A C Y A h • • • • a A R R A 1 r ( i a o a a h w h 0 Nuclear Engineering and Design 276 (2014) 216–227 Contents lists available at ScienceDirect Nuclear Engineering and Design jou rn al hom epage : www.elsev ier .com/ locate /nucengdes ssessment of ASSERT-PV for prediction of critical heat flux in ANDU bundles .F. Rao ∗, Z. Cheng, G.M. Waddington tomic Energy of Canada Limited, Chalk River, ON, Canada K0J 1J0 i g h l i g h t s Assessment of the new Canadian subchannel code ASSERT-PV 3.2 for CHF prediction. CANDU 28-, 37- and 43-element bundle CHF experiments. Prediction improvement of ASSERT-PV 3.2 over previous code versions. Sensitivity study of the effect of CHF model options. r t i c l e i n f o rticle history: eceived 1 May 2014 eceived in revised form 11 June 2014 ccepted 14 June 2014 a b s t r a c t Atomic Energy of Canada Limited (AECL) has developed the subchannel thermalhydraulics code ASSERT- PV for the Canadian nuclear industry. The recently released ASSERT-PV 3.2 provides enhanced models for improved predictions of flow distribution, critical heat flux (CHF), and post-dryout (PDO) heat transfer in horizontal CANDU fuel channels. This paper presents results of an assessment of the new code version against five full-scale CANDU bundle experiments conducted in 1990s and in 2009 by Stern Laborato- ries (SL), using 28-, 37- and 43-element (CANFLEX) bundles. A total of 15 CHF test series with varying pressure-tube creep and/or bearing-pad height were analyzed. The SL experiments encompassed the bundle geometries and range of flow conditions for the intended ASSERT-PV applications for CANDU reactors. Code predictions of channel dryout power and axial and radial CHF locations were compared against measurements from the SL CHF tests to quantify the code prediction accuracy. The prediction statistics using the recommended model set of ASSERT-PV 3.2 were compared to those from previous code versions. Furthermore, the sensitivity studies evaluated the contribution of each CHF model change or enhancement to the improvement in CHF prediction. Overall, the assessment demonstrated significant improvement in prediction of channel dryout power and axial and radial CHF locations in horizontal fuel channels containing CANDU bundles. . Introduction ASSERT-PV (advanced solution of subchannel equations in eactor thermalhydraulics, pressure–velocity solution procedure) Carver et al., 1990; Carlucci et al., 2004; Rao and Hammouda, 2003) s a computer code developed at AECL mainly for thermalhydraulic nalysis of CANDU reactor fuel bundles. The code is also capable f modeling other reactor fuel bundles, including PWR and BWR ssemblies in vertical or horizontal channels. As well, the code can ccommodate a range of fluids, including single-and two-phase eavy water, light water, various Freons, and an air–water mixture. ∗ Corresponding author. Tel.: +1 613 584 3311x46346. E-mail addresses:
[email protected] (Y.F. Rao),
[email protected] (Z. Cheng),
[email protected] (G.M. Waddington). ttp://dx.doi.org/10.1016/j.nucengdes.2014.06.004 029-5493/Crown Copyright © 2014 Published by Elsevier B.V. All rights reserved. Crown Copyright © 2014 Published by Elsevier B.V. All rights reserved. The main use of ASSERT-PV is to compute thermalhydraulic parameters in a horizontal CANDU fuel channel, including pres- sure drop, critical heat flux (CHF) location, dryout power, and post dryout (PDO) fuel sheath temperature, for steady-state or slow transient flows. With financial support from utilities through CANDU Owners Group, Atomic Energy of Canada Limited (AECL) has developed a new version of ASSERT-PV, v3.2 released in 2012, with significant improvements/enhancements in flow-distribution, CHF and PDO heat transfer models. A series of four papers are devoted to presenting major results of the code-development project. The history of the ASSERT-PV code and the development of ASSERT- PV 3.2 flow-distribution, CHF and PDO model sets are described in Rao et al. (2014, Part 1). Code and model assessment for prediction of subchannel flow distributions is presented in Nava-Dominguez et al. (2014, Part 2). The assessment for prediction of dryout power and CHF locations in CANDU bundles is reported in this paper dx.doi.org/10.1016/j.nucengdes.2014.06.004 http://www.sciencedirect.com/science/journal/00295493 http://www.elsevier.com/locate/nucengdes http://crossmark.crossref.org/dialog/?doi=10.1016/j.nucengdes.2014.06.004&domain=pdf mailto:
