Investigation of failure behavior of two different types of Zircaloy clad tubes used as nuclear reactor fuel pins

Engineering Failure Analysis 18 (2011) 2042–2053

Contents lists available at ScienceDirect

Engineering Failure Analysis journal homepage: www.elsevier.com/locate/engfailanal

Investigation of failure behavior of two different types of Zircaloy clad tubes used as nuclear reactor fuel pins M.K. Samal a,⇑, G. Sanyal b, J.K. Chakravartty b a b

Reactor Safety Division, Bhabha Atomic Research Centre, Mumbai 400 085, India Mechanical Metallurgy Division, Bhabha Atomic Research Centre, Mumbai 400 085, India

a r t i c l e

i n f o

Article history: Received 29 December 2010 Received in revised form 7 June 2011 Accepted 8 June 2011 Available online 6 July 2011 Keywords: Zircaloy Fracture resistance behavior Fuel clad tube Load normalization technique Non-standard fracture mechanics specimen

a b s t r a c t Nuclear reactor fuel pins act as barriers to the release of radioactive fission products to the coolant flowing around these thin-walled tubes and hence they prevent the leakage of radioactivity to the surroundings of reactor core. These tubes are of small thickness in order to have less resistance in the path of heat flow from the fuel to the coolant. Investigation of failure behavior of these fuel clad tubes is of utmost importance to the designers and plant operators in order to ensure the maximum residence time of the fuel bundles inside the reactor core as well as to ensure minimal activity during operation and refueling activities. Various types of zirconium based alloys are used to manufacture these pins. The focus is to obtain better strength, ductility, corrosion resistance, oxidation resistance, and minimal creep including those due to irradiation-assisted damage and deformation processes. Two number of such types of alloys, namely, re-crystallization annealed Zircaloy-2 and stress-relief annealed Zircaloy-4, have been investigated in this work for their fracture behavior. As standard fracture mechanics specimens cannot be machined from these thin-walled tubes, non-standard specimens with axial cracks have been used in this work. Load normalization technique has been used to evaluate crack growth during loading of these specimens. It was observed that the re-crystallization annealed Zircaloy-2 specimens have higher initiation fracture toughness as well as higher resistance to crack growth compared to the other type of specimens. In order to understand the micro-structural aspects of the fracture resistance behavior of these materials, further investigation incorporating optical and transmission electron microscopy have also been carried out. It was concluded that the higher fracture resistance behavior of the re-crystallization annealed Zircaloy-2 specimens can be attributed to the presence of finer grain and subgrain micro-structure, very low dislocation density and other defects in the material. Ó 2011 Elsevier Ltd. All rights reserved.

1. Introduction Zirconium alloys are widely used as the fuel cladding tubes and other core internals in light water reactors such as boiling water reactors (BWR), pressurized water reactors (PWR) as well as pressurized heavy water reactors (PHWR). These alloys have favorable properties such as high mechanical strength, good ductility, very low neutron absorption cross-section and good corrosion resistance. A review of the development and application of these alloys is presented in Ref. [1]. The development of zirconium alloys for nuclear reactor applications started with addition of 2.5% Sn to pure zirconium producing Zircaloy-1 which had good corrosion resistance, strength, and formability. Later, addition of Fe and Cr to Zircaloy-1

⇑ Corresponding author. Tel.: +91 22 25591523; fax: +91 22 25505151. E-mail addresses: [email protected], [email protected] (M.K. Samal). 1350-6307/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.engfailanal.2011.06.009

