, and found to be Fe203 by XRD analysis (Fig 9)

Fig 5 Stereomicrogaph of the cracked portion of the sight port flange x 16

1 Fig 6 Optical micrographs of sight port flange (a) Flange portion (does not correspond to AISI 304 type)

282

Fig 7 Optical micrographs of sight port flange (a) Interface between flange and pipe portion, showing

corrosion attack on the interface and propagation intergranularly x 120 (b) Cracked area in the flange

portion, showing individual grain dissolution x 80 (c) Weld region (dendritic type), flange (left top) and pipe

(left bottom), showing corrosion attack in flange only x 120

283

f

e-

Fig 8 SI:M fractographs ol'fai1r.d right port rlangc (a) Radial cmck, \tarring I'rorn thc pip-to-llangc intcri,ice

(h) Intergranular cracks with grain dissolution (c) 1ntergranul;ir cracks nith prefcrenti.ll a t t x k :ildng 'Ilp bands (d) Corrosion product (Fe,O,)

have started at the flange-pipe interface, and propagated along the radial direction in the flange The pipe portion, however, did not reveal any cracks The grain boundary thickening in the flange portion is due to sensitization of the microstructure during welding of the flange to the pipe The sensitized microstructure contains Cr&-type carbides precipitated along the grain boundaries in

a fine network The area adjacent to the grain boundaries becomes depleted of chromium, and the corrosion resistance in this area deteriorates Two dissimilar metal compositions are in contact, and

a large unfavorable area ratio is present The net effect is rapid attack in the impoverished areas, resulting in intergranular fracture [3] The use of Teflon as gasket material can also enhance the

corrosion attack It is reported [4] that, in the system consisting of austenitic stainless steel, N204

and Teflon, the latter can enhance the corrosion rate if any moisture is present in the system

3.3 Conclusion

Usage of incompatible materials for different components has resulted in the failure of the component

3.4 Recommendations

(1) Joining by welding of dissimilar materials should be avoided

( 2 ) Chrome plating of components enhances the resistance to corrosion attack

(3) In oxidizing atmospheres like HNO,, care should be exercised to avoid moisture while using

Teflon as the gasket material in combination with austenitic stainless steels

Acknowledgement-The authors are grateful to Dr S Srinivasan, Director, Vikram Sarabhai Space Centre, ISRO,

1 Metals Handook, Vol 13: Corrosion, 9th edn ASM, Metals Park, OH, 1987, p 348

2 Metals Handbook, Vol 7: Atlas of Microstructures, 8th edn ASM, Metals Park, OH, 1972, p 133

3 Fontana, M G and Greene, N D , Corrosion Engineering McGraw-Hill, New York, 1967, p 59

4 Liberto, R R., Titan II Storable Propellant Hand Book Revision A, AFBSD-TR-62-2 Airforce Ballistic Systems Division CA, 1962, p 5-2