[email protected] mailto:
[email protected] mailto:
[email protected] dx.doi.org/10.1016/j.nucengdes.2014.06.004 Y.F. Rao et al. / Nuclear Engineering an Nomenclature AECL Atomic Energy of Canada Limited ASSERT-PV advanced solution of subchannel equations in reactor thermalhydraulics, pressure–velocity solu- tion procedure AVG mean difference BLA boiling length average CHF critical heat flux CANFLEX CANDU FLEXable fuel (43-element bundle) CANDU CANadian Deuterium Uranium LUT look-up table OPG Ontario power generation PT pressure tube PDO post-dryout RMS root-mean-square SL Stern Laboratories STD standard deviation Subscripts CW cold wall effect enh enhancement size subchannel size (diameter) ( i p t m t v o o i c o o C i m t m o b a e d d t m o N p c p A m ( A C x quality Part 3); the assessment for prediction of PDO sheath temperatures s presented in a subsequent paper (Cheng et al., 2014, Part 4). The ASSERT-PV 3.2 development focused on improving code rediction of (i) flow distribution, (ii) dryout power and CHF loca- ion, and (iii) PDO sheath temperature distribution. The significant odel changes or additions in ASSERT-PV 3.2 (v3.2), compared o the previous version ASSERT-PV V3R1 (the format of the code ersion identifier has changed during the course of new code devel- pment) are described in Rao et al. (2014). For the improvement f the CHF model set, the development focused on incremental mprovements to the existing code implementation rather than onsidering completely new models. The effort was focused mainly n implementing a new CHF look-up table (LUT) and on the devel- pment and improvement of CHF correction factor models for the HF table look-up method. The CHF table look-up method, which s subchannel local condition based, has been the recommended odel option in previous code versions due to its wider applica- ion range compared to the other CHF correlations or empirical odels available in the code. ASSERT-PV 3.2 continues to use this ption in developing an improved model set, with additions of the oiling length average (BLA) model, the cold-wall effect model, nd improved models for the CHF quality/gap effects and the CHF nhancement effect (by bundle appendages) (Rao et al., 2014). Assessments of the improved capabilities for flow distribution, ryout power and CHF location, and PDO sheath temperature pre- ictions have been completed using experiment data sets including he Stern Laboratories’ 28-, 37- and 43-element bundle experi- ents. Significant improvement has been confirmed for all key utput parameters and over all the three CANDU bundles. Refs. ava-Dominguez et al. (2014) and Cheng et al. (2014), respectively, resent the assessment results for the flow-distribution and PDO alculations, whereas this paper focuses on the assessment of CHF redictions. Since this paper is Part 3 of the four-paper series on SSERT-PV 3.2, readers who are interested in the details of the code odels are advised to read this paper together with Ref. Rao et al. 2014, Part 1). Subsequent sections in the paper are arranged as follows: (i) SSERT-PV code assessment guidelines and methodology; (ii) SL HF experiments and ASSERT idealization; (iii) ASSERT-PV results d Design 276 (2014) 216–227 217 and accuracy; (iv) sensitivity study of each model change; and (v) conclusions. 2. Code assessment guidelines and methodology 2.1. Code assessment guidelines The following principles are followed in the current ASSERT code assessment: Data qualification should be examined for the selected datasets, test series, and each selected individual test for code assessment. The following data qualifications should be satisfied: • the experimental parameters were measured and recorded at suf- ficient accuracy and sampling frequency, and the measurement uncertainties are known or can be estimated, • test boundary and initial conditions to be applied to the ASSERT input model were recorded at sufficient detail and accuracy to allow simulation of the tests, • test facility configuration and geometry are described with suffi- cient detail and accuracy for the ASSERT input model, • test results are reasonable