M.K. Samal et al. / Engineering Failure Analysis 18 (2011) 2042–2053

2043

produced Zircaloy-2. This alloy has improved corrosion resistance compared to Zircaloy-1. Zircaloy-2 is now popularly used as fuel cladding material in BWRs and pressure tubes (the latest alloy contains 2.5% Nb in addition to the existing composition of Zircaloy-2) for PHWRs. Addition of more Sn and a little reduction in Fe content led to the generation of Zircaloy-3. One of the problems with these alloys is the hydrogen pickup and formation of hydride during reactor operation. The hydride formation process is accelerated because of the presence of Ni. By drastically reducing Ni content and increasing the content of Fe lead to development of Zircaloy-4 which has lesser tendency for hydrogen absorption (pickup) compared to Zircaloy-2. Zircaloy-4 is now employed as fuel element cladding in PWRs and PHWRs, spacer grid structural material in PWRs and channel box structural material in BWRs. Discussions regarding the development of these alloys, their property evaluation and the development of Zircaloys for new generation reactors can be found in several references such as Refs. [2–5]. The fuel clad tubes act as a barrier to release of radioactivity to the surrounding coolant and a conduit for transfer of heat generated due to nuclear fission from the nuclear fuel pellets (i.e., uranium oxide fuels) to the coolant (i.e., water in light water reactors and heavy water in pressurized heavy water reactors). In order to enhance the heat transfer, the tubes are manufactured as thin-walled tubes (thickness ranging from 0.38 to 0.9 mm approximately). The inner diameter varies from 10 to 15 mm approximately. Increasing the thickness of the tubes can also lead to decrease in available reactivity because of the decreased moderation ratio achieved by the moderator [6]. Kromar and Kurincic [6] estimated that fuel enrichment of the order of 0.06–0.15% is required in order to compensate for this loss in the moderator-to-fuel ratio. In case of PHWRs which use natural uranium oxide as fuel, this enrichment of fuel cannot be achieved unless any other changes in the reactor physics design parameters are carried out. It is important to ensure the structural integrity of these fuel clad tubes under various types of operational transients as well as accident scenarios. One of the postulated design-basis accidents for the fuel clad tubes is the reactivity initiated accident (RIA). In such a situation, the control rod drops or ejects very rapidly from the reactor. This in turn deposits a large amount of energy in the fuel pellets and leads to adiabatic heating and large fission gas release in the fuel pins [7]. The fuel pellets expand thermally and may cause fast straining of the surrounding Zircaloy clad tube through pellet–clad mechanical interaction. The fast loading imposed by the mechanical expansion of fuel pellet may cause rapid propagation of pre-existing cracks [8–10]. These pre-existing defects may have been initiated during the service life due to mechanical and chemical interaction of fuel pellets with the clad. The fracture behaviors of these clad tubes are further complicated due to several conditions such as hydrogen pickup, high temperature, irradiation environment, prior cold work and heat treatment conditions [11–16]. In order to ensure that the clad remains intact for a longer period in the reactor, fracture mechanics principles can be applied for the fitness-for-purpose of service assessment of these fuel pins. However, it may not be possible to evaluate the fracture behavior of these thin-walled tubular components using standard ASTM test techniques [17]. This is because of the stringent requirements of the ASTM test methods for satisfaction of plane strain condition at the crack tip. A number of test techniques have been proposed in research literature [18–24] that can approximate the state of stress experienced by the cladding for crack initiation and propagation in the axial direction (hoop stress being the dominant stress component) and provide a measure of the crack propagation resistance. A round-robin exercise to this effect has been carried out very recently and the results are documented in Ref. [25]. From a review of the literature, it was observed that attempts for evaluation of fracture resistance behavior of thin-walled Zircaloy tubes are very limited. Many times, a qualitative comparison between different alloys are carried out using Jmax values (which is evaluated at maximum load of the load–displacement diagram). It was pointed out in Ref. [25] that Jmax is not a proper indicator of fracture toughness values of these alloys. Moreover, the geometric functions of standard fracture mechanics specimens are usually used to evaluate the J-integral from the area under the load–displacement diagram in the absence of suitable geometric functions for these axially-cracked thin-walled fuel tubes. In this work, the fracture resistance behavior (in terms of J-integral vs. crack growth or J–R curve) of two types of alloys, namely, re-crystallized annealed (RXA) Zircaloy-2 and stress-relief annealed (SRA) Zircaloy-4, have been investigated. The RXA Zircaloy-2 is used as the fuel pin material of the Indian BWR whereas the SRA Zircaloy-4 is used as the fuel pin material of Indian PHWR. As standard fracture mechanics specimens cannot be machined from these thin-walled tubes, non-standard specimens with axial cracks have been machined from the actual fuel pins. These cracked specimens have been loaded with the help of two semi-cylindrical mandrels. These mandrels are inserted into the cracked pins and loaded at one end so as to open the cracks. The other ends of the mandrels are contact with each other through a central pin and they load the specimen by rotating through the pin. The load normalization technique [26–36] has been used to evaluate crack growth during loading of these specimens. The J– R curves are evaluated with the help of geometric functions obtained through finite element analysis in a recent study [22–24]. It was observed that the RXA Zircaloy-2 specimens have higher initiation fracture toughness as well as higher resistance to crack growth compared to the other type of specimens (i.e., SRA Zircaloy-4). In order to understand the microstructural aspects of fracture resistance behavior of these materials, further investigation incorporating transmission electron microscopy have been carried out. It was concluded that the higher fracture resistance behavior of the RXA Zircaloy-2 specimens can be attributed to the presence of finer grain and sub-grain micro-structure, very low dislocation density and other defects in the material. The paper is divided in four sections. The details of chemical composition, heat treatment and mechanical properties of the materials studied in this work are presented in Section 2. The method of evaluation of crack growth using

2044

M.K. Samal et al. / Engineering Failure Analysis 18 (2011) 2042–2053

load-normalization technique, evaluation of J–R curves, details of micro-structural investigation and a discussion regarding the difference in fracture behavior of both types of fuel pin specimens are presented in Section 3 followed by concluding remarks in Section 4.

