Annals of Nuclear Energy 37 (2010) 13411350

Contents lists available at ScienceDirect

Annals of Nuclear Energy

journal homepage: www.elsevier.com/locate/anucene

Assessment of a RELAP5 model for the IPR-R1 TRIGA research reactor

Patrcia A.L. Reis a,b, Antonella L. Costa a,b,*, Clubia Pereira a,b, Maria A.F. Veloso a,b, Amir Z. Mesquita c, Humberto V. Soares a,b, Graiciany de P. Barros a,b
a Departamento de Engenharia Nuclear Escola de Engenharia da Universidade Federal de Minas Gerais, Av. Antonio Carlos No. 6627, Campus Pampulha, PCA 1, CEP 31270-901, Belo Horizonte, MG, Brazil b Instituto Nacional de Cincias e Tecnologia de Reatores Nucleares Inovadores/CNPq, Brazil1 c Centro de Desenvolvimento da Tecnologia Nuclear CDTN/CNEN, Av. Antnio Carlos, 6627, Campus UFMG, Belo Horizonte, Brazil

a r t i c l e

i n f o

a b s t r a c t
RELAP5 code was developed at the Idaho National Environmental and Engineering Laboratory and it is widely used for thermal hydraulic studies of commercial nuclear power plants and, currently, it has been also applied for thermal hydraulic analysis of nuclear research systems with good predictions. This work is a contribution to the assessment of RELAP5/3.3 code for research reactors analysis. It presents steadystate and transient calculation results performed using a RELAP5 model to simulate the IPR-R1 TRIGA research reactor conditions operating at 50 and 100 kW. The reactor is located at the Nuclear Technology Development Centre (CDTN), Brazil. The development and the assessment of a RELAP5 model for the IPRR1 TRIGA are presented. Experimental data were considered in the process of code-to-data validation. The RELAP5 results were also compared with calculation performed using the STHIRP-1 (Research Reactors Thermal Hydraulic Simulation) code. The use of a cross ow model has been essential to improve results in the transient condition respect to preceding investigations. 2010 Elsevier Ltd. All rights reserved.

Article history: Received 18 February 2010 Received in revised form 14 May 2010 Accepted 18 May 2010 Available online 15 June 2010 Keywords: RELAP5 IPR-R1 TRIGA Thermal hydraulic

1. Introduction Interest in safety issues of nuclear research reactors is nowadays increasing due their enlarged commercial exploitation commonly devoted for generation of neutrons for different scientic and social purposes (Adorni, 2007; Bokhari et al., 2002; Khater et al., 2007; Khedr et al., 2005). The main activity of a nuclear research reactor is not for energy generation reaching maximum power operation of about 100 MW. In spite of this, specic features are necessary to ensure safe utilization of such installations. Therefore, several codes have been used focusing special attention for research reactors safety analysis and valuation of specic perturbation plant processes. The RELAP5 system code was developed to simulate transient scenarios in power reactors such as PWR and BWR but recent works as, for example (Antariksawan et al., 2005; Khedr et al., 2005; Marcum et al., 2010), have been

* Corresponding author at: Departamento de Engenharia Nuclear Escola de Engenharia da Universidade Federal de Minas Gerais, Av. Antonio Carlos n 6627, Campus Pampulha, PCA 1, CEP 31270-901, Belo Horizonte, MG, Brazil Tel.: +55 31 34096688. E-mail addresses: patricialire@yahoo.com.br (P.A.L. Reis), lombardicosta@gmail. com (A.L. Costa), claubia@nuclear.ufmg.br (C. Pereira), dora@nuclear.ufmg.br (M.A.F. Veloso), amir@cdtn.br (A.Z. Mesquita), betovitor@ig.com.br (H.V. Soares), graicibarros@yahoo.com.br (Graiciany de P. Barros). 1 http://www.cnpq.br/programas/inct/_apresentacao/inct_reatores_nucleares. html. 0306-4549/$ - see front matter 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.anucene.2010.05.013

