Poolside equipment equipped with linear variable differential transformers (LVDTs) can be used to determine the dimensions of fuel assembly com­ponents. As an example, LVDTs may be used to measure the dimensions of the outer flow channel of BWR fuel assemblies. In this case the position of a channel relative to a reference surface is measured. Each of the four channel sides is measured by three transducers and consequently 12 axial traces are obtained simultaneously over the circumference of the channel. From these measurements, the bulge, bow, and twist may be calculated over the total length of the channel. Bulge is an outward deformation of the channel faces which results from the difference in pressure between the inside and the outside of the flow channel and from the effects of fast neutrons on channel creep. Twist is an angular deformation over the length of an assembly.

The length of the fuel channel is also of interest and can be measured with in-pool devices that range from tape measures to caliper-like gauges equipped with LVDTs or similar sensors. Length data are used to assess the effects of differential growth in the axial direction on the fit and remaining growth margin of the channel relative to its fuel bundle. Length data can also be used to estimate lateral bow in cases where time or equipment avail­ability prevents more accurate, explicit measurements.

1.2.1 PWRs

A large variety of structural metals present in primary and secondary circuits of PWRs suffer corrosion: [1]

used to achieve better hardenability. These alloys have a limited resis­tance to corrosion.

• Stainless steels are iron-base alloys containing more than 13% chro­mium. An austenitic structure is obtained when they contain a large nickel content. Austenitic stainless steels have excellent resistance to corrosion despite the fact that they can suffer pitting and SCC under some conditions. A martensitic structure is obtained when the nickel content is low. Martensitic stainless steels have high mechan­ical strength but they are less resistant to corrosion than austenitic structures.

• Ni alloys have 15-30% Cr, with a high resistance to uniform and pitting corrosion, but they are susceptible to SCC, except for those with the highest range of chromium content. Ni alloys are expensive, especially with 30% chromium metals.

In the primary circuit, the water is at a temperature ranging from 270°C to 345°C. Boron is introduced as boric acid to absorb and to control the core reactivity. In order to counteract the general corrosion of materials, lithium hydroxide is added to the water, in order to reach a slightly alkaline pH of 7.2 at 300°C. Finally, hydrogen is dissolved to counteract the radiolytic decomposition of water into oxidizing compounds that may lead to SCC of stainless steels. In addition, a few ppb of zinc can be introduced to mitigate the activation of cobalt.

In the secondary circuit, the pH of the boiling water is also made slightly alkaline in both liquid and steam phases to limit corrosion. The original operating chemistry combined ammonia (for its slightly alkaline pH) with phosphate (to buffer the various potential contaminants that may enter the system through the condenser). Current chemistries involve all vola­tile treatment (AVT) without any phosphate addition in high quality water. Ammonia is added to get a pH25°C higher than 9.8 in plants without any copper alloys to avoid FAC of carbon steel. In other situations, an amine is preferred such as ammonia, morpholine or ethanolamine which exhibit a high thermal stability. Last, hydrazine is introduced to obtain a reducing environment and to limit the SCC of Alloy 600 tubing.

In the section on diffusion creep mechanisms we discussed the importance of grain size, temperature and stress in determining the rate controlling

mechanism. Coble creep is dominant at very fine grain sizes whereas N-H creep is rate controlling at larger grain sizes. Also at relatively low homolo­gous temperatures Coble creep is rate controlling and N-H creep takes pre­cedence at high homologous temperatures. This approach to understanding diffusion creep is quite valid for metals and alloys. Similar phenomena in ceramics become complex due to ambipolar diffusion and stoichiometry. The diffusion flux of both cations and anions constituting the ceramic must be considered to estimate the net diffusion rate. In monovalent materials the vacancies in diffusion creep regime can get transported along the grain boundaries or the lattice and the total strain rate of deformation is given by the sum of N-H and Coble creep mechanisms. But in a ceramic of the type ApBq, where A is the cation and B the anion, both the anions and cat­ions participate in the diffusion process and might adopt different transport paths. In this case the total strain rate of deformation in the Coble creep model is given by

composite > Dcomposite

(1/ p)[d+l + 12 (s+D+h/d)

(q/p)(DL +1. ifrnk/dyDi+l^S-D-Jd))’

The transport path of the anions and cations was originally considered by Gordon100 who suggested that the total transport of vacancies from the hor­izontal to the vertical boundaries should be in the appropriate stoichiomet­ric ratio. This leads to the prediction that creep would be controlled by the diffusion of the slower moving species along the faster diffusion path. In this scenario, it is possible for the cations and anions to be transported predom­inantly along different paths as depicted in Fig. 3.25a.