and self-consistent with all anomalies been explained, problems identified and reported issues resolved, and • experimental data used in the current code assessment are independent of those for developing correlations or models in ASSERT-PV 3.2 and in the previous code versions. The selected experimental data should be close to the ASSERT intended application range. This is to ensure that the ASSERT code is assessed for the qualification of safety analyses under normal operating conditions and various accidental scenarios. The ASSERT intended application range for light water is: • channel outlet pressure: 6–11 MPa • channel inlet flow rate: 7–25 kg/s • channel power: up to 15% overpower beyond the onset of CHF • channel outlet qualities: up to 100% • channel inlet temperature: 230–270 ◦C • pressure tube diametric creep: 0–5.1%. One recommended model set is to be applied to cover all CANDU bundle geometries and operating conditions; i.e., user input would remain the same except for the bundle geometry information. This model set for the code assessment is based on the knowledge and experience accumulated in recent ASSERT code development. No “tuning” of model constants or coefficients is allowed during the assessment so as to increase the degree of confidence and credibil- ity in the assessment (see Section 3.2). This is also to ensure that the ASSERT-PV 3.2 code can be used for the new bundle design in the future. 2.2. Code assessment methodology The ASSERT code accuracy is determined by the quantification of the closeness of the code prediction to the measured value. The prediction error, or residual, is the difference between the predic- tion and the corresponding experimental (measured) value. The mean value of a set of residuals over a range of a key parameter is calculated to provide an estimate of the code bias applicable over that range. The uncertainty or variability of the code bias is deter- mined statistically based upon observation of the distribution of the residuals about the mean. The standard deviation of the resid- uals is used to express this variation in the code bias. The mean (AVG) and standard deviation (STD) can be combined into a single 2 ring a s i q c L o t b m m A 2 c m t c V C 3 f 3 u ( a 2 o t s p u w n e p n u n ( c b c i r m t i t c # c d ( p e N 18 Y.F. Rao et al. / Nuclear Enginee tatistic that reflects both the average difference and the variation n the differences: the root-mean-square (RMS) of the residuals. These statistics (AVG, STD, RMS) are calculated and used to uantify the ASSERT code accuracy for the CHF predictions. They are alculated for a subset of the tests in each exercise (e.g., for a Stern ab test series representing a single pressure-tube profile), but the verall bias, standard deviation and RMS of the differences between he predicted and measured dryout power are also reported, and ased on as many individual tests and measurements as possible. The contributions of the newly developed (or modified) CHF odels to the CHF prediction improvement are assessed by the odel sensitivity studies. Table 1 shows the comparison of the SSERT-PV 3.2 CHF model set and the V3R1 model set (Rao et al., 014). The separate effect of each model change is quantified by omparing the prediction result obtained by using all the recom- ended v3.2 CHF model changes with that obtained by excluding hat particular model change. The overall effect of the CHF model hanges is assessed too by excluding all CHF model changes to 3R1, resulting in an ASSERT-PV model set that combines the V3R1 HF model set with the v3.2 flow-distribution model set. . Experiment data and ASSERT idealization The assessment of the recommended CHF model set uses the ull-scale SL CHF experiments. .1. Stern Lab bundle tests Several experimental programs were conducted at SL to eval- ate CHF and thermalhydraulic behaviors of the 28-element Fortman, 2010), 37-element (Fortman et al., 1997; Fortman, 2011), nd CANFLEX (43-element) (Dimmick et al., 1999; Leung et al., 001) fuel designs. Single- and two-phase tests were performed ver a range of thermalhydraulic conditions using one uniform and wo non-uniform internal diameter flow channel profiles, which imulate, respectively, an uncrept and two crept pressure tube (PT) rofiles of maximum