performed to investigate the applicability of the code to research reactors operating conditions (TRIGA 2000, MTR, Oregon State TRIGA), respectively. Specically, the Training, Research, Isotope, General Atomic (TRIGA) reactors are constructed in a variety of congurations and capabilities, with steady-state power levels ranging from 20 kW to 16 MW offering true inherent safety. TRIGA is a pool type reactor that can be installed without a containment building being designed for use by scientic institutions and universities for purposes such as graduate education, private commercial research, non-destructive testing and isotope production. In the present work, the IPR-R1 TRIGA reactor, Mark-I model, installed in Brazil, in operation since 1960, has been modeled for RELAP5 code with the aim of to reproduce the measured steady-state as well as transient conditions. The development and the calculation for the validation of a RELAP5 model for the IPR-R1 TRIGA research reactor have been presented. The version MOD3.3 was used to perform the simulations. The model validation at 50 kW was presented in a preceding work (Costa et al., 2010) using experimental data and also comparisons with data calculated using the STHIRP-1 code (Veloso, 2004). However, the modeling was not able to predict transient conditions adequately. Modications were then performed in the model mainly respect to the pool nodalization adding a cross ow model to permit a better heat removal from the core in natural circulation conditions. The current results obtained with this new nodalization demonstrate that the IPR-R1

1342

P.A.L. Reis et al. / Annals of Nuclear Energy 37 (2010) 13411350

TRIGA model is representative of the reactor behavior considering steady-state and transient operation conditions as it is being described in the next sections. 1.1. RELAP5 and STHIP-1 codes RELAP5 computer code is a LWR transient analysis code developed mainly by the Idaho National Engineering Laboratory (INEL) for the U.S. Nuclear Regulatory Commission (NRC) for use in rulemaking, licensing audit calculations, evaluation of operator guidelines, and as a basis for a nuclear plant analyzer (US NRC, 2001). Specic applications have included simulations of transients in LWR systems such as loss of coolant, anticipated transients without scram (ATWS), and operational transients such as loss of feedwater, loss of offsite power, station blackout and turbine trip. However, some recent works have been performed to access the applicability of the code to research reactors operating conditions. The STHIRP-1 computer program was developed at the Nuclear Engineering Department in the Federal University of Minas Gerais, Brazil, to simulate the IPR-R1 reactor. It uses the sub-channels analysis methodology to simulate, under steady-state and transient conditions, the thermal and hydraulic phenomena occurring inside the core of water cooled research reactor under a natural convection regime (Veloso, 2004). The models and empirical correlations necessary to describe the ow phenomena which cannot be described by theoretical relations were selected according to the characteristics of the IPR-R1 TRIGA research reactor operation. The program was validated against the IPR1 TRIGA model. The calculation results in comparison with the experimental data indicate that the program reproduces the experimental data with good agreements. 1.2. IPR-R1 TRIGA research reactor The IPR-R1 is a reactor type TRIGA, Mark-I model, manufactured by the General Atomic Company and installed at Nuclear Energy Development Centre (CDTN) of Brazilian Nuclear Energy Commission (CNEN), in Belo Horizonte, Brazil. It is a light water moderated and cooled, graphite-reected, open-pool type research reactor.

