Thermal Performance Analysis of Novel UO2 Pellet containing Tungsten Matrix for Pressurized Water Reactor

Transcription

1 Thermal Performance Analysis of Novel UO Pellet containing Tungsten Matrix for Pressurized Water Reactor Suwardi PTBN National Nuclear Energy Agency, Serpong, Indonesia Abstract THERMAL PERFORMANCE ANALYSIS OF NOVEL UO PELLET CONTAINING TUNGSTEN MATRIX FOR PRESSURIZED WATER REACTOR. Sintered UO exhibits very stable in reactor core and has been used as LWR and BWR fuel for more than 30 year, with continuous improvement from only 8 MWd/kgU up-to 6 MWd/kgU. However, its thermal conductivity is very low, that limits its performance. UO pellet containing Tungsten network is US patent in 005, the network having a high melting point and excellent thermal conductivity is continuously formed around UO grains, which improves considerably the fuel conductivity. The paper presents analysis of thermal performance of the fuel for PWR.. The analysis takes into account the pore and temperature distribution and high burn up effect to pellet conductivity. The thermal diffusivity ratio of novel to conventional fuel has been used as correction factor for the novel fuel conductivity. The gap conductance has been considered invariable since decrease caused by wider gap size related to lower pellet expansion is compensated by increase caused by fewer of refractory fission gas released. The results (LHGR = 47 kw/m, burn-up 40%) show lowering pellet temperature along the radius until 440 K at the hottest position. The analysis is underestimates since the gap conductivity has been treated as decreased by % fission gas released, that is not really since the use of lower temperature, and also decreasing thermal conductivity by porosity formation will much lower. The analysis shows that the novel fuel has very good thermal properties which able to pass the barrier of 65 MWd/kgU, the limit to day commercial fuel. The primary advantage of the novel fuel is cold fuel, the extension of burnup over the barrier which needs fewer natural uranium per energy produced. temperature below 1100 C the threshold of 1% fission gas release at high burnup]. Less fuel loading and improve safety, reliability and operability of PWR.. As a result the usage for PWR will increase PWR competitively. Keywords: Tungsten matrix, UO pellet, thermal performance, PWR fuel. 1. Introduction Fuel cycle cost reduces as discharge burnup increases. The reduction is greater with higher backend cost and much higher with discharge burnup when back-end cost is based spent fuel mass, not energy produced. The economic potential rises the interest for developing high burnup fuel. It is a goal of fuel development to increase the rod power and the burn up with a view to optimizing the economics of fuel assemblies. Overpower Margin Increase, Fuel Reliability Enhancement, Operation Flexibility. However, this causes increased amounts of fission gases to be released, which can have the effect of restricting the burn up. Basically the increase of burnup of the fuel has some influence on the safety of reactor operation and on fuel cycle facilities. The most relevant effect of increased burnup on reactor operation is related to the integrity of the cladding material of the fuel rods. In addition shutdown margin, reactivity behavior and increased heat are relevant. [1]. Some solution for increasing pellet temperature caused burn up increase are by diameter reduction [] pellets with central hole [3] duplex pellets [4] graphite discs between pellets [5,6]. The two first solutions minimize the fuel, so they are not favorable solution. Other pellet having high thermal conductivity has been invented by Song [7]. It is an oxidic fuel containing tungsten. The tungsten metal having high melting temperature and better dissipate thermal energy than the fuel (UO, PuO, ThO). Particularly, the nuclear fuel body comprises tungsten network that is continuously formed between fuel grains and thus envelops fuel grains (see Fig. 1). As shown in Fig.1.b, tungsten network is continuous between fuel grains throughout the entire uranium oxide sintered body and acts as a heat- 71

