Fission gas release and temperature data from instrumented high burnup LWR fuel XA0202217 T. Tverberg, W. Wiesenack Institutt for Energiteknikk, OECD Halden Reactor Project, Norway Abstract.The in-pile performance of light water reactor fuels with high burnup is being assessed as part of the experimental programme of the Halden Reactor Project. To this end, fuel segments pre-irradiated in commercial LWRs are instrumented with fuel centre thermocouples and pressure transducers in order to obtain data on two key performance parameters, namely fuel temperature and rod pressure or fission gas release. The paper describes the results of a re-irradiation of fuel with burnup 59 MWd/kgUO 2 as related to the initial re-irradiation startup and power cycle in the Halden reactor. With emphasis on fission gas release behaviour, one can determine from the observations: the point of onset of fission gas release; restricted axial gas transport in high burnup, bonded fuel which is a characteristic feature of such fuel; the release of trapped fission gas to the plenum volume during power reduction; the response of the fuel temperatures to a degraded gap conductance as the released fission gases mix with the fill gas. The fission gas release data indicate that the onset of release is lower than extrapolated from medium burnup experience. During operation, when the fuel-cladding gap is tightly closed, only minor amounts of released gas reach the plenum. Fuel temperatures remain unaffected in this case. Even during power reduction, when gap conduction is changed, the fuel temperatures are not significantly affected by fission gas release due to a small gap. In total, the data provide valuable insight into the in-pile performance of high burnup fuel and extend the basis for model development and verification. 1. Introduction The experimental programme of the Halden Reactor Project (HRP) has for several years focused on high burnup effects. The objectives of the test programme include extending the data base of UO2 fuel performance, assessing the influence of fuel microstructure on pellet-cladding mechanical interaction (PCMI) and fission gas release in medium and high burnup fuel, investigating integral fuel rod behaviour at high burnup, investigating rim effects. The work in the areas mentioned above is mainly performed through re-instrumentation of pre-irradiated LWR fuel segments, taken from PWRs as well as BWRs, for irradiation in the Halden Boiling Water Reactor (HBWR) and then collecting on-line data of measured parameters for subsequent storing in the Halden Test Fuel Data Bank (TFDB) system. The instrumentation that has been used by the Halden Project over the years in this respect include: fuel thermocouples (TF) for investigation of thermal behaviour of the fuel, cladding extensometers (EC) for studies of PCMI, pressure transducers (PF) for internal rod pressure measurements, fuel stack extensometers (EF) for densification and swelling assessment, diameter gauges for in-pile measurements of fuel rod diameter changes.
This paper will address one such experiment containing a fuel segment previously irradiated in a commercial LWR and re-instrumented with a TF and a PF. The discussion and analysis will mainly focus on the determination of point of onset of fission gas release and comparing with known relations and the influence of fission gas release on the thermal behaviour of high burnup fuel. 2. Description of experiment The irradiation rig contained two fuel segments from a 8 x 8 BWR rod. When discharged from the BWR, the segments had achieved a burnup of ~59 MWd/kgUO2. Towards the end of the commercial irradiation, the segments were running at a very low power: -12 kw/m. Non-destructive post irradiation examination (PEE) was performed before shipment to Halden. Of the main results from this examination it is worth to mention about 3.3% fission gas release and an outer oxide layer of ca. 43 urn. Cold gap measurements were also performed and an average cold gap (diametral) of ca. 30 urn was found. This corresponds to a closed gap at around 11 kw/m. These and further data on the rods are summarized in Table I. Table I. Properties of the fuel segment Base power (commercial irradiation) [kw/m] Burnup after commercial irradiation [MWd/kgUC>2] Outer oxide layer after comm. irr.[um] Enrichment at BOL [w/o 2i5 U] Density [% of T.D.] Active length [mm] TF centre hole diam.[mm] Pellet outer diam. [mm] Clad, inner diam. [mm] Clad, thickness [mm] Diametral gap at BOL [urn] Diametral gap before HBWR irradiation (from PIE) [um] Filler gas/pressure [bar] Free volume at start of HBWR irradiation [cc] Fuel weight [kg] 12 59 43 3.35 95.7 327.5 2.5 10.44 12.25 0.8 210 30 He/5.0 5 0.287
Before loading in the HBWR, both rods were re-instrumented with a TF and a PF, hence allowing monitoring both temperature and pressure data. Figure 1 shows a schematic of the rig that was used for the irradiation. The TF is situated at the top of the rod in a ~35 mm deep centre hole, 2.5 mm in diameter. The remaining fuel pellets are solid. The instrumentation in the rig include 5 vanadium neutron detectors (ND) for power monitoring. The NDs are positioned at different axial and radial positions hence allowing for calculating a power distribution, hi addition, the rig is equipped with inlet and outlet thermocouples and flow turbines which together with the calibration valve at the inlet is used for the calorimetric power calibration which was performed at an initial stage in the HBWR irradiation. r)5*itutt for e OECO HALOEN REACTOR PROJECT INTEGRAL FUEL ROD BEHAVIOUR TEST IFA-597.2 Outlet Turbine Flowmeter Outlet Coolant Thermocouples Fuel Centerline Thermocouples Vanadium Neutron Detectors Shroud Fuel Rods Bellows pressure transducers - Inlet Coolant Thermocouples Inlet Turbine Flowmeter JLM Calibration Valve FIG. 1. Schematic of the rig. 3. Operation in the HBWR Figure 2 shows the operation history of the rod during the irradiation in the HBWR. We note the power cycles during the first 3 to 7 days, which are due to power calibration of this and other rigs during the start of the irradiation cycle. Three such short cycles can be seen during the early stages of irradiation and the average linear heat rate (ALHR) of the rod reaches -20.5 kw/m, 18.5 kw/m and 18 kw/m respectively as a maximum during these three ramps before being reduced to zero. The corresponding fuel centre temperatures, as measured by the thermocouples, are 830 C, 780 C and 760 C. The internal rod pressure follows the rod power during these cycles. Little or no indication of fission gas release can be seen.