However the transport paths suggested by Gordon might lead to the development of local non-stoichiometry101 which has not been observed in ceramics. Hence Chokshi102 suggested that it would be appropriate to con­strain diffusion fluxes along each path to be in the appropriate stoichiometric ratio, as depicted schematically in Fig. 3.25b. In this scenario, it is necessary to find the slower moving species along each path, and the rate controlling process is then determined from the faster diffusion path. The difference in transport paths suggested by Gordon100 and Chokshi102 has implications for the transitions in diffusion creep mechanisms. Plots of strain rate against TmIT, for a fixed grain size are shown in Fig 3.26a and 3.26b. The symbols C and N, in these figures, represent Coble and N-H creep, and the superscripts + and — represent cation and anion, respectively. Figure 3.26a, correspond­ing to transport paths suggested by Gordon, indicates that there will be transitions with an increase in temperature from diffusion creep controlled by cation grain boundary diffusion (C+) to cation lattice diffusion (N+) to

3.25 ( a) Transport paths in ceramics as suggested by Gordon where the total flux from horizontal to vertical boundaries is in the appropriate stoichiometric ratio. (b) Transport paths in ceramics as suggested by Chokshi where total flux along each transport path is in the appropriate stoichiometric ratio.

anion grain boundary diffusion (C-) to anion lattice diffusion (N-). Figure 3.26b, corresponding to transport paths proposed by Chokshi, indicates that over the same temperature range, there will only be a single transition from Coble creep controlled by cation grain boundary diffusion to N-H creep controlled by anion lattice diffusion.

The temperature of a component during irradiation is an important variable. In a BWR, temperature variation along the length of the core is relatively small: water rods and spacers are near the temperature of the boiling water (288°C, 561K); fuel cladding material operates at slightly higher temperatures due to heat generation from the fuel and buildup of oxide and crud; but the range is between 288°C (561K) and 320°C (593K). However, in a PWR the components all operate with a substantial axial temperature gradient due to increase in the water temperature as it rises through the core. Depending on core design and duty, material temperatures could be as low as 280°C (553K) at the bottom and nearly 400°C (673K) at the top of the core. Therefore, the temperature depen­dence of irradiation growth and creep must be accounted for.

Growth as a function of temperature is not straightforward, as shown schematically in Fig. 4.65 (Holt, 1988) for RXA Zircaloy. In general for Zircaloy at low fluence, growth decreases with increasing temperature and at high fluences growth increases with temperature, with the critical tem­perature (T1 in Fig. 4.65) being near 360°C (633K). It is seen that at less than about 2 x 1025 n/m2 (E>1MeV) (which is before the region of break­away growth and before c-component dislocations form) growth peaks at about 300°C (573K) and then steadily decreases at higher temperatures. At post-breakaway fluences, growth rate (the slope of the growth vs fluence

Longitudinal,

Transverse

0 —

О О

1

5 10 15 Neutron fluence, E > 1 MeV (1025nm-2)

4.64 Dependence of growth on neutron fluence in three orthogonal directions at irradiation temperatures in the range 651-669K (Tucker et a!., 1984).

4.65 Schematic diagram showing the growth of annealed Zircaloy in the longitudinal direction (FL <0.1) as a function of temperature (Holt, 1988).

curve) steadily increases with increasing temperature. Also, as indicated in Fig. 4.65 , the fluence at which growth breakaway occurs gets smaller as temperature increases. Also note in Fig. 4.66 (Rogerson, 1988) that at temperatures above about 360°C (633K) the growth rate is the same for recrystallized (RXA) and cold worked (SRA or CWSRA) materials.

Temperature effect information is given by Gilbon & Simonot (1994) and Gilbon et al. (2000) and confirms that growth dramatically increases between the temperatures 350°C and 400°C (623-673K). Gilbon et al. (2000) also give data showing that growth of the M5 alloy at 5 x 1025 n/m2 decreases from 0.08% at 280°C (553K) to 0.04% at 350°C (623K).