2 conducting channel. The lines metal network--shown in two-dimensional photograph mean the planes in three-dimensions, and the polygon fuel grains in twodimensional photograph are the polyhedrons in threedimensions. Thus, the metal tungsten has a shape which envelopes the polyhedral grains. As tungsten network is completely formed, nuclear fuel materials are in polyhedral tungsten and such tungsten polyhedrons are continuously gathered together to constitute a sintered body. The sizes of the tungsten polyhedrons are corresponding to those of fuel grains. The fuel grain size depends on the conditions of its preparation method. The fuel grains become larger as additives are added to nuclear fuel materials. As an additive, Nb, Ti, Al, Si or Mg oxide can be used. The grain size of nuclear fuel sintered body provided by the present invention is between 5 and 500 μm in diameter in two-dimensional photograph. Thus, the size of one unit of tungsten network is also between 5 and 500. μm in diameter and the tungsten channel between two neighboring fuel grains has thickness ranging from 0.1 to 0. μm. The amount of tungsten in the fuel body increases with increasing the thickness of tungsten channel and with decreasing the size of fuel grain. On the contrary, the amount of nuclear fuel material in sintered body decreases with increasing the amount of tungsten therein and so does the fission energy produced in fuel body, causing a drop in economical efficiency. Therefore, it is desirable to enhance the thermal conductivity of fuel body by a small amount of tungsten. The nuclear fuel sintered body provided by this invention has the amount of tungsten ranging from 0. to 50% by weight of fuel body. The melting point of tungsten is 3400.degree. C. and the melting point of uranium oxide is 800.degree. C. That is the reason why tungsten is chosen as a network material. Tungsten alloys containing other metals up to about 10% by weight can be used as network materials unless they do not significantly decrease the melting point of tungsten alloys. The nuclear fuel body of the present invention is characterized by being composed of nuclear fuel materials and tungsten network. Uranium oxide or uranium oxide mixture, prepared by mixing one selected from a group consisting of gadolinium oxide, plutonium oxide and thorium oxide with uranium oxide, can be used for the nuclear fuel material. If tungsten is dispersed in a form of isolated particles with no tungsten network, tungsten cannot act as a heat-conducting channel. The nuclear fuel body of the present invention is characterized by having tungsten network, not by tungsten dispersed in a form of isolated particles. Figure 1. Microstructure of UO pellet containing tungsten (a) W dispersion in UO, (b) W network for UO grains. There are two steps of transferring the heat generated from nuclear fuel materials in the nuclear fuel body of the present invention. The heat generated in the nuclear fuel grains by nuclear fission of uranium is conducted to the neighboring tungsten network through uranium oxide (step 1), and the heat is further conducted through the tungsten network having relatively high thermal conductivity and being continuous throughout the entire sintered body, and thereby the heat generated in the center of the body can be conducted to the surface of the body (step ). The thermal conductivity of tungsten is higher by 5 times than that of uranium oxide. Thus, the nuclear fuel body of the present invention has enhanced thermal conductivity. In order to supply the low thermal conductivity of conventional nuclear fuel bodies, every possible factor that might decrease the thermal conductivity have been restricted. Since the nuclear fuel body of the present invention comprises tungsten network, which means heat is conducted through the tungsten network, thermal conductivity does not much affected by the change in the oxygen/uranium ratio and the density of the fuel body. That is, the 7