diametral creep of 3.3% and 5.1%. Each sim- lated fuel string was comprised of electrically heated elements ith simulated end plates and appendages to represent the exter- al geometry of a fully aligned string of twelve fuel bundles. The xperiments were set up to obtain the onset-of-dryout power, ressure drop, and PDO sheath temperature for fuel bundles with on-uniform power profiles in uncrept and crept flow channels nder conditions relevant to CANDU reactors. Figs. 1–3 show the bundle images (left) and ASSERT subchan- el modeling (right) for 28-element, 37-element and CANFLEX 43-element) bundles, respectively. The bundle images show the ross-views of the bundles with appendages (end-plates, spacers, uttons and bearing pads). Each fuel element ring in the bundles an be classified as an inner, intermediate or outer ring, as shown n the left figure of Fig. 2. For the 28-element bundle, the inner ing contains elements #1–4, the intermediate ring contains ele- ents #5–12 and the outer ring contains elements #13–28. For he 37-element bundle, the inner ring contains elements #2–7, the ntermediate ring contains elements #8–19 and the outer ring con- ains elements #20–37. For the CANFLEX bundle, the inner ring ontains elements #2–8, the intermediate ring contains elements 9–22 and the outer ring contains elements #23–43. The full-scale 28-element bundle simulator was designed and onstructed at SL to simulate twelve aligned Pickering fuel bun- les (Fig. 1). The outside diameter of each rod was 15.27 mm cold cold measurements were made at a temperature of 20 ◦C and a ressure of 103 kPa, with no power or flow). The full-scale 37- lement bundle was similar to the Bruce and Darlington CANDU uclear Generating Station fuel bundles (Fig. 2), each having two nd Design 276 (2014) 216–227 endplates, five planes of bearing pads and one plane of spacer pads. The outside diameter of each rod was 13.06 mm cold. The full-scale 43-element (CANFLEX) bundle simulator was designed and constructed to simulate an Advanced CANDU Reactor (Fig. 3). The length of each simulated fuel bundle was 495.3 mm cold. The overall bundle outer diameter (OD) was 102.5 mm (cold). Two ele- ment diameters were utilized in the CANFLEX design. The center and seven inner ring elements had a nominal diameter of 13.5 mm (cold) whereas the 14 intermediate and 21 outer ring elements had a nominal diameter of 11.5 mm (cold). The experiments covered bearing pad heights of 1.4 (Phase 1), 1.7 and 1.8 mm (Phase 2). In SL tests, the experimental fuel string simulated 12 identi- cal, fully aligned bundles (Bundles A–L as shown in Fig. 4), each of which is 495.3 mm long. The experimental fuel string was elec- trically heated with a nominal heated length of 5.76 m The wall thickness of the elements varied axially to produce a non-uniform, symmetrical axial heat flux distribution (AFD) for 28-element bun- dles, and a downstream-skewed cosine-shaped AFD for 37-element and CANFLEX bundles, as shown in Fig. 5. The radial heat flux distribution (RFD) increased from the cen- ter to the outside of the bundle. The specified inner to outer ring linear to average power ratios were based on design calculations to simulate reactor fuel flux depression for fresh natural uranium fuel. For the RFD of 28-element bundle (Fortman, 2010), the spec- ified ring to average power ratios were 0.78/0.902/1.104 from the inner to the outer ring of elements. The fuel string had a maxi- mum operating power level of 9.5 MW with a design pressure of 11 MPa and a design maximum local sheath temperature of 650 ◦C. For 37-element bundles (Fortman et al., 1997; Fortman, 2011), the ratios of ring to average heat flux were 0.828/0.860/0.932/1.101 from the center element to the outer ring of elements. The design parameters of the fuel strings were maximum operating power of 12.5 MW at 220 V DC with a design pressure of 13.5 MPa and a maximum local sheath surface temperature of 650 ◦C. For CAN- FLEX bundles (Dimmick et al., 1999; Leung et al., 2001), the RFD ratios of ring to average heat flux, were 0.910/0.951/0.902/1.090. Hot dimensions were calculated assuming an inlet coolant tem- perature of 268 ◦C, exit pressure of 11 MPa, flow of 18 kg/s and bundle power of 9.0 MW. The design parameters of the fuel strings were maximum operating power of 13.5 MW at 240 V DC, a maxi- mum operating pressure of 11.5 MPa, and a maximum local sheath surface temperature of 600 ◦C. SL also simulated pressure tube (PT) aging effect during nor- mal reactor operation. Fig. 6 shows the two crept PT profiles of maximum diametric creep of 3.3% and 5.1%. Table 2 lists the SL CHF tests used in the assessment of the ASSERT-PV 3.2 CHF model set, which is recommended to be used with the new v3.2 flow distribution models (Nava-Dominguez et al., 2014). The test conditions are consistent with the ASSERT intended application range. 