IPR-R1 works at 100 kW but it will be briey licensed to operate at 250 kW. It presents low power, low pressure, for application in research, training and radioisotopes production. The reactor is housed in a 6.625 m deep pool with 1.92 m of internal diameter and lled with demineralized light water. A schematic reactor diagram is illustrated in Fig. 1. The main aim of the water in the pool is for cooling, as well as moderator, neutron reector and it is able to assure an adequate radioactive shielding. The reactor cooling occurs predominantly by natural convection, with the circulation forces governed by the water density differences. The heat removal generated from the nuclear ssions is performed pumping the pool water through a heat exchanger. The core has a radial cylindrical conguration with six concentric rings (A, B, C, D, E, F) with 91 channels able to host either fuel rods or other components like control rods, reectors and irradiator channels. There are in the core 63 fuel elements constituted by a cylindrical metal cladding lled with a homogeneous mixture of zirconium hydride and Uranium 20% enriched in 235U isotope. There are 59 fuel elements covered with aluminum and four fuel elements with stainless steel. Details of IPRR1 fuel elements are presented in Fig. 2. They have three axial sections, upper and lower reector (graphite), and the central portion lled with fuel (U-ZrHx) which is 0.356 and 0.381 m in length for elements coated with aluminum and stainless steel, respectively. The radial power distribution was calculated in preceding works using the WIMSD4C and CITATION codes (Dalle, 2003; Dalle et al., 2002) and also experimental data (Souza, 1999). The radial factor is dened as the ratio of the average linear power density in the element to the average linear power density in the core. Fig. 3 shows the radial relative power distribution. The main features of the TRIGA IPR-R1 are presented in Table 1 (CDTN/CNEN, 2000). General Atomics developed the manufacturing process of fuel containing high concentrations of hydrogen using a metal alloy of uranium and zirconium, to fulll a core requirement of intrinsic safety (General Atomics, 1959). The characteristic of this fuel is the prompt negative temperature coefcient (1.1 0.2/C) that provides safety and automatically limiting the power when excess reactivity is suddenly inserted (Mesquita e Souza, 2010). The aver-

Fig. 1. Schematic representation of the IPR-R1 (out of scale, measure in mm).

P.A.L. Reis et al. / Annals of Nuclear Energy 37 (2010) 13411350

1343

Fig. 2. IPR-R1 TRIGA fuel element design (measure in mm).

age energy of thermal neutrons follows the temperature in the fuel elements. Therefore, an increase in power causes an increase in the fuel temperature and, consequently, an increase in neutron energy and a decrease in reactivity due to: reduction of the cross sections of 235U due to increased energy of the neutrons. increase of parasitic absorption of neutrons due to Doppler broadening in the resonances of 238U, present in large quantities in the nucleus. increased leakage of neutrons. Therefore, an increase of fuel temperature causes immediate reduction in the neutron population, reducing the reactor power faster than any device engineering. The fuel acts as an automatic power regulator. This propriety of the fuel elements characterizing a high negative prompt temperature coefcient is the main reason of the inherent safety behavior of the TRIGA reactors.

2. Nodalization description Each of the 63 fuel elements was modeled separately and 63 heat structure (HS) components were associated with 13 corresponding hydrodynamic pipe components constituting 13 hydrodynamic channels (201213), as can be veried in Fig. 4. Table 2 presents some characteristics of the 13 regions. Fig. 5 shows the RELAP5 general nodalization developed to simulate the IPR-R1. The reactor pool was modeled using two pipe components, each one composed by ten volumes. As it can be veried in the Fig. 5, both components (0 2 0 and 0 5 0) have their volumes connected by single junctions to characterize a cross ow model. The pool nodalization has been modied respect to a preceding work (Costa et al., 2010) to improve transient predictions as it is being clearly demonstrated in Section 4. A time dependent volume was used to simulate the atmospheric pressure on the pool surface. The natural convection system and the primary loop circulation have been modeled. The secondary loop, composed mainly

1344

P.A.L. Reis et al. / Annals of Nuclear Energy 37 (2010) 13411350

Fig. 3. Radial relative power distribution.

by the external cooling tower was not modeled in the present nodalization because the primary circuit was sufcient to guaranty the heat removal of the coolant. The point kinetics model was used in the current simulations. A detailed representation of each element is, however, essential to properly take into account the radial power distribution associated with the position of the fuel elements. The axial power distribution was calculated considering a cosine prole and taking into account also that the power is cut off in the extremes of the element due the presence of the graphite as it is sketched in Fig. 6. Although the above modeling procedure is approximated, it is used here to conserve the actual axial and radial power distribution xed.