3 nuclear fuel body of the present invention can be used in the wide range of oxygen/uranium ratio and density. Methodology Calculation of coolant water temperature along rod side related to the model of axial power distribution that is considered as cosines type. The cladding surface to water coolant heat transfer is calculated according to simplified model of Dittus- Boelter. The temperature of cladding surface is determined without taken into account thermal conductivity of CRUD. The gap conductance is calculated as given parameter partial contact of typical rod at given heat rate and high burnup. The gap conductance has been considered invariable since decrease caused by wider gap size related to lower pellet expansion is compensated by increase caused by fewer of refractory fission gas released. It permit to calculate the pellet surface temperature. The radial power distribution is modeled as a linear combination function, fitted from typical density from experimental data. Rod heat transfer in the axial direction is omitted. Heat transfer equation in the fuel pellet is approached by a combination of finite element and finite differences Saturn-FS1 [8]. The temperature distribution in the pellet has been obtained by applying a simple model of steady-state heat transfer of fuel rod at particular level of high burn-up. The model takes into account thermal properties dependent of pellet to local temperature, pore, and burn-up. The thermal analysis of cladding is the same as mentioned in reference above. Since the gap conductance is considered no significant difference, so do for temperature of pellet surface. The model takes into account the correction factor for pellet conductivity related to both pore and temperature distribution and high burn up effect. The thermal conductivity changes of irradiated UO pellets may be related to different phenomena such irradiationinduced point defects, fission products and irradiation-induced micro bubbles. The degradation of thermal conductivity related to high burnup has long been studied, but still all be fully understood. The formation of HBS (rim structure at high burnup) that contains more pore has been studied for long time. Choice pore distribution at high burn-up, model pore and temperature dependent of pellet conductivity, and temperature on cladding thermal conductivity. Pore contribution is modeled as factor to pellet thermal conductivity. Pore distribution is obtained by linear combination fitting an experimental data, and modification for this parametric study. The dependent of pellet thermo-physical properties has been modeled by UO of 95% TD, MATPRO properties. The dependent of thermal conductivity to current temporal and space position to burnup and porosity The pore distribution has been modeled by curve fitting of experimental measurement. por( y) := 0.05 Power( y) if y < Power( y) + ( y.8) 1.6 if 0.8 y Power ( ) if 0.9< y (1) Two models of thermal conductivity of irradiated fuel is applied: a. Correction factor for fuel pore that is modeled as pure pore variable of ( 1 por) Fp := 1 + ( por + 10 por ) () b. Correction factor of pore that depends on pore and a coefficient depending temperature. The measured thermal conductivities were normalized to the values of 96.5%TD (TD: theoretical density) by using the Loeb's equation: λn=λ m ( ε)/(1-εP) (3) where: λn is the thermal conductivity normalized to that of 96.5%TD; μm, the measured thermal conductivity; ε, the parameter which express the effect of pore shape on the thermal conductivity of pellets; is P, the porosity evaluated from the sample density. The parameter expressed as follows [9]: ε = (T(K)-73.15), (4) P = 1 TD (5) The last two models may be unified as correction factor of Waisenak Fp = (1-β.P) / (1-0.05*β) (6) 3 β = T T=Tc (7) K = K95 % TD (1 βp) /( β) ( W / mk) (8) MatPro v 9.0 [10] model of temperature dependent of thermal conductivity of fresh/un-irradiated UO fuel has been chosen. The (qv(r)) is volumetric power density profile according to radial coordinate, be modeled as polynomial eq.10 that is fit of typical power distribution. qv(r) = p. v(r) (9) For high burn-up the polynomial is expressed by eq.- 11., 0,373 p = 0, 0,410 ( x) v 1 1 = x 4 x and (10) 73