3.2. ASSERT idealization The ASSERT model for the SL bundle channel is set up based on the test section dimensions and the test conditions from the experiments described in Section 3.1. 3.2.1. Subchannel geometry The SL test section geometry is modeled in ASSERT-PV by a number of discrete control volumes that, taken together, cover the whole channel. The channel is divided into 41 subchannels for 28-element bundle, 60 subchannels for 37-element bundle, and 70 subchannels for CANFLEX bundle. These subchannels are defined as the coolant flow area bounded by the rod surfaces and imaginary lines joining adjacent rod centers. The right pictures in Figs. 1–3 show the subchannel discretization of the SL 28-element, Y.F. Rao et al. / Nuclear Engineering and Design 276 (2014) 216–227 219 Table 1 ASSERT-PV v3.2 CHF model set compared to V3R1. Mod V3R1 Model 3.2 Comments Flow distribution models Previous models (V3R1) v3.2 flow distribution model (Rao et al., 2014; Nava-Dominguez et al., 2014) The recommended v3.2 flow distribution model set is superior to V3R1 model set (Nava-Dominguez et al., 2014). CHF look-up table 1995 2005 LUT 2005 (Groeneveld et al., 2005) reduces prediction STD when the “right” flow and CHF correction models are used (Rao et al., 2014). BLA model No BLA model with lower bound KBLA = qlocal/qBLA, KBLA lower-bounded by 0.8 in expectation of reduced BLA effect in an “open” subchannel with bundle appendages (Rao et al., 2014). Subchannel size (diameter) correction exponent coefficient n 1/3, but 0 in effect (Rao et al., 2014) 1/2 Ksize = (0.008/Dh)n . Recent CHF look-up tables are associated with n = 1/2 (Rao et al., 2014; Groeneveld et al., 2005). Cold-wall effect and relaxation factor �CW (0.0–1.5) N/A KCW with default, �CW = 1.0 KCW = 1 − �CW(1 − �1/2), where � = Dh/Dheated. The cold-wall model is similar to others in literature in using Dheated/Dh to correlate the effect. �CW: relaxation factor (Rao et al., 2014). CHF enhancement by bundle appendages Existing CHF enh. model (Rao et al., 2014) Modified CHF enh. model (Rao et al., 2014) Based on a previous R&D work. The model reduces CHF enhancement effect (Doerffer et al., 2000) by considering an “enlarged” subchannel area including regions of gaps connecting neighboring subchannels (Rao et al., 2014). Model constant �enh is set to 1.0 (default). Quality effect, relaxation factor: �x 0.25 0.5 Kx = 1 − �x(1 − K∗x ), K∗x is the CHF penalty ratio obtained based on experiment data. Justification of an under-relaxation of the effect was based on the consideration that data 3 e 3 e o ( 7-element and CANFLEX bundles, respectively, together with the lement and subchannel numbering used by ASSERT-PV. .2.2. Axial nodalization The length of the channel simulated by ASSERT is 6.439 m for ach bundle test. This length is equivalent to the 13-bundle length f the pressure drop measurements described in Refs. Fortman 2010, 2011) and Dimmick et al. (1999). Note that the simulated Fig. 1. Image of a 28-element bundle and ASS used in the original model (Groeneveld et al., 1992) included the cold-wall effect (Rao et al., 2014). 13-bundle length consists of the 12 heated bundles and approxi- mately half an unheated bundle at each end. The total bundle length (6.439 m) is divided into 156 (12 per bundle) non-uniformly sized axial zones using a user specified axial zone table. 3.2.3. Axial pressure-tube creep profiles The axial variation in pressure-tube diameter is specified in the ASSERT input by a table of axial variation factors. This affects the ERT subchannel modeling (uncrept PT). 