The effect of temperature on growth rates of cold-worked Zr-2.5Nb is given by Holt et al. (2000). In the range of interest for CANDU reactors, that is 290-317°C (563-590K), the growth rate decreases with increasing temperature at high or low fluences. This is slightly different and opposite to that observed for Zircaloy, but the range of temperature data available is smaller than for Zircaloy.

The practical implications of the discussed temperature dependence should be clear. For BWRs and CANDUs variations in growth should be small and predictable, at least for the alloys currently used, and for PWRs variation may be large and, particularly above about 350°C (623K), may be unpredictable with our current state of knowledge.

Insulation resistance, ‘Hi-Pot,’ partial discharge, AgeAlert™, and quality/ dissipation factor are electrical measurement tests for the entire cable cir­cuit (cable, connections, and end device) to identify cable insulation degra­dation, failed end devices, and moisture intrusion on the cable.

Insulation resistance (IR) — the simplest and most common test for moni­toring cable ageing — quantifies the quality of cable insulation by energizing the cable conductor and measuring for leakage current through degraded insulation (AMS Corp., 2011; U. S. NRC, 1990). One of two fundamental wire insulator properties, insulation resistance is the resistance to current leakage through and over the surface of the cable material. Insulation can also be impacted by cable length; humidity or moisture in the cable and insulation as well as dirt, oil, and other surface contaminants (U. S. NRC, 2010a). IR changes in a progressive, ongoing basis as a cable is exposed to these environmental stressors (IAEA, 2011).

In an IR test, a high voltage (e. g. 100 V DC), is applied between each cable lead and the cable shield, and also between the cable shield and ground. The principle of the IR test is that the application of a DC voltage to an insu­lated conductor induces a small current in the insulation to ground (U. S. NRC, 2010a). As the high voltage is applied, the leakage current through the insulating material is measured to establish the quality of the cable insula­tion and to determine if there are any contaminants (moisture, grease, dirt, etc.) in the cable (Hashemian, 2010). The IR test is easy to perform with inexpensive equipment, but it is a simple pass/fail test for cable dielectric and results are too inconsistent for trending purposes (U. S. NRC, 2010a; U. S. NRC, 2010b).

Two other insulation quality tests, polarization index (PI) and dielec­tric absorption ratio (DAR), provide the trendability that the IR test lacks. This is essential because if insulation is badly deteriorated, wet, or contami­nated, the leakage current will exceed acceptable levels and could continue to increase over time. Insulation resistance is therefore typically measured at different time intervals, with the ratio of these two measurements con­stituting the polarization index (PI) or polarization ratio (PR) (U. S. NRC, 2010a). The PI test detects cracking induced by heat, radiation, moisture, and surface contamination. Although the PI test is trendable, easy to per­form and does not require access to the entire cable, it requires that the end terminations be disconnected and is insensitive to insulation degradation (U. S. NRC, 2010b).

DAR is another index of the quality of cable insulation over time. To determine DAR, the IR is measured 60 s after applying the test voltage, and that result is then divided by the IR measurement after 30 s. Depending on how fast the system polarizes, the IR will increase and then start to plateau over time. DAR is somewhat subjective and should be considered in the con­text of IR; it is not an absolute indicator of insulation quality (Hashemian, 2010). DAR, PR, and PI values of less than 1.0 usually indicate degradation in the insulating material, which may be due to dirt, moisture, cracking, age­ing, or other problems.

Another insulation quality test, the direct current (DC) high-potential test (Hi-Pot), is a pass/fail test used on medium-voltage power cables and all insulation and jacket materials to detect embrittlement and cracking caused by heat and radiation, mechanical damage, water treeing, moisture intrusion, and surface contamination. The principle of the test is that if a cable contains defects, the high test voltage will force the defects to fail. In the DC Hi-Pot test, a high-voltage potential is applied to the insulation to see if it can withstand higher DC potential than it normally experiences in operation. Because cable insulation can normally endure sustained DC potential without damage, the Hi-Pot test is typically used to repetitively test insulation at sufficient voltage to indicate whether insulation is weak enough to begin failing in service (breakdown voltage) but without damag­ing sound insulation (withstand voltage). The Hi-Pot test is easy to perform, provides trendable data, and does not require access to the entire length of a cable. However, the cable must be disconnected in order to perform the test, and the high voltages applied may damage the cable insulation (U. S. NRC, 2010b).