4 The correlation between linear power density / LHGR qr(r) and volumetric power density qv(r) is: qr ( r) =. π. r. qv( r). dr (11) A finite element approach is applied for the radial distribution of fuel temperature. The radial space is discretized into nr element of linear LaGrange type. Fuel temperature in each element is defined as according to SATURN- FS1 algorithm, that give a solution of pellet temperature as eq.1 T Qv Tk Dk. nr i i nr i BB λ ( T ( r), por( r)).[ Qv. R. { D.( AA + A. pm) }] k = + where Q, R, A, B, AA, BB, F and G are numerical variable mentioned some where [8,11] f.p ( por) por 0.6 Figure 4. Model of pore correction factor Figure 5a shows plot of thermal diffusivity data from room temperature to 1500 o C for UO pellet (noted UO square red) and for UO pellet containing tungsten network (noted UOW diamondblue). The data for Figure 5a has been obtained by using the same measurement method [7]. 3. Results and Discussion Figure shows the effect of discretizing the spatial dimension to the numerical results. It suggests the use discretization of 50 to 100 radial elements for good results Txxx.0inR Th.Conduktivity, W/m/K UO/W 3 UO Temperature, C nrr inr 10 Figure. Effect of discretization on calculated temperature Figure 3 visualizes the pore distribution at the state, a fitting data of high burnup pore distribution on UO fuel por( xr) xr 1 Figure 3. Fitting pore radial distribution. The pore correction factor equation is plotted in Figure 4. For 0% porous pellet near surface zone, the conductivity drops about 50%. Figure.5. Thermal diffusivity of fresh pellet containing and not containing metal network (upper) 9%w W matrix [7] (lower) different alloy matrix [1] It appears that the thermal conductivity of fuel without metal network is lower than usual about absis scale of Figure 5a. This thermal property is comparable to that s composite fuel invented by Dorr et. al shown in Figure 5b [1]. Note, that Fig.5b is thermal conductivity in radial direction. Dorr et all invented UO pellets containing alloy network. The ratio between the two conductivities of pellet with to without tungsten network, has been fitted to a square fitting the ratio resulted equation for 74

5 the diffusivity ratio to temperature as : kth(t,oc) = - 3.E-07.T^ T , with a good correlation coefficient of The diffusivity ratio increases with increasing temperature. This property is beneficial since by implementing this network the hot zone will be better lowering than in the colder zone. The ratio of thermal conductivity of fuel invented by Dorr also have same tendency, i.e., increase with temperature. The radial distribution of pore correction for the high burnup tungsten matrix UO fuel is shown in Figure 6 (y1 axis) together fresh UO fuel (Y axis). It shows an important conductivity drop in the rim zone for high burnup fuel, but the tungsten matrix still dissipates heat better than by UO fresh fuel. The high burnup tungsten matrix UO fuel of upper curve of Fig.6 includes overall corrections of temperature, burnup, and pore distribution. Corrected kth, W/m/s kth_p_bu_w i ( ) 0 i r r Radial position, mm k.ref ( Tp i ) Figure 6. Radial distribution of thermal conductivity (a) New fuel high burnup, (b) UO fresh. The thermal analysis of fuel comprising 9%w of tungsten network has been carried out by using typical data of UO pellet and the thermal conductivity has been calculated by applying thermal conductivity ratio to correct the new fuel conductivity. Pellets partially containing tungsten network also have been analyzed. The results is presented in Figure 7. Temperature, oc Comparison Radial Distribution of Temperature TW00 i 150 TW100 i i % Radial Position Figure 7. Radial temperature distributions of UO pellets(tw00) and UO W matrix (TW100), LHGR 47 kw/m, burnup 40 MWd/kgU 100 kth(t(r)), W/m/s The upper curve noted TW00 in Fig.7 shows the radial distributions of pellet temperature for in UO pellet without tungsten network. The lower curve noted TW100 is for pellet fully, all radius, contained 9%w tungsten. In this case the temperature is lower and maximum lowering temperature at the fuel center is about 440 K. The upper curve noted TW00 in Fig.4 shows the radial distributions of pellet temperature for in UO pellet without tungsten network. The lower curve noted TW100 is for pellet fully, all radius, contained 9%w tungsten. In this case the temperature is lower and maximum lowering temperature at the fuel center is about 440 K. Since the tungsten reduces U contain and U contain in the rim zone of pellet will be burn much more than other zone, it is interesting to apply W network in the out-side of rim zone. UO pellets contained only partially in radial fraction. In addition of temperature limit of fuel in fuel safety criteria, a lower pellet temperature reduces the mobility of the fission gases in the fuel and thereby lowers the rate at which fission gases are released. The lower overall heat content of pellets with