Following this initial 'conditioning' phase, after a ~4 day power shutdown the power is again increased to about 25 kw/m where it is kept for a period of ca. 10 days, before the final shutdown cycle. During this shutdown, there is an intermediate power increase (from 11 to 17 kw/m) before the final shutdown after a total of 21 days of irradiation. It is seen that while the fuel temperature essentially remains constant throughout the steady-state phase of the cycle, the rod internal pressure increases significantly during the shutdown. This development will be discussed in more detail below. o 15 10l. n N A, 2000; i500 S 1000 Q. 2 it 500 0 35 \ S\ 25r 20 T 8 10 12 14 16 18 20 22 Time (days). 2. Irradiation history. The dotted temperature curve shows the calculated peak temperature. Note the pressure increase coinciding with the power shutdown at around 21 days. 4. Pressure data Figure 3 (pressure versus power for the first power cycle) shows that there is little fission gas release. The pressure change indicates a release of about 0.5%. No further release is registered for the next two cycles which went to powers slightly below those of the initial cycle. For the last cycle, this changes however, as was seen in Fig. 2. The pressure increases slightly during the last few days at power, and then increases rapidly when the power is reduced during the shutdown. In Figure 4, the same period is shown in the pressure power domain. Initially at zero power, the pressure is at 9 bar and at the peak power of-26 kw/m the pressure is ~ 10. 2 bar. During the steady state period with about 25 kw/m, the pressure increases only little by about 0.5 bar. The pressure continues to increase to -15 bar as the power is reduced further to ~12 kw/m. When the power is kept at this intermediate power level for a short period, the pressure still 10
increases (~0.5 bar). Finally the power is again increased to about 18 kw/m before the final shutdown occurs and the power is reduced to zero upon which the pressure is ~ 14.8 bar. The total increase in pressure at zero power is thus 5.8 bar during the period. This behaviour is typical of what is often referred to as delayed measured fission gas release which is seen for high burnup fuel rods. In this case the gap will be tightly closed at power, thus leaving little room for the gas to diffuse into the plenum and hence to the pressure detector. Only when the power is reduced and the gap opens, can the released gas be detected by the transducer. The pressure increase at 20 C is 4.1 bar between the initial startup and the final shutdown. Using a burnup of 59 MWd/kgUC>2 and a value for produced fission gas of 31 cc per MWd together with the values for fuel mass and initial free volume from Table I, we obtain a fission gas release of 3.9%. It can be inferred that the onset of fission gas release occurs at an average linear heat rate somewhere between 18 and 25 kw/m, corresponding to calculated maximum fuel temperatures in the solid pellets of -880 and 1170 C, respectively. 17 18-: I'" Power increase - - Power decrease to Q. T5 12- O c to4 10 20 25 30 Average heat rate (kw/m) FIG. 3. Pressure versus rod power during first cycle. 17-1-4 15 20 Average heat rate (kw/m) 25 so FIG. 4. Pressure versus rod power during the last operation cycle. 11
The two release fractions which can be deduced from this fuel rod, together with the result from a sibling rod are shown in Fig. 5. The interpolation curve indicates that the temperature of 1 % release is below the Halden 1 % fission gas release treshold curve defined as [1] 9800 T =-( W] t c] (i) lnl l 1000. 1000. where T is the fuel centreline temperature for 1% fission gas release and BU is the burnup in MWd/kgUO2. Eq. (1) predicts a treshold temperature of 1050 C for fuel with a burnup of 59 MWd/kgUO 2. 