The partial discharge (PD) test is another insulation quality assessment method. Partial discharges are small electrical sparks that occur at voids, gaps, and similar defects within the insulation in medium and high voltage cables. Over time, these partial discharges will erode the insulation and ulti­mately break down the cable completely, resulting in embrittlement and cracking, mechanical damage, and water treeing (U. S. NRC, 2010b). The lower the PD inception voltage, the greater the degradation of the insula­tion material (IAEA, 2011).

Measurements of partial discharge are performed in both time and fre­quency domain by a monitor connected to a cable circuit. If a sufficiently high voltage (called the inception voltage) is applied across a cable insula­tion, an electrical discharge (partial discharge or corona) can occur in small voids or air gaps in the insulation or between insulation and a ground plane or shield (U. S. NRC, 2001). The monitor or oscilloscope measures the peak magnitude of the partial discharge pulse, phase angle, and pulse shapes of the partial discharge signals acquired. PD test equipment can determine the location of the voids or gaps by measuring the time lag between direct and reflected pulses from the discharge site or by using acoustic emission monitoring techniques (U. S. NRC, 2001). These measurements can be made continuously or intermittently and identified on — or off-line (Hashemian, 2010).

The PD test does not require access to the full length of the cable and enables both the quantification of the severity of insulation defects and identification of their location in the cable. However, the end terminations of the cable must be disconnected to perform the test, the test itself requires highly skilled personnel, and the high testing voltage can weaken or damage cable insulation (U. S. NRC, 2010b).

Another insulation quality test — AgeAlert™ — is a wireless microsensor that measures ageing or degradation of electrical insulation. Constructed out of cable insulation and nano-size conductive particles, it is installed in multiple locations along cables or embedded in motors and allowed to age together with the insulation material being monitored (IAEA, 2011). The parameter that is measured and correlated to cable condition is the resistiv­ity of the microsensor as a function of its age. Because it is constructed out of the same material as the cable it measures, it responds to temperature, humidity and radiation environments much as the insulation does (IAEA, 2011). Under thermal-oxidative conditions, the polymer material becomes denser and loses volume. This change results in a change in resistance of the AgeAlert™ microsensor. The resistance data can be transmitted wirelessly using radio frequency identification (RFID) technology for short distances and can be acquired by a handheld RFID reader or a personal digital assis­tant (PDA) type device. If needed for long distance transmission, active wireless transmission technology could also be used. The AgeAlert™ has no internal power source, receiving power from the radio frequency (RF) sig­nal that is used to interrogate it and read the resistivity data from the sensor (AMS Corp., 2010). The AgeAlert™ sensors can be installed by wire/cable manufacturers during manufacture of the cable or bonded to the cable after installation (IAEA, 2011).

Dissipation factor (DF) and quality factor (QF) represent a final insu­lation quality test. The ratio of the energy loss in a dielectric to the total energy transmitted through the dielectric, DF represents the departure of a cable from an ‘ideal’ capacitor. If the cable is free of defects or contami­nants, its dielectric properties are similar to a perfect capacitor. If the cable dielectric contains impurities, the resistance of the insulation decreases, and it no longer acts as a perfect capacitor (Hashemian, 2010). Similarly, QF represents the departure of a cable from an ‘ideal’ inductor. Quality factor applies to electrical circuits that contain resistance, inductance, and capaci­tance and is the ratio of energy stored to energy dissipated in a system at a specific frequency.

It is possible to identify a particular micromechanism of creep through knowledge of the stress exponent (n), the activation energy (Qc) and the grain size exponent (p). Table 3.1 describes the different mechanisms of creep and their relationship to the creep parameters n, Qc and p. In addition to these three parameters, the relevant mechanism of creep can be identi­fied through knowledge of the creep constant A given by K2 in Equation

[1.16] . Each mechanism of creep possesses a distinct value of A.

The mechanisms of creep can be broadly classified into two types: diffusion-based processes and dislocation-based processes. Coble creep and Nabarro-Herring (N-H) creep are mechanisms of deformation that fall under the category of diffusion-based processes. Harper-Dorn (H-D), vis­cous glide and dislocation climb are mechanisms of creep that fall under the category of dislocation-based processes. Grain boundary sliding (GBS) appears to proceed by a combination of diffusion — and dislocation-based

processes. Power-law breakdown (PLB) occurs at relatively high stresses equal to or greater than around 10-3E (E is modulus of elasticity); this has also been correlated with strain rates at or greater than 10-9D (D is diffu — sivity).16 In this high stress and high strain rate regime, the creep-rates vary with stress via an exponential function. While this region is observed in all materials under appropriate conditions, the underlying mechanism of PLB is still a moot factor.