an increased thermal conductivity improves the fuel assembly performance under accident conditions (LOCA and RIA) by lengthening the time before the fuel assembly is destroyed. A lower central temperature with otherwise identical fuel properties also reduces what is known as the hour-glass effect, which has an adverse effect on the pellet cladding interaction (PCI) properties of a pellet. It seem the potential use of the new pellet that may change fuel peformance. Tungsten network channel is not a line but a plane in three-dimensions, so point damage of a plane does not much degrade the heat conduction through the tungsten channel. The nuclear fuel body of the present invention includes the tungsten network throughout the entire fuel body. Some variations are possible in the spirit of this invention. The tungsten network can be formed in local regions of the fuel body. Especially, it is possible to prepare a cylindrical fuel body in which an inner cylinder has tungsten network but the outside ring does not. This design may has beter performace, since at high burnup, the rim zone contain nearly half of burned fuel and the quantity of fuel is not change because without W network, where the temperature is low. Result obtained by Tulenko et al. for improving thermal performance of fuel rod by applying metal liquid bond between pellet and cladding for 6 kw/ft ~ kw/m power rating is showed A lowering temperature arround 350 K [13]. It is roughly comparable to application of tungsten network about half portion of inner pellet. 75

6 4. Conclusion Thermal performance of new fuel UO containing tungsten network invented by Song et al. has been carried out for steady state at high burnup. The thermal conductivity of the fuel is based by thermal conductivity ratio of new to typical UO for different temperature which tends increasing with temperature. Beside local composition and temperature, the analyses take into account the influent of porosity and burnup to conductivity. The novel pellet containing tungsten network permit lowering 440 K of pellet center temperature. That give more safety and reliability and permet to be burned over 65 mwd/kgu. It will lower the generating cost, and need less natural uranium as increasing burnup. Since important part of burned fuel located in pellet rim zone, it is recommended to evaluate the thermal performance of pellet fuel where only inner part containing W network for diference portion of purely UO fuel. A detailled analysis will give more lowering temperature of pellet, i.e, advantage since less fission gas released at lower pellet temperature that is less refractory gas in the gap and plenum. References 1. KIM KYU TAE, (1990), Nuclear Fuel Design and Fabrication, in Nuclear Fuel Engineering,, KHNP. W.D. MANLY, (1956), "Fundamentals of Liquid Metal Corrosion", Corrosion, 1, 46-5,. 3. J. REST,( 1975), "SST: A Computer Code to Predict Fuel Response and Fission Product Release from Light Water Reactor Fuels During Steady-State and Transient Conditions," Trans. Am. Nucl. Soc. (1), , November. 4. M.G. ANDREWS, H.R. FREEBURN, AND S.R. PATI, (1976), "Light Water Reactor Fuel Rod Modeling Code Evaluation, Phase II Topical Report", CENPD-18, Combustion Engineering, Inc., Appendix A, April. 5. P.E. MACDONALD, ed., (1976), "MATPRO: A Handbook of Materials Properties for Use in the Analysis of Light Water Reactor Fuel Rod Behavior", ANCR-163, NRC-5, Aerojet Nuclear Company, February. 6. PA , (1975) "Derivation of a Burnup Dependent Fission Gas Release Model for Use in the PAD Fuel Performance Code", Westinghouse, August. 7. United States Patent (005). 8. H.J., RITZHAUPT, et all., (1993), SATURN=FS 1 A computer code for Thermo-mechanical Fuel Rod Analysis, Kernforschungszentrum Karlsruhe 9. J.S. TULENKO AND R.G. CONNELL, An Innovative Fuel Design Concept for Improved Light Water Reactor Performance and Safety, Technical Report No. DE-F60-9ER insc.anl.gov/matprop/uo/cond/solid/index.php 11. SUWARDI, (006), Influence of Pore Distribution Model on Temperature Prediction of High Burn-up UO Fuel, TM. On Economy and Modelling High Burnup Nuclear Fuel, IAEA,.September 1. United States Patent Application (005). 13. J.S. TULENKO AND R.G. CONNELL, An Innovative Fuel Design Concept for Improved Light Water Reactor Performance and Safety, Technical Report No. DE-F60-9ER