220 Y.F. Rao et al. / Nuclear Engineering and Design 276 (2014) 216–227 Fig. 2. Image of a 37-element bundle and ASSERT subchannel modeling (5.1% (max) crept PT). Fig. 3. Image of a CANFLEX (43-element) bundle and ASSERT subchannel modeling (uncrept PT). ion in s s d t g a i t 3 i c r t s Fig. 4. Test sect ubchannel flow areas, and heated and wetted perimeters, inter- ubchannel gap widths, and the subchannel centroid-to-centroid istances and associated angles for the outer subchannels next to he flow tube. These subchannel geometries are computed by the eometry pre-processor using the relative pressure tube diameter t each relative axial location (The relative pressure tube diameter s defined as the local crept diameter over the nominal pressure ube diameter as shown in Fig. 6). .2.4. Boundary conditions The ASSERT-PV input file provides the information for the exper- mental axial and radial heat flux distributions (AFD and RFD) to ompute the heat flux over each axial zone and each element ing. Other boundary conditions, such as the channel fluid inlet emperature, mass flux (inlet mass flow rate), channel outlet pres- ure and applied channel average heat flux (input power) are Stern Lab tests. determined from the reported measurements, and specified for each test case through a series of stacked input records. The enthalpy (temperature) distribution at the inlet is assumed uniform across all subchannels. The mass flow rate at the inlet is split by the ASSERT code to equalize the axial pressure drop in each subchannel at the inlet region. 3.2.5. Fluid properties Light water was used in the SL tests. The ASSERT HLWP package is used to determine the saturated and unsaturated thermodynamic properties of light water and vapour. 3.2.6. Flow distribution modeling The CHF calculations are inevitably involved with flow dis- tribution calculations. The friction factors, multipliers and heat transfer correlations are specified in the input file with the Y.F. Rao et al. / Nuclear Engineering and Design 276 (2014) 216–227 221 Fig. 5. Axial heat flux distribution (AFD). profile r 2 C h T S Fig. 6. Flow channel ecommended ASSERT flow distribution model set (Rao et al., 014; Nava-Dominguez et al., 2014). This CHF assessment uses the olebrook–White single-phase turbulent friction factor (with no eated wall viscosity correction factor) and the Friedel correlation able 2 ummary of experiment datasets. Bundle PT creep or bearing pad Test series 28-E Uncrept (with/without 90◦ bundle rotation) R1 – 96 cases R2 – 42 cases (90◦ rotation) 3.3% creep (90◦ bundle rotation) C1 – 124 cases 37-E 1990s Uncrept R2 – 40 cases 3.3% creep C1 – 48 cases 37-E 2009 3.3% creep C4 – 200 cases 5.1% creep C3 – 163 cases CANFLEX (43-E) Phase 1 Uncrept, BP 1.4 mm R1 – 48 cases 3.3% creep BP 1.4 mm C2 – 63 cases 5.1% creep BP 1.4 mm C1 – 70 cases CANFLEX (43-E) Phase 2 Uncrept, BP 1.8 mm B1 – 84 cases 5.1% creep, BP 1.8 mm B2 – 49 cases 5.1% creep, BP 1.7 mm B3 – 45 cases 3.3% creep, BP 1.7 mm B4 – 50 cases Uncrept, BP 1.7 mm B5 – 75 cases s for Stern Lab tests. for the two-phase friction multiplier. The homogeneous two-phase multiplier is used for the form loss calculations. The wall-to-fluid heat transfer model is selected such that the Chen correlation is applied for void fractions up to 20% (the Chen correlation is used Operating conditions Pressure (MPa) Inlet temperature (◦C) Channel flow rate (kg/s) 5–11 225–275 7–24 6–11 225–280 7–23 6–11 200–290 10–29 222 Y.F. Rao et al. / Nuclear Engineering and Design 276 (2014) 216–227 .2 and f fi e t i m t e m f D 3 a t • • • • • Fig. 7. Comparison of ASSERT-PV 3 or the liquid convective and nucleate boiling heat-transfer coef- cient, and the Ahmad correlation for higher void fractions, Rao t al., 2014. For single-phase flow, the Chen correlation reverts to he Dittus–Boelter correlation). The subchannel mixing models and parameters are the mod- fied Carlucci thermal and momentum mixing and void diffusion ixing model, Rowe’s equilibrium void model, Rehme’s correla- ion for thermal and momentum mixing of subchannel pairs (Rao t al., 2014; Nava-Dominguez et al., 2014). Default incremental void ixing multiplier and thermal and momentum mixing obstruction actors are used in the code assessment (Rao et al., 2014; Nava- ominguez et al., 2014). .2.7. CHF modeling The v3.2 recommended CHF model set used in the current CHF ssessment is briefly described below, with details