3. Steady-state predictions The validation of a RELAP5 nodalization implicates that the model reproduces the measured steady-state conditions of the system with acceptable margins. The nodalization may be considered qualied when it has a geometric delity with the system, it reproduces the measured steady-state condition of the system, and it demonstrates satisfactory time evolution conditions (DAuria et al., 1999). However, sometimes a nodalization qualied to simulate determined condition may not be suitable to simulate other type of situation being necessary modications and re-qualication.

The RELAP5 steady-state calculation has been performed at 50 and 100 kW. The temperature values at the inlet and outlet of the thermal hydraulic channels three, eight and 13 calculated using RELAP5 can be veried in Tables 3 and 4, for 50 and 100 kW, respectively. The calculated values were compared with the available experimental data (inlet and outlet channel temperature) and with STHIRP-1 calculation data (outlet channel temperature). Chromelalumel calibrated thermocouples were used to collect the coolant temperature data and the measured values have a maximum error of 1 C (Veloso, 2004). As it can be veried in the Table 3, considering operation at 50 kW, the results for both codes are in good agreement in comparison with the experimental data. The error obtained using the RELAP5 calculation is into the range of the maximum acceptable error suggested for coolant temperature (0.5%) by the RELAP5 users (DAuria and Galassi, 1998). STHIRP-1 code reached values of outlet temperature with minimum error demonstrating an excellent reproduction of the reactor behavior. STHIRP-1 was developed specially according with the IPR-R1 TRIGA characteristics and the core region was modeled with 104 TH channels against 13 TH channels in the present RELAP5 model. STHIRP-1 inlet temperature values are not available. The inlet temperature values calculated using the RELAP5 code presented the same values for the three considered channels.

P.A.L. Reis et al. / Annals of Nuclear Energy 37 (2010) 13411350 Table 1 Main thermal hydraulic and kinetic IPR-R1 core data (CDTN/CNEN, 2000). Core power (kW) Delayed neutron fraction Prompt neutron generation time (s) Isothermal coefcient (cents/kW) Temperature reactivity feedback (cents/C) Pressure of operation (kPa) Main moderator Cladding Coolant Reector 250 0 (in process of licensing) 0.0079 100 4 0.44 11 158 7 Zirconium hydride (hydrogen) Aluminum or stainless steel (SS) Light water Graphite Al cladding Number of fuel elements Fuel Zr concentration (% weight) U concentration (% weight) H concentration (%w eight) 235 U enrichment Cladding material Fuel diameter (mm) Total height (mm) Gap width (mm) Total diameter (mm) Active height (mm) Graphite reector height (mm) Gap material 59 UZrHj o 9.10 8.0 1.0 20% A11100-F 35.6 722.4 0.09 37.3 355 6 101 6 Helium SS cladding 4 UZrH16 899 8.5 1.6 20% SS AISI304 36.3 720.6 0.14 37.6 381.0 88.1 Helium 6 5 206 0.002779 226230 Table 2 Main features of the 13 thermal hydraulic (TH) regions. TH channel 1 2 3 4 5 Number of fuel elements 6 5 5 4 5 TH channel identier 201 202 203 204 205 Mass ow area (nf) 0.003595 0 002779 0.002779 0002759 0.002779 HS identier 201206 207211 212216 217220 221225

1345

HS position (see Fig. 3) B1, B2, B3, B4, B5, B6 C2, D2, C3, D3, D4 C4, C5, D5, D6, D7 C6, D8, D9, D10 C8, C9, D11, D12, D13 C1O, C11, D14, 15, D16 C12, D17, D18, D1 E2, E3, E4, E5, F6 E6, E7, E8, E9, F11 E1O, E11, E12, E13 E14, E15, E16, E17, F21 E18, E19, E20, E21, F26 E22, E23, E24, E1, F1 Total = 63

7 8 9 10 11

4 5 5 4 5

207 208 209 210 211

0.002756 0005477 0.005495 0.005436 0.005477

231234 235239 240244 245248 249253

12

212

0.005477

254258

13 Total = 63

213

0.005477

259263

Fig. 4. Representation of the 13 TH channels in RELAP5 model.