2. Fuel pins, materials and experiment Two types of Zirconium alloys have been investigated in this work for their fracture behavior. These are designated as RXA Zircaloy-2 (fuel pin of Indian BWR) and SRA Zircaloy-4 (fuel pin of Indian PHWR) respectively. The fuel pins of the Indian PHWR (500 MWe capacity) are taken from the typical fuel bundles are shown in Fig. 1. These fuel bundles consist of 37 cylindrical fuel pins, which are held together by welding the pins to end plates at both the ends. The pins are arranged in concentric rings of 1, 6, 12 and 18 elements in different rings. Each fuel pin contains a stack of sintered natural UO2 pellets in a thin Zircaloy-4 cladding with end caps welded at both ends. Due to the pellet–clad mechanical and chemical interactions, the crack can initiate and propagate in the axial direction (as shown in Fig. 2) due to the dominant hoop stress. The crack initiation can be because of pellet–clad interaction as shown in Fig. 3 or fracture of hydrides as shown in Fig. 4. In order to study the fracture behavior of these thin-walled tubes, specimens have been machined from the as received fuel pins. Co-axial cracks of various sizes have been machined on these specimens and the specimens have been pre-fatigued so as to create sharp initial cracks. The typical axially-cracked fuel pin specimen in shown in Fig. 5a. Four specimens of RXA Zircaloy-2 and five specimens of SRA Zircaloy-4 have been tested for evaluation of their fracture resistance behavior. The specimens have been designated as RXA_Sp1, RXA_Sp2, RXA_Sp3, RXA_Sp4 and SRA_Sp1, SRA_Sp2, SRA_Sp3, SRA_Sp4, SRA_Sp5 respectively. The RXA specimens have internal diameter and wall thickness of 12.4 mm and 0.9 mm respectively. The SRA specimens have internal diameter and wall thickness of 14.2 mm and 0.4 mm respectively.

Fig. 1. Fuel bundle with 37 Zircaloy-4 fuel pins used in 500 MWe Indian PHWR.

Fig. 2. Crack initiation from a local pellet–clad interaction defect and subsequent propagation in the axial direction.

Fig. 3. Crack initiation and propagation from a initial defect caused by pellet–clad mechanical and chemical interaction.

M.K. Samal et al. / Engineering Failure Analysis 18 (2011) 2042–2053

2045

Fig. 4. Crack initiation and propagation from zirconium hydrides formed due to pickup of hydrogen by Zircaloy from the surrounding fluid.

The loading fixture consists of a pair of split semi-cylindrical mandrels which form the cylindrical holder when put together. The fixture halves are loaded through the pins after being inserted into the tubular cracked specimen. The loading arrangement is shown in Fig. 5b. The specimen is aligned such that the plane of axial cracks is co-planar with the openings of the split-halves of the mandrel. A pin at the back end of the fixture is inserted into a groove machined in the mandrels. It serves as an anchor point about which the semi-cylindrical fixtures rotate during the loading process. The tubular specimen is prevented from slipping from the loading mandrels by inserting a clip at the rear end of the mandrels (opposite to loading grooves). The loading pins at the right end of the fixtures in Fig. 5b are connected to the machine frames through the loading attachments. A notch is provided at the left end of the specimen in order to prevent any possible buckling during the loading. The initial crack length a0 and width W (19 mm in the present study) of the PLT setup are labeled in Fig. 5b. The initial crack length to width ratios (a0/W) of the RXA specimens are 0.526, 0.578, 0.631 and 0.684 respectively. The initial crack length is taken as average of crack lengths on both sides of the specimen. The a0/W ratios for the SRA specimens are 0.553, 0.555, 0.56, 0.568 and 0.569 respectively. The RXA Zircaloy-2 is an alloy of zirconium with 1.5 wt.% of Sn, 0.09% of Fe, 0.1% of Cr, 0.05% of Ni and 0.12% of O. The microstructure contains equi-axed grains with an average grain diameter of approximately 5 lm. Though the tube is manufactured by cold-pilgering process which subjects the materials to around 70% cold-work, the micro-structure is subsequently refined because of the annealing process (565 °C for 1.5 h). It not only relieves the residual stresses because of the cold-working, but also leads to re-crystallization where new sub-grains are formed making the grains more or less equi-axed. The matrix is more or less uniformly dispersed with inter-metallic precipitates of Zr(Fe, Cr)2 type as second phase particles. The SRA Zircaloy-4 is an alloy of zirconium with 1.5 wt.% of Sn, 0.21% of Fe, 0.1% of Cr, 0.007% of Ni and 0.12% of O. It can be noted that the Ni content has been reduced drastically in this alloy in order to have less hydrogen pickup and Fe content is increased. The micro-structure contains elongated grains due to the cold-working process. The stress relieving heat treatment is done at 480 °C for 3.5 h. This relaxes the residual stresses. However, the grain micro-structure is kept intact. The fuel pins tested in this work are un-irradiated and have not been subjected to any material degradation (due to reactor operating conditions). The tests have been conducted at room temperature. The yield stress and ultimate tensile strength for the RXA Zircaloy-2 are 427 and 576 MPa respectively. These data for the SRA Zircaloy-4 are 535 and 732 MPa respectively.

Fig. 5. (a) Axially cracked specimen machined from a fuel pin and (b) pre-cracked specimen loaded further by two split mandrels.