In this table, Qgb is the grain boundary diffusion activation energy, QL the activation energy for lattice diffusion, and Qs the activation energy for solute diffusion. As Table 3.1 suggests, a stress exponent value of unity (Newtonian viscous) imply that the deformation mechanism could be Coble, N-H or H-D creep. However knowledge of the grain size exponent or the activa­tion energy would establish the right mechanism of creep. For example, a stress exponent of 1 and activation energy equal to that for lattice diffusion would suggest the mechanism of creep to be either N-H or H-D. But if the grain size exponent is equal to 2, it would establish that the mechanism of deformation is N-H. On the other hand, if the steady-state strain rate is found to be independent of the grain size (p = 0), the mechanism of creep is H-D. The fact that the Coble creep mechanism is more sensitive to the grain size and is controlled by the grain boundary diffusivity, it becomes dominant at lower temperatures and/or smaller grain sizes while N-H creep becomes predominant at relatively larger grain sizes and higher tempera­tures. H-D creep becomes significant at large grain sizes and bulk single crystals. Thus, knowledge of the creep parameters would help in identifying the exact mechanism of creep.

RBMKs are basically vertical channel-type boiling water reactors (Cox et al, 2004). They use E110 alloy (Zr-1%Nb) fuel cladding in the fully recrystallized conditions. The fuel assembly consists of two fuel bundles of 18 rods supported on a central stainless steel rod, see Fig. 4.5, with 11 stain­less steel grids on each bundle. The overall length of the assembly is 10.01 m (Kupalov-Yaropolk et al., 1998).

The loss of coolant flow decreases heat transfer from the fuel, increases the fuel temperature and causes a significant temperature rise of the clad­ding (Strasser et al, 2010b). The decrease in system pressure causes a pres­sure drop across and a hoop stress in the cladding. The result is the creep

deformation or ballooning of the cladding. Depending on the temperature, the cladding ductility and the rod internal pressure, the cladding will either stay intact or may burst. Ballooning of the fuel rods may result in a block­age of the coolant sub-channel that, in turn, may impact the fuel coolability. If large fuel clad burst strains occur at the same axial elevation, co-planar deformation in the fuel assembly can result and the coolability may be sig­nificantly degraded. The extent of the ballooning is also dependent on the fuel clad hydrogen content (picked up during the water-zirconium alloy corrosion reaction during reactor operation prior to LOCA). Hydrogen decreases the ala+в phase transformation temperature, which means that increasing the hydrogen content in the fuel cladding will lower its ductility and result in more fuel rod bursts during a LOCA.

The five power-law creep regime is observed at higher stresses and lower temperatures. This creep regime is generally controlled by dislocation climb but describing it as dislocation climb-controlled creep is not appropriate. This is because some of the other creep regimes such as GBS could also be dislocation climb-controlled.

The five power-law creep regime is commonly displayed by class-M alloys and can be described by

[3.27]

where A isa constant and Qc is the activation energy corresponding to the rate controlling mechanism. The activation energy indicated in Equation [3.27] is the apparent activation energy. This is because the effect of tem­perature on the elastic modulus is not included here and that could have a substantial effect.62 The true activation energy can be obtained from the following equation

[3.28 ]

Here E is temperature-dependent modulus of elasticity,

[3.29]

The temperature-normalized stress (a/E) term includes the effect of tem­perature, and thus the value of Q’c obtained from Equation [3.28] is the true activation energy. The activation energy thus obtained has been found to be equivalent to the lattice self-diffusion activation energy.

Careful metallographic examinations show hydrides to be short, thin plate­lets that have precipitated along a variety of crystallographic planes, very

commonly on grain boundaries or intragranularly. For pure zirconium the most common habit plane is near [10І0] and for Zircaloy or alloys it is [10І7], which is ~15° from the basal plane. Intragranular precipitation is less com­mon and is more likely to occur in materials with large grain size, at inter­metallic particles, at dislocations or as a result of very rapid cooling rates. Factors that determine the orientation of the precipitating hydrides in addi­tion to grain size include stress, texture and cold work. As a result, the fab­rication process will have a strong effect on the hydride orientation. Good reviews of the various hydride precipitates are given by Ells (1968), Cox and Rudling (2000) and Coleman (2003). Figure 4.37 illustrates hydrides in 3-dimensions after the metallographic etching process, which exaggerates the actual length of the hydrides.