being referred o Rao et al. (2014). CHF LUT 2005 is used instead of the previous version CHF LUT 1995. The new LUT resolved a prediction accuracy issue of the old LUT in the so-called “Limiting Quality Region (LQR)” (Groeneveld et al., 2005) characterized by a fast decrease of CHF with an increase in steam quality. The new LUT has an enhanced quality of database leading to better predictions of CHF. The subchannel-size correction model is activated and the expo- nent coefficient is set to 1/2 for consistency with the use of CHF LUT 2005. This correction factor often has significant effect on predicted dryout power and CHF location since there are many subchannels with hydraulic diameters significantly different from 8 mm, which is used for generating the CHF look-up-table 2005. The modification to the CHF enhancement correction factor reduces the CHF enhancement effect by considering an “enlarged” subchannel area (for calculation of this factor only) that includes regions of gaps connecting neighboring subchannels. The BLA correction model is chosen as the recommended option against the local-condition CHF approach, and a lower bound of 0.8 (or 80%) is imposed to limit its effect because CANDU bundle appendages are expected to make upstream flow history less rel- evant compared to a tube geometry without these appendages, again because of the presence of adjacent subchannels. The new cold-wall effect correction factor in ASSERT-PV 3.2 accounts for the CHF penalty effect due to existence of a cold (or unheated or adiabatic) wall or rod segment that causes increased imbalance in flow and enthalpy distribution within a subchannel. In most CANDU applications, the relevant V3R1 for dryout power prediction. subchannels are the outer subchannels adjacent to the unheated pressure tube. • The quality/gap effect (gap-correction) model is revised with the relaxation coefficient changed from a previous value of 0.25 (75% discount of the effect compared to that in a heated annular chan- nel) to 0.5 (50% discount). 3.2.8. Calculation control The calculation control options are specified in the input file. The maximum numbers of energy (enthalpy) iterations, inner (mass) iterations, outer (flow) iterations, are set conservatively to 20, 20 and 150, respectively. The relative residual change criteria for enthalpy, flow, mass equation, lateral and axial momentum are set to 10−6, 10−4, 10−6, 10−3 and 10−3, respectively. Note that the test cases are executed using the stacked runs and these numbers of iterations ensures good convergence for all the test cases stacked in a single input file (for a certain test series). 4. ASSERT-PV results and accuracy The current assessment is made by the comparison of the pre- diction accuracy of dryout power, using the prediction statistics including the mean difference (AVG), the standard deviation of the differences (STD), and the root-mean-square difference (RMS) in dryout power. Improvement of CHF location prediction is one of the model development targets and the assessment also includes the comparison of the prediction accuracy of CHF axial and radial locations. In comparing radial CHF location, emphasis is placed on accurately predicting both the element ring and the subchannel ring, i.e., CHF at the correct element ring facing the correct side (facing inside toward bundle center or facing outside toward the PT, rather than on individual elements or subchannels. Table 3 lists the statistics of assessment of the ASSERT-PV 3.2 model set (herein referred to as Model 3.2) against the ASSERT V3R1 model set (V3R1) for the predication of dryout power. The related v3.2 code bias is −0.4% (AVG) for 28-element bundles, −2.5% for 37-element 1990s tests, −3.1% for 37-element 2009 tests, −2.3% for CANFLEX Phase 1 tests and −0.7% for CANFLEX Phase 2 tests. The overall code bias is −1.1% for all three bundle types. The assessment confirms that the v3.2 code bias is significantly smaller than the V3R1 code bias. The RMS differences obtained from Model 3.2 are also smaller than those from the ASSERT V3R1 model set, indicating that Model 3.2 is superior to the previous model set for the dryout power predictions. The most significant improvement in dryout power can be iden- tified for the 28-element and CANFLEX bundle tests, as