2046

M.K. Samal et al. / Engineering Failure Analysis 18 (2011) 2042–2053

3. Results and discussion 3.1. Fracture experiments on axially-cracked specimens machined from fuel pins The fracture specimens machined from the two types of fuel pins as (as discussed in the previous section) were tested in the experimental setup as depicted in Fig. 5b. The load-line displacement is obtained through a COD gauge which is mounted at the right end opening of the two semi-cylindrical mandrels as shown in Fig. 5b. Measurement of load-line displacement by COD eliminated the need for correction due to machine compliance. The load data is directly taken from the load cell of the machine. The load–displacement data for all the four specimens of type RXA and five specimens of types SRA are plotted in Fig. 6. As the thickness of RXA specimens are higher (i.e., 0.9 mm) compared to those of SRA specimens (i.e., 0.4 mm), the maximum load is higher for RXA specimens. Moreover, the load reaches a maximum and reduces further due to crack growth as well as plastic collapse (necking) of the tubes in the remaining ligament. The load–displacement data for the five SRA tubes are very close to each other compared to those of RXA specimens. This is because of very close a0/W ratios of the SRA specimens as compared to those of RXA specimens. These data from Fig. 6 is normalized using a load-normalization technique in order to evaluate the crack growth information during loading of these specimens. This technique is discussed as follows. 3.2. Evaluation of crack growth using load normalization technique Direct evaluation of crack growth during loading of small and thin tubular specimens is often cumbersome. Especially, in high rate loading conditions, aggressive environments such as high temperature, hydrogen atmosphere and irradiation, it may not be possible to directly measure the crack growth using conventional techniques such as ACPD (Alternating Current Potential Difference), DCPD (Direct Current Potential Difference), acoustic emission and ultrasonic techniques. Use of multispecimen test techniques may be expensive and time consuming. Moreover, there can be variability in the toughness properties among the specimens. The unloading compliance method may also incur significant error in crack growth estimation if there is significant nonlinearity in the unloading load–displacement curve due to various conditions such as change in contact conditions, nonlinear material relaxations and precision of instrument for compliance measurement. In order to circumvent the above problem, a load normalization method was at first proposed by Herrera and Landes [26]. In this method, the load is normalized by a geometry function and then a deformation function can be expressed as the relationship between the normalized load and the normalized plastic displacement. To determine the instantaneous crack growth during loading, the deformation function can be described by a power function [27] with two calibration points or by LMN function [28] with three calibration points. The calibration points are the data points at which the load, displacement and crack length must be known simultaneously. These data points are usually determined from the fracture surfaces of tested specimens. In this method, the load–displacement curves (Fig. 6) obtained from the test, are normalized using suitable normalization functions as discussed below. The load Pi at any point i, up to the maximum load is converted to a normalized load PNi as [29–30]

PNi ¼



Pi

2tW ðW  a0bi Þ=W

gpl

ð1Þ

where t is the thickness of the specimens, W is defined in Fig. 5b, gpl is the geometric factor which is also used in the evaluation of plastic part of the J-integral. The term a0bi is the blunting-corrected crack size at the ith data point and is written as 2000 RXA_Sp1 RXA_Sp3

RXA_Sp2

RXA_Sp4

load (N)

1500 SRA_Sp1 SRA_Sp3 SRA_Sp4

1000 SRA_Sp5

SRA_Sp2

500

0 0

1

2

3

4

displacement (mm) Fig. 6. Load–displacement curves of different types of specimens obtained from fracture tests.

M.K. Samal et al. / Engineering Failure Analysis 18 (2011) 2042–2053

a0bi ¼ a00 þ J i =2rY

2047

ð2Þ

where a00 is the crack length at the start of the test and rY is material flow stress (i.e., average of yield and ultimate tensile stress). The parameter Ji is J-integral value for the ith loading point based on initial crack length a0. The J-integral has two parts, i.e., Jel and Jpl and these are calculated as

J el ¼ K 2I =E0

ð3Þ 0

where KI is the mode-I elastic stress intensity factor, E is equal to Young’s modulus E corresponding to plane stress condition and E/(1  t2) corresponding to plane strain, t is Poisson’s ratio. For these thin-walled tubes, plane stress condition usually prevails at the crack tip. The other part, i.e., Jpl is calculated using plastic area under the load–displacement curve and its evaluation for a growing crack situation will be discussed later. The load-line displacement mi at the ith point, up to the maximum load, is also normalized to mpli using the compliance data Ci at the point of loading as

v 0pli ¼ ðv i  Pi C i Þ=W

ð4Þ

where Ci is the elastic compliance of the specimen based on the blunting corrected crack size a0bi and it is calculated using the following equation [22] for the RXA specimens

C ¼ 0:01101 þ 0:06113  ða=WÞ  0:11156  ða=WÞ2 þ 0:0691  ða=WÞ3

ð5Þ

The corresponding compliance vs. a/W relation for the SRA specimens is given by [23]