The orientation of the hydride platelets that form during normal reac­tor operation, preferentially near the cooler cladding OD, usually have axial-circumferential orientation in tubes (respectively axial-tangential ori­entation in strips) and they remain so during wet storage of the spent fuel. The hydrides can become oriented in the radial (through thickness) direc­tion if they are precipitated during operation under high tensile stresses, for example from fuel swelling, or precipitated from solid solution by cooling the alloy from a higher temperature under a tensile hoop stress. Reorientation could occur during reactor operation during cool-down or power cycling, although it is generally unlikely. However, it can occur during dry storage if the internally pressurized cladding is at a high temperature, holds sufficient hydrogen in solution and is then cooled under a sufficiently high hoop stress. The hydrides in solution will precipitate in the radial orientation, while the hydrides that did not dissolve will remain in their original circumferential orientation.

This is most likely to occur during rapid cool-down from high tempera­tures when a cask drying or evacuation procedure is applied rather than during storage.

4.30 Uniform (top) and total elongation (middle) and reduction of area (bottom) as a function of hydrogen for unirradiated and irradiated Zircaloy-2 tested at 332°C (605K) (circumferential hydrides) (Wisner & Adamson, 1998).

250	500 Hydrogen, ppm	750	1000
0

4.31 Uniform (top) and total elongation (middle) and reduction of area (bottom) as a function of hydrogen for unirradiated and irradiated Zircaloy-2 tested at 22°C (295K) (circumferential hydrides) (Wisner & Adamson, 1998).

4.32 Local fracture strain versus hydride blister thickness for both cold worked stress relieved (CWSR) and recrystallized (RX) Zircaloy-4 sheet tested at either 25°C or 300°C. All data are for 3 mm blisters (Pierron et a!., 2003).

4.33 Strength as a function of hydrogen content for irradiated and unirradiated Zircaloy-2 tested at 332°C (605K) (Wisner & Adamson, 1998).

The factors that affect hydride reorientation in irradiated cladding are:

• hoop stress, or tensile and compressive stresses;

• maximum temperature;

• cool-down rate and final temperature;

4.34
Stress-strain response of hydrided Zircaloy-4 tubes stressed in circumferential direction at room temperature. Specimen R32AC has the most radial hydrides (Hong and Lee, 2005).

4.35 Hydride distribution in the radial-circumferential plane of SRA Zircaloy-4 (a) as received, (b) R21AC, (c) R32AC and (d) R43AC (Hong & Lee, 2005).

100 pm

4.36 Typical hydride orientation in a cold worked and stress relieved Zircaloy-4 cladding (~230 ppm hydrogen) (Chu et al., 2008). [2]

• texture;

• time.

The radial hydrides visible in metallographic cross sections can be present in a wide variety of sizes and distributions as well as fractions of the total hydrides present.

Radial hydrides in zirconium alloy cladding are undesirable because they reduce the critical stress intensity required to propagate a radial crack through the wall of the cladding during handling or transportation, as shown by mechanical property data in the previous section. This is illustrated in Fig. 4.38, where it is seen that cracks propagate along radial hydrides, but are blunted in the circumferential hydride region. For this reason considerable attention and effort is expended to define the conditions for radial hydride formation and evaluate their effect on mechanical properties and the per­formance of the fuel, particularly during hypothetical accidents. One of the objectives of the dry storage regulations in the United States is to limit the conditions that could result in hydride reorientation and affect fuel recon­figuration during handling and transport.

Since one of the preferred hydride sites is the grain boundary, RXA material with equi-axed grains is more susceptible to radial hydride forma­tion compared to SRA material with grains elongated in the axial direction. This is illustrated in Fig. 4.40 for SRA Zircaloy-4 and RXA Zircaloy-2.

Hydride distribution in fuel rods having very high heat flux can be quite complicated, as illustrated for an extreme case in Fig. 4.39 for a Zircaloy-2 BWR cladding with a liner of zirconium. A dense hydride rim is seen at the outer surface (a condition more common in PWR rods than BWR ones), mixed radial and circumferential in the outer interior, a zone denuded of hydrides on the inner interior and substantial hydriding of the inner zirconium liner. High burnup performance of both BWR and PWR rods may be affected by such hydride distribution. One evaluation is given by Garzarolli et al. (2010).