5. Temperature data Figure 7 shows the temperature-power relation for the whole of the last power cycle seen in Fig. 2. It is seen that the release of fission gas during the last cycle induces only a small effect of thermal feedback on the temperatures between the startup and the shutdown. Included in the figure are lines of 2nd order least squares fit through the startup and shutdown data, respectively. At 15 kw/m, the temperature at the thermocouple position is ca. 25 K higher for the down-ramp compared to the same power of the up-ramp. In analysing the measured fuel temperatures, it is important to have a good estimate of the fuel-to-clad gap. As mentioned above, PEE was performed on these segments after unloading from the BWR and a cold (diametral) gap of 30 urn was found. Gap closure can also be confirmed by looking at clad elongation measurements on a sibling rod that was irradiated in a different loading of the same IFA. This is shown in Fig. 6, where clad elongation data and gap prediction are plotted versus rod average linear heat rate for the first 3 power ramps above the power level the rod saw at its final stage of commercial irradiation. For the first ramp, the cladding elongation curve starts to deviate from the calculated curve of cladding free thermal expansion indicating onset of PCMI for powers > 10 kw/m. LWRFTEMP, a modified version of the HRP's FTEMP2 steady state fuel modelling code, was used with this input for analysing the measured fuel temperatures. LWRFTEMP is specially tuned to properly handle the radial burnup and plutonium distribution in rods preirradiated in a LWR to burn-ups beyond where rim structure formation occurs. The code uses the TUBRNP model [2] for radial distribution of plutonium and burnup in high burnup UO2 fuel, and the conductivity degradation model derived from other Halden data [3]. Figure 8 shows the LWRFTEMP calculations for the last power cycle. For the final shutdown sequence a Xenon content of 45% is assumed, as derived from the pressure data. Complete mixing of the gas is assumed. Good agreement between measured and calculated temperatures is achieved. It should be noted that the small difference between the case with pure helium and the case with considerable admixture of fission gas can only be obtained if the roughness of fuel and cladding is not increased as is the case in some gap conductance models. In the upper curve, the calculated diametral fuel-clad gap is shown. The predicted power for gap closure is about 11 kw/m, which is about the same power at which the rod was running during the end of commercial irradiation. 12
1500 <F 1200 I 900- E c 600 ' Peak temperatures A A A Vilanaa-cuive at 59 MVv'd'kgUC, 2 3 Fission gas release {%) FIG. 5. Comparison of measured fission gas release data with the release treshold curve at 59Wd/kgUO2..6-:.5 Sscond tarns: SMWAUO 300-.4-.2-.1 0.8T Down ramp -.]+ 0 10 20 Linear heat rate (kw/m) 30 FIG. 6. Cladding elongation versus average rod power for sibling rod in later loading. The elongation curve starts to deviate from the calculated curve of free thermal expansion, indicating PCMI, at about the same power as L WRFTEMP predicts gap closure. 13
r & 10-E 1100 10 15 20 25 1000- Shut down Start up 200 10 15 Local heat rate (kw/m). 7. Measured fuel centre temperatures versus local power during the last operating cycle. The curves are second order least squares fits for the startup and shutdown sequences as indicated in the figure. Because of the poisoned gap, temperatures during the shutdown sequence are higher than during the startup: -15 "C at a linear heat rate of 15 kw/m. 14
10 15 Local heat rate (kw/m) 25 FIG. 8. Fuel centre temperature calculated and measured at thermocouple position versus local power during the last operating cycle. SUMMARY AND CONCLUSIONS For the first period of operation, a small amount of FGR is observed; At high power, the fission gas is trapped inside the fuel rod, with no communication to the plenum and hence the pressure transducer; A significant release can only be detected when the power is reduced and the pelletcladding gap opens; Because of the tightly closed gap at power, an almost 50% fission gas mix has only a small effect on the fuel centre temperature. 15
REFERENCES [1] VITANZA C, KOLSTAD E., GRAZIANI U., "Fission gas release from UO 2 fuel at high burnup", ANS topical meeting on light water reactor fuel performance, Portland 1979. [2] LASSMANN K., OVARROLL C, van de LAAR 1, WALKER C. T, "The radial distribution of plutonium in high burnup UO2 fuels", Journal of Nuclear Materials, 208 (1994), 223-231. [3] WIESENACK, W., "Assessment of UO 2 conductivity degradation based on in-pile temperature data", International Topical Meeting on Light Water Reactor Fuel Performance, Portland, March 1997. 16