C ¼ 0:01752 þ 0:10473  ða=WÞ  0:20424  ða=WÞ2 þ 0:13573  ða=WÞ3

ð6Þ

The final load–displacement pair corresponding to fracture is also normalized using the same equations as above except that the final measured crack size a0f (after the specimen is break-open) is used. 3.3. Evaluation of geometric factors for evaluation of plastic part of J-integral Once the load–displacement data and crack size estimates are available at each data point, the plastic part of J-integral at each data point i, i.e., Jpl is evaluated using the following expression

   a0ðiÞ  a0ði1Þ gði1Þ  AplðiÞ  Aplði1Þ  1  cði1Þ J plðiÞ ¼ J plði1Þ þ bði1Þ 2t bði1Þ

ð7Þ

with g and c are the geometric factor and the correction factor respectively for a growing crack situation. The g factor is defined as a non-dimensional parameter that takes into account of the effect of geometry and loading condition at the crack-tip of a cracked specimen on the differential energy that is required for crack growth. It can be derived as a function of limit load function FL(a/W) as [23]



g¼ 1

a  1 @F L W F L @ða=WÞ

ð8Þ

For a growing crack, the solution for Jpl is corrected with a c factor. It is defined as a non-dimensional parameter that accounts for correction in the J-integral calculation due to crack propagation, which is otherwise defined for a stationary crack. The expression for c factor as a function of g(a/W) can be written as [23]



c¼g1 1

a  1 @g W g @ða=WÞ

ð9Þ

For evaluation of the above g and c functions, the limit load solutions are required for the different geometries which have been obtained here using a detailed 3D elastic–plastic finite element (FE) analysis of the test setup. For the FE analysis, the whole experimental setup (i.e., the tubular specimen as well as the mandrel) has been discretized in 3D domain. This is very essential as the load is applied at the end of the mandrel and this in turn loads the cracked tubular specimen in an opening mode. Taking into account of the symmetry in the specimen geometry, crack and loading configuration, one-fourth of the test setup has been modeled as shown in Fig. 7. A prescribed displacement is applied at the end of the mandrel is y-direction and the movement of the loading point is restricted in the x-direction (i.e., direction of length of the tube) as this point is attached to the loading axis of the machine. The un-cracked ligament of the tubular specimen is fixed in y-direction (however, this is free to move in x and z-directions) in order to take into account of the symmetry condition. The line of the loading pin (which acts like a fulcrum) is also fixed in y-direction only. The symmetric plane of the tubular specimen as well as of the mandrel is fixed in z-direction only in order to consider the symmetric boundary condition. The FE discretization consists of 20-noded iso-parametric brick elements with 3  3  3 Gauss point integration (full integration scheme). Elastic–plastic analysis with von-Mises yield surface and isotropic plastic hardening constitutive equations have been carried out in this work. The analysis has been carried out using an in-house FE code. The material true stress–plastic strain curves for both the materials are shown in Fig. 8. A

2048

M.K. Samal et al. / Engineering Failure Analysis 18 (2011) 2042–2053

Split mandrel for loading

Tubular specimen

Point for measurement of COD Crack front

Pin

Fig. 7. Typical 3D finite element mesh used for elastic–plastic analysis of the loading setup.

RXA specimen SRA specimen

1000

true stress (MPa)

800

600

400

200

0 0.0

0.1

0.2

0.3

0.4

0.5

true plastic strain Fig. 8. Stress–strain curve of the two alloys used in elastic–plastic FE analysis.

typical load–displacement for the RXA specimen with initial crack length to width ratio (a0/W) of 0.566 as obtained from FE analysis is shown in Fig. 9. This corresponds to the load–displacement of the specimen with stationary crack. The experimental data for the specimen with growing crack is also shown in Fig. 9. Both the specimens have same initial crack lengths. The load–displacement curves of the identical specimens with stationary and growing cracks start deviating after a displacement of about 1 mm. This is due to crack propagation in the actual experiment compared to the FE analysis where the crack is stationary. The limit load has been obtained as the load corresponding to the applied displacement at which the remaining ligament of the specimen goes into plastic deformation. The variations of limit load function FL with a/W for the two types of specimens are shown in Fig. 10. Using these data, the expressions for g(a/W) and c(a/W) have been derived using Eqs. (8) and (9) respectively. These functions are also plotted in Figs. 11 and 12 for both the types of fuel pin specimens. 3.4. Evaluation of fracture resistance curves for both types of alloys The representative normalized load–displacement data for specimens of both the materials are shown in Fig. 13. A line has been drawn from the final force–displacement pair, which is tangent to the remaining data points. Data to the right hand side of the point of intersection of tangent with remaining data are excluded. These data corresponds to the crack growth region. Data with v 0pli 6 0:001 are also discarded due to absence of crack growth. The remaining data points are fitted with the following analytical normalization function. 0

PN ¼

a þ bv pl þ cv 0pl 2 d þ v 0pl

ð10Þ

M.K. Samal et al. / Engineering Failure Analysis 18 (2011) 2042–2053

2049

2500

load (N)

2000

1500

a0/W=0.566 stationary crack (FE) growing crack (exp.)