1346

P.A.L. Reis et al. / Annals of Nuclear Energy 37 (2010) 13411350

Fig. 5. IPR-R1 TRIGA nodalization in the RELAP5 model.

Table 3 Experimental and calculated results at 50 kW of power operation. TH channel Outlet channel temperature (K) Exp. 3 8 13
a

Inlet temperature (K) Exp. RELAP5 Error (%)a 0.1 0.5 0.5

RELAP5 Error (%)a 0.5 0.5 0.5

STHIRP- Error 1 (%)a 299.0 297.6 298.5 0.4 0.1 0.1

300.0 298.4 298.0 296.4 298.0 296.4

294.1 294.7 296.1 294.7 296.1 294.7

Error = 100 (Calculation Experimental)/Experimental.

Table 4 Experimental and calculated results at 100 kW of power operation. TH channel Outlet channel temperature (K) Exp. 3 8 13
a

Inlet temperature (K) Exp. RELAP5 Error (%)a 0.6 0.1 0.3

RELAP5 Error (%)a 0.9 0.8 1.1

STHIRP- Error 1 (%)a 305.3 300.1 303.6 0.4 0.2 0.6

304.0 301.3 300.5 298.8 301.5 298.8

294.0 295.7 295.5 295.7 296.5 295.7

Error = 100 (Calculation Experimental)/Experimental.

Fig. 6. Prediction of the axial power distribution function in a TRIGA fuel element.

Results performed at 100 kW of power operation are shown in Table 4. The error found for RELAP5 calculation is a few overestimated in comparison with the error suggested for coolant temperature (0.5%) by the RELAP5 users. However, considering the error from the experimental data (1 C) the values predicted using RE-

LAP5 are perfectly acceptable for the present model validation process for operation power up to 100 kW. Figs. 7 and 8 show the RELAP5 calculation for the inlet and outlet temperature for the TH channel 1, at 50 and 100 kW of power, respectively. Such channel was chosen because it concentrates the HS with higher values of radial power. As it can be veried, after about 2500 s of calculation, the temperatures reach steady-state condition. The temperature stable values are in good agreement with the experimental available data. In addition, Fig. 9 shows the fuel and cladding temperature evolution for the heat structure 1 (HS-B1) in the channel one at the mid high predicted using RELAP5. As it can be observed, these parameters are completely stable and the difference of temperature between the two curves is approximately 40 K. The HS-B1 corresponds to the fuel element in the position B1 according to Fig. 3. Fig. 10 illustrates the time evolution for the HS-B1 fuel temperature at four different axial levels. The axial fuel temperature distribution follows the cosine prole function, reaching higher temperatures in the central parts of the element following the same power prole, as it is demonstrated in Fig. 11 for HS-B1.

P.A.L. Reis et al. / Annals of Nuclear Energy 37 (2010) 13411350

1347

Fig. 7. Inlet and outlet coolant temperature for the channel 1 at 50 kW predicted by the RELAP5.

Fig. 8. Inlet and outlet coolant temperature for the channel 1 at 100 kW predicted by the RELAP5.

4. Transient simulation Several transient events are possible to occur in research reactors, as for example, loss of electrical power supplies, insertion of excess reactivity, loss of ow, loss of coolant, erroneous handling or failure of equipment or components, special internal events, external events and human errors (IAEA, 2005). In spite of the IPR-R1 to be inherently safe, situations that can disturb the normal reactor operation are possible to occur. The RELAP5 model presented in this work demonstrated to reproduce very well the steady-state conditions. Therefore, in addition to the validation process of the modeling, a transient event was investigated using the code and the results has been compared with available experimental

data. The investigated event is the forced recirculation off and may be caused by the recirculation pump failure, bringing the reactor to operate in natural circulation conditions. In the experiment, the reactor operated during about 2.5 h with the forced cooling system switched off and with an indication of 100 kW at the linear neutronic channel (Mesquita et al., 2009). The measurements have demonstrated an average temperaturerise rate of about 4.8 C/h. At inlet and outlet of a thermal hydraulic channel the temperature values were veried to increase about 5.3 C/h in both cases (Mesquita et al., 2009). To perform the simulation using the RELAP5, the valve in the primary system (number 600 in the nodalization) has been closed at 3000 s of calculation after the system to reach steady-state