1000

500

0 0.0

0.5

1.0

1.5

2.0

2.5

load line displacement (mm) Fig. 9. Comparison of load–displacement curve obtained from FE analysis (for stationary crack) with the experimental data for the RXA Zircaloy-2 specimens with initial a0/W ratio of 0.566.

3000 RXA specimen SRA specimen

2500

limit load (N)

2000

1500

1000

500

0 0.4

0.5

0.6

0.7

0.8

0.9

a/W Fig. 10. Limit load values obtained from FE analysis for the two types of specimens and their variation with a/W ratio.

3.5 RXA specimen SRA specimen

η factor

3.0

2.5

2.0

1.5

1.0 0.45

0.50

0.55

0.60

0.65

0.70

0.75

0.80

a/W Fig. 11. The factor ‘eta’ as a function of a/W ratio used in the J-integral calculation for both types of fuel pin specimens.

where a, b, c and d are fitting constants. Using these values of the constants, the crack growth information of the remaining data points are evaluated. The process involves an iterative procedure. The crack growth at a point is guessed and the normalized load and displacement data are re-calculated using Eqs. (1) and (4). The procedure is repeated till the data point falls on the curve defined by Eq. (10) within a pre-defined tolerance.

2050

M.K. Samal et al. / Engineering Failure Analysis 18 (2011) 2042–2053

2.0 RXA specimen SRA specimen

1.5

γ factor

1.0 0.5 0.0 -0.5 -1.0 0.45

0.50

0.55

0.60

0.65

0.70

0.75

0.80

a/W Fig. 12. The factor ‘gamma’ as a function of a/W ratio used in the J-integral calculation for both types of fuel pin specimens.

fitted curve

normalized load (MPa)

250 RXA data

fitted curve SRA data

200

150

0.000

0.025

0.050

0.075

0.100

0.125

normalized displacement Fig. 13. Normalized load–displacement data obtained for typical specimens of both type of materials used in this study which have been used for crack growth calculation during loading.

It can be observed from Fig. 13 that the normalized load–displacement curve for the RXA specimens are steeper compared to the SRA specimens. This indicated that the crack growth is slow in case of RXA specimens compared to the SRA specimens for the same amount of applied energy. This is also reflected in the J–R curves which are obtained from the plastic area under the load–displacement data using Eq. (7), crack growth information from load-normalization method and use of suitable g and c functions. The J–R curves for both types of specimens are plotted in Fig. 14. The J–R curves of SRA specimens are very close to each other due to the similar a0/W ratios whereas these are slightly different for the specimens of RXA alloy due to different a0/W ratios. RXA specimens with higher a0/W ratios have lower J–R curves. Moreover, it can be observed that RXA specimens have a markedly high fracture resistance behavior compared to the SRA specimens. The reason for the above difference lies in the initial micro-structure of these alloys and this aspect of the difference in the fracture resistance behavior between these two materials is investigated further as follows. 3.5. Micro-structural investigation using optical and transmission electron microscopy In order to investigate the higher fracture resistance of RXA Zircaloy-2 specimens compared to the SRA Zircaloy-4 specimens, we have carried out optical and transmission electron microscopy of these alloys. The specimens for optical microscopic examinations were cut from the as-received cladding tubes as rings and were mounted using Bakelite powder with the help of a mounting press. The specimens were ground using 300 grade emery papers and were then polished first using alundum (an abrasive composed of fused alumina) and finally using 0.25 lm diamond paste. A set of polished specimens was examined for characterizing the microstructure using 5 ml HF and 2 ml AgNO3 in 100 ml water as etchant. The micro-structures of both these alloys are shown in Fig. 15a and b for RXA Zircaloy-2 and SRA Zircaloy-4 respectively. The RXA Zircaloy contains equi-axed grains due to the re-crystallization annealing heat treatment as discussed in Section 2, whereas the SRA Zircaloy-4 contains elongated grains. The higher fracture toughness value obtained for the RXA specimens is because of the presence of finer grain and sub-grain structure compared to the SRA specimens.

M.K. Samal et al. / Engineering Failure Analysis 18 (2011) 2042–2053

2051

RXA_Sp1 RXA_Sp2 RXA_Sp3 RXA_Sp4

500

J-integral (N/mm)

400 SRA_Sp1 SRA_Sp2 SRA_Sp3 SRA_Sp4 SRA_Sp5

300

200

100

0 0.0

0.2

0.4

0.6

crack growth (mm) Fig. 14. J-integral vs. crack growth curves of all the pre-cracked specimens of both materials.

Fig. 15. Micro-structure of (a) RXA Zircaloy-2 and (b) SRA Zircaloy-4 obtained using optical microscopy.