1348

P.A.L. Reis et al. / Annals of Nuclear Energy 37 (2010) 13411350

Fig. 9. Cladding and fuel temperatures time evolution for an element in the TH channel 1 predicted using RELAP5.

Fig. 10. RELAP5 axial fuel temperature prediction at four levels of HS-B1.

condition. After the beginning of the transient at 3000 s, the temperatures increase as consequence of no energy removal from the pool since the primary was off. The coolant temperature at inlet and outlet of the TH channel one then increases gradually with rates of about 4.9 C/h and 4.6 C/h, respectively, demonstrating very good agreement with the experimental available data. Both curves are illustrated in the Fig. 12. The modication of the pool nodalization represented an essential process to improve the transient predictions using the RELAP5. The insertion of the cross ow model makes possible better removal

of heat from the core during natural circulation condition due improvement on the coolant ow between the pool pipe volumes. Fig. 13 illustrates the coolant temperature predicted using the code considering the nodalization presented in this paper with cross ow model and that in the preceding work without cross ow model (Costa et al., 2010), both at 100 kW of power operation. The curves show clearly that the model using cross ow presents a temperature-rise rate (4.9 C/h) much more approximated to the experimental rate (4.8 C/h) than the model without cross ow model (30.0 C/ h), demonstrating the efciency of the changes in the nodalization.

P.A.L. Reis et al. / Annals of Nuclear Energy 37 (2010) 13411350

1349

Fig. 11. Axial fuel, cladding and coolant temperature distributions predicted using RELAP5.

Fig. 12. Inlet and outlet coolant temperature for the channel 1 at 100 kW predicted by the RELAP5 after forced recirculation off at 3000 s.

5. Conclusions In this work, a nodalization for the IPR-R1 TRIGA research reactor performed using the RELAP5/MOD3.3 code was presented as a contribution to the assessment of such code for research reactor safety analysis. The nodalization was validated against experimental data from steady-state conditions at 50 and 100 kW of power operation. RELAP5 results have been also compared with data obtained using the STHIRP-1 thermal hydraulic code. The results demonstrate good agreement between the codes with little discrepancies which could be explained by the different empirical correlations embedded within each code. The little discrepancies can be also related to the differences on the nodalization methodologies adopted for each code. Future investigations will be performed to

verify the effect of the number of TH channels in the calculation results as a sensitivity investigation. Moreover, other transient calculations must be performed and compared with experimental data aiming to complete the code validation process. Considering the three basic aspects necessary to qualify a nodalization for a system (geometric delity, reproduction of the measured steady-state conditions and satisfactory time evolution conditions), it is possible to conclude that the RELAP5 model presented in this work was qualied to represent adequately the IPR-R1 TRIGA research reactor in steady-state as well as in transient situations. Future works about the present model include sensitivity investigations including core cross ow calculation and also analysis adding the coupled neutron kinetic calculation using, for example, the RELAP53D code.

1350

P.A.L. Reis et al. / Annals of Nuclear Energy 37 (2010) 13411350

Fig. 13. Forced recirculation off transientprediction using two different types of pool nodalization.