Transmission electron microscopic (TEM) examinations have also been carried out for both the materials in order to understand their sub-structural features (e.g., information regarding initial dislocation density). Thin disk shaped specimens (3 mm diameter) were punched out from the tubes and were thinned by grinding with emery papers. These specimens were next subjected to electrolytic thinning by jet polisher. Perchloric acid (HClO4) was used as the electrolyte. After achieving a small hole at the center of the disk, these were washed with ethanol and water, and were finally dried for observation in TEM. The microstructures of the selected alloys could not be revealed satisfactorily by optical microscopy or by SEM examinations, primarily due to its internal structure governed by the parameters of pilgering plus re-crystallization/stress relief annealing heat treatment. Transmission electron microscopic examinations were, therefore, carried out to understand the characteristic of the grain structure and the substructures. Typical TEM micrographs are shown in Fig. 16. Prominent

Fig. 16. (a) Transmission electron microscope picture of RXA Zircaloy-2 depicting the very low density of initial dislocations and (b) enlarged view.

2052

M.K. Samal et al. / Engineering Failure Analysis 18 (2011) 2042–2053

Fig. 17. (a) Transmission electron microscope picture of SRA Zircaloy-4 depicting presence of a very high density of initial dislocations and (b) enlarged view.

dislocation lines are observed in Fig. 16b, while a typical grain boundary triple point junction associated with high angle sharp grain boundaries can be noted in Fig. 16a for RXA Zircaloy-2. Typical results obtained through TEM for the SRA Zircaloy-4 are shown in Fig. 17. A dense network of dislocation tangles can be seen in Fig. 17b. In Fig. 17a, distinct sub-grains are seen with some precipitates within the matrix. These precipitates have not been identified but are probably some intermetallics of Zr with the alloying elements. Comparing Figs. 16 and 17, it can be confirmed that the initial dislocation density in SRA Zircaloy-4 is higher than that in RXA Zircaloy-2. By stress relief annealing, the dislocation network generated due to the cold pilgering operation does not annihilate unlike that in the case of re-crystallization annealing. It can be concluded that the higher fracture resistance behavior of the RXA Zircaloy-2 specimens can be attributed to the presence of finer grain and sub-grain micro-structure, very low dislocation density and other defects in the material.

4. Conclusions In order to increase the resident time of fuel bundles in the nuclear reactor and prevent the release of radioactivity into the coolant, it is crucial to ensure the integrity of these thin-walled fuel pins. Though standard fracture mechanics specimens cannot be machined from these tubes, non-standard specimens can be machined from the as-received tubes. Using machined and pre-fatigued cracks in the axial directions, the actual fracture resistance behavior can be evaluated. Though these are not true fracture properties according to the ASTM standard, these can be used for integrity assessment of the particular tubes concerned. In this study, the fracture resistance behaviors of two types of fuel pin specimens were studied. The crack growth during loading was evaluated using load-normalization techniques. The RXA Zircaloy-2 specimens were found to have better fracture resistance behavior due to the presence of finer grain and sub-grains in the micro-structure and absence of large initial dislocation density (due to the re-crystallization annealing heat treatment procedure). However, the fracture properties of these fuel pins degrade due to irradiation, high temperature and various material degradation processes in the reactor environment. The future research will focus on these aspects of material fracture behavior. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11]

[12] [13] [14] [15] [16] [17]

Northwood DO. The development and applications of zirconium alloys. Mater Des 1985;6(2):58–70. Nikulina AV. Zirconium–niobium alloys for core elements of pressurized water reactors. Met Sci Heat Treat 2003;45(7–8):287–92. Nikulina AV. Zirconium alloys in nuclear power engineering. Met Sci Heat Treat 2004;46(11–12):458–62. Seok CS, Bae BK, Koo JM, Murty KL. The properties of the ring and burst creep of ZIRLO cladding. Eng Fail Anal 2006;13:389–97. Lee KW, Kim SK, Kim KT, Hong SI. Ductility and strain rate sensitivity of Zircaloy-4 nuclear fuel claddings. J Nucl Mater 2001;295:21–6. Kromar M, Kurincic B. Impact of thicker cladding on the nuclear parameters of the NPP Krsko fuel. Nucl Eng Des 2011;241(4):1087–9. Meyer RO, McCardell RK, Chung HM, Diamond DJ, Scott HHA. Regulatory assessment of test data for reactivity initiated accidents. Nucl Safety 1996;37(4):271–88. Leclercq S, Parrot A, Leroy M. Failure characteristics of cladding tubes under RIA conditions. Nucl Eng Des 2008;238(9):2206–18. Chung H, Kassner T. Cladding metallurgy and fracture behavior during reactivity-initiated accidents at high burnup. Nucl Eng Des 1998;186:411–27. Saux ML, Besson J, Carassou S, Poussard C, Averty X. Behavior and failure of uniformly hydrided Zircaloy-4 fuel claddings between 25° C and 480° C under various stress states, including RIA loading conditions. Eng Fail Anal 2010;17:683–700. Arsène S, Bai J, Bompard P. Hydride embrittlement and irradiation effects on the hoop mechanical properties of Pressurized Water Reactor (PWR) and Boiling-Water Reactor (BWR) Zircaloy cladding tubes. Part III. Mechanical behavior of hydride in stress-relieved annealed and recrystallized Zircaloys at 20° C and 300° C. Metall Mater Trans A 2003;34A:579–88. Bertolino G, Meyer G, Perez Ipiña J. Effects of hydrogen content and temperature on fracture toughness of Zircaloy-4. J Nucl Mater 2003;320:272–9. Une K, Ishimoto S. EBSP measurements of hydrogenated Zircaloy-2 claddings with stress-relieved and recrystallized annealing conditions. J Nucl Mater 2006;357(1–3):147–55. Bertolino G, Meyer G, Ipiña JP. Degradation of the mechanical properties of Zircaloy-4 due to hydrogen embrittlement. J Alloy Compd 2002;330– 332:408–13. Bertolino G, Meyer G, Perez IJ. Effects of hydrogen content and temperature on fracture toughness of Zircaloy-4. J Nucl Mater 2003;320:272–9. Bertolino G, Meyer G, Perez IJ. In-situ crack growth observation and fracture toughness measurement of hydrogen charged zircaloy-4. J Nucl Mater 2003;322:57–65. ASTM Standard E 1820-01. Standard test method for measurement of fracture toughness. Philadelphia, PA: ASTM; 03.01.2000, p. 1–21.