Acknowledgements The authors are grateful to CAPES, CDTN, CNPq and FAPEMIG for the support. References
Adorni, M., 2007. Accident Analysis in Research Reactors. Laurea Thesis, University of Pisa, Italy. Antariksawan, A.R., Huda, M.Q., Liu, T., Zmitkova, J., Allison C.M., Hohorst, J.K., 2005. Validation of RELAP/SCAPSIM/MOD3.4 for research reactor applications. In: 13th International Conference on Nuclear Engineering, Beijing, China, May 16 20, 2005, pp. 18. Bokhari, I.H., Israr, M., Pervez, S., 2002. Analysis of reactivity induced accidents at Pakistan research reactor-1. Annals of Nuclear Energy 29, 22252234. CDTN/CNEN, 2000. Relatrio de Anlise de Segurana do Reactor TRIGA IPR-R1. RASIN/TRIGA-IPR-R1/CDTN, restrict document, Belo Horizonte, Brazil, pp. 1 321. Costa, A.L., Reis, P.A.L., Pereira, C., Veloso, M.A.F., Mesquita, A.Z., Soares, H.V., 2010. Thermal hydraulic analysis of the IPR-R1 research reactor using a RELAP5 model. Nuclear Engineering and Design 240, 14871494. Dalle, H. M., 2003. Avaliao Neutrnica do Reator TRIGA IPR-R1 com Congurao de 63 Elementos Combustveis e Barra de Regulao em F16. Restrict document, CDTN/CNEN (NIEC3-01/03), Belo Horizonte, Brazil. Dalle, H.M., Pereira, C., Souza, R.M.G.P., 2002. Neutronic calculation to the TRIGA IPR-R1 reactor using the WIMSD4 and CITATION codes. Annals of Nuclear Energy 29, 901912. DAuria, F., Frogheri, M. and Giannoti, W., 1999. RELAP5/MOD3.2 Post test analysis and accuracy quantication of lobi test BL-44. International Agreement Report, NUREG/IA-0153.

DAuria, F., Galassi, G.M., 1998. Code validation and uncertainties in system thermalhydraulics. Progress in Nuclear Energy 33, 175216. General Atomics, 1959. TRIGA Mark I Reactor Description. Gulf Energy & Environmental Systems, San Diego, CA, 1959 (GA-952). IAEA, 2005. Safety of Research Reactors. IAEA Safety Standards Series, N NS-R-4, IAEA, Vienna, 2005. Khater, H.A., El-Di El-Morshedy, S., Ibrahim, M.M.A., 2007. Thermalhydraulic modeling of the onset of ow instability in MTR reactors. Annals of Nuclear Energy 34, 194200. Khedr, A., Adorni, M., D0 Auria, F., 2005. The Effect of code user and boundary conditions on RELAP calculations of MTR research reactor transient scenarios. Nuclear Technology & Radiation Protection 1, 1622. Marcum, W.R., Woods, B.G., Reese, S.R., 2010. Experimental and theoretical comparison of fuel temperature and bulk coolant characteristics in the oregon state TRIGA reactor during steady-state operation. Nuclear Engineering and Design 240, 151159. Mesquita, A.Z., Souza, R.M.G.P., 2010. On-line monitoring of the IPR-R1 TRIGA reactor neutronic parameters. Progress in Nuclear Energy 52, 292297. Mesquita, A.Z., Rezende, H.C. and Souza, R.M.G.P., 2009. Thermal power calibrations of the IPR-R1 TRIGA nuclear reactor. In: Proceedings of the 20th International Congress of Mechanical Engineering, COBEM 2009, Gramado, RS, Brazil, November 1520, pp. 110. Souza, R.M.G.P., 1999. Resultados Experimentais da Calibrao das Barras de Controle e do Excesso de Reatividade em Cinco Conguraes do Ncleo do Reator IPR-R1. Restrict document, NI-AT 4-001/99, CDTN/CNEN, Belo Horizonte, Brazil. US NRC, 2001. RELAP5/MOD3.3 Code Manuals. Idaho National Engineering Laboratory, NUREG/CR-5535. Veloso, M. A. F., 2004. Anlise Termouidodinmica de Reatores Nucleares de Pesquisa Refrigerados a gua em Regime de Conveco Natural. Doctor Thesis, Universidade Estadual de Campinas, Brazil.