M.K. Samal et al. / Engineering Failure Analysis 18 (2011) 2042–2053

2053

[18] Grigoriev V, Josefsson B, Lind A, Rosborg B. A pin-loading tension test for evaluation of thin-walled tubular materials. Scr Metall Mater 1995;33(1):109–14. [19] Grigoriev V, Jakobsson R. Delayed hydrogen cracking velocity and J-integral measurements on irradiated BWR cladding. J ASTM Int 2005;2(8):1–16. [20] Grigoriev V, Josefsson B, Rosborg B. Fracture toughness of Zircaloy cladding tubes. Zirconium in nuclear industry. In: Bradely ER, Sabol GP, editors. Proc of 11th Int Sympos. ASTM STP 1295; 1996, p. 431–47. [21] Bertsch J, Hoffelner W. Crack resistance curve determination of tube cladding material. J Nucl Mater 2006;352:116–25. [22] Samal MK, Sanyal G, Chakravartty JK. An experimental and numerical study of the fracture behaviour of tubular specimens in a pin-loading-tension set-up. J Mech Eng Sci 2010;224(1):1–12. [23] Samal MK, Sanyal G, Chakravartty JK. Estimation of fracture behavior of thin walled nuclear reactor fuel pins using Pin-Loading-Tension (PLT) test. Nucl Eng Des 2010;240(12):4043–50. [24] Sanyal G, Samal MK, Ray KK, Chakravartty JK, Suri AK, Banerjee S. Prediction of J-R curves of thin-walled fuel pin specimens in a PLT setup. Eng Fract Mech 2011;78(6):1029–43. [25] Yagnik SK, Ramasubramanian N, Grigoriev V, Sainte-Catherine C, Bertsch J, Adamson R, et al. Round-robin testing of fracture toughness characteristics of thin-walled tubing. J ASTM Int 2008;5(2):1–21. [26] Herrera R, Landes JD. A direct J–R curve analysis of fracture toughness tests. J Test Eval 1988;16(5):427–49. [27] Landes JD, Herrera R. A new look at J–R curve analysis. Int J Fract 1988;36:9–14. [28] Landes JD, Zhou Z, Lee K, Herrera R. Normalization method for developing J–R curves with the LMN function. J Test Eval 1991;19(4):305–11. [29] Sharobeam MH, Landes JD. The load separation criterion and methodology in ductile fracture mechanics. Int J Fract 1991;47(2):81–104. [30] Sharobeam MH, Landes JD. The load separation and gpl development in precracked specimen test records. Int J Fract 1993;59(3):213–26. [31] Cassanelli AN, Ortiz H, Wainstein JE, de Vedia LA. Separability property and Load normalization in AA 6061-T6 aluminum alloy. Fatigue Fract Mech. ASTM STP 1406, American Society for Testing and Materials; 2001. p. 49–72. [32] Dzugan J, Viehrig HW. Application of the normalization method for the determination of J–R curves. Mater Sci Eng A 2004;387–389:307–11. [33] Zhu XK, Joyce JA. J-resistance curve testing of HY80 steel using SE (B) specimens and normalization method. Eng Fract Mech 2007;74(14):2263–81. [34] Zhu XK, Lam PS, Chao YJ. Application of normalization method to fracture resistance testing for storage tank A285 carbon steel. Int J Press Vessels Pip 2009;86(10):669–76. [35] Baldi F, Riccò T. High-rate J-testing of toughened polyamide 6/6: applicability of the load separation criterion and the normalization method. Eng Fract Mech 2005;72(14):2218–31. [36] Morhain C, Velasco JI. Determination of J–R curve of polypropylene copolymers using the normalization method. J Mater Sci 2001;36(6):1487–99.