shown in Y.F. Rao et al. / Nuclear Engineering and Design 276 (2014) 216–227 223 Table 3 Prediction statistics: dryout power. Bundle PT creep (%) Statistics V3R1 (%) Model 3.2 (%) 28e (262 cases) 0,3.3 AVG 10.6 −0.4 STD 5.8 3.7 RMS 12.1 3.7 37e-1990s (88 cases) AVG −5.3 −2.5 0,3.3 STD 4.5 5.7 RMS 7.0 6.2 37e-2009 (363 cases) 3.3,5.1 AVG −6.9 −3.1 STD 6.7 8.7 RMS 9.6 9.2 43e-Phase 1 (181 cases) 0,3.3,5.1 AVG 9.1 2.3 STD 5.2 4.2 RMS 10.5 4.8 43e-Phase 2 (303 cases) 0,3.3,5.1 AVG 6.6 −0.7 STD 6.4 4.5 RMS 9.2 4.5 All (1197 cases) AVG 2.9 −1.1 0,3.3,5.1 STD 9.7 6.3 F o a a p m p p U f C C t b P e a M e ig. 7. In the 28-element and CANFLEX bundle tests, most CHF ccurred on the outer ring, where the heat transfer is significantly ffected by the cold wall of the pressure tube. The newly developed nd modified CHF models significantly reduce the dryout power rediction errors for these tests. The current assessment shows that the V3R1 and v3.2 CHF odel sets predict the axial location of CHF similarly well, but v3.2 redicts the radial location significantly better. Figs. 8–12 show the rediction statistics of CHF radial locations for each bundle test. nlike the CHF axial location predictions, the V3R1 CHF model set ails in the CHF radial location predictions for the 28-element and ANFLEX bundle tests. However, Model 3.2 accurately predicts the HF radial locations for most test cases regardless of their bundle ypes. For 28-element tests, Model 3.2 accurately predicts the num- er of cases of CHF at outer-ring elements facing outside (toward T), which accounts for 87% of the test points (Fig. 8). For 37- lement tests, Model 3.2 accurately predicts most cases of CHF at n inner-ring element (Figs. 9 and 10), whereas for CANFLEX tests, odel 3.2 accurately predicts most cases of CHF at an outer-ring lement (Figs. 11 and 12). Model 3.2 is assessed to be superior to Fig. 8. Prediction of CHF radial locations for 28-E tests. RMS 10.1 6.4 the V3R1 model set in the prediction of CHF radial locations. This capability is important in applications for bundle geometry opti- mization or for bearing pad height design. 5. Model sensitivity studies Sensitivity studies are performed to quantify the contribution of each CHF model change and the overall effect of the CHF model changes to the improvement in dryout power and CHF location predictions. Following the code assessment methodology (Section 2), the effect of each model change is assessed using the recom- mended v3.2 CHF model set but excluding that particular model change. Similarly, the overall effect of the CHF model changes can be assessed by excluding all CHF model changes to V3R1, resulting in an ASSERT-PV model set that combines the V3R1 CHF model set with the v3.2 flow-distribution model set. 5.1. Dryout power prediction Table 4 lists prediction statistics of dryout power for each sensitivity-study case, designed to assess the effect of a par- ticular model change. The expectation is that each and every Fig. 9. Prediction of radial CHF locations for 37-E 1990s tests. 224 Y.F. Rao et al. / Nuclear Engineering and Design 276 (2014) 216–227 Table 4 Sensitivity study: summary of prediction statistics of dryout power. Bundle Statistics Model 3.2 (%) LUT 1995 (%) No BLA model (%) V3R1 CHF enhancement (%) No sub-channel size correction (%) No cold-wall effect �CW = 0.0 (%) V3R1 quality effect �x = 0.25 (%) All CHF model changes, not used (%) 28-e AVG −0.4 0.0 4.2 −0.1 1.9 5.0 4.8 11.7 STD 3.7 4.0 3.6 3.7 4.1 3.6 3.8 4.8 RMS 3.7 4.0 5.6 3.7 4.5 6.2 6.1 12.6 37-e 1990s AVG −2.5 −2.7 0.9 −2.2 −5.0 −2.5 1.2 2.7 STD 5.7 6.3 6.1 5.6 6.8 5.7 6.2 7.3 RMS 6.2 6.8 6.2 6.0 8.5 6.2 6.3 7.8 37-e 2009 AVG −3.1 −3.0 0.1 −2.9 −4.5 −3.1 0.5 2.8 STD 8.7 9.1 9.6 8.7 9.9 8.7 9.8 10.7 RMS 9.2 9.6 9.6 9.2 10.9 9.3 9.8 11.0 43-e Phase 1 AVG 2.3 1.9 6.5 4.5 2.8 2.3 6.9 12.4 STD 4.2 4.3 4.2 4.3 4.1 4.2 4.0 4.4 RMS 4.8 4.7 7.8 6.3 5.0 4.8 8.0 13.2 43-e Phase 2 AVG −0.7 −1.1 3.8 2.4 0.4 −0.7 4.6 10.6 STD 4.5 5.0 4.3 4.1 4.5 4.5 4.1 5.1 RMS 4.5 5.1 5.7 4.8 4.6 4.5 6.2 11.7 Overall AVG −1.1 −1.1 3.0 0.2 −0.8 0.1 3.5 8.2 6.6 6.6 s p p t m o m o o t s a T q b b m t f STD 6.3 6.6 6.8 RMS 6.4 6.7 7.4 ensitivity case should show an increase in RMS difference com- ared to the base (v3.2) case, confirming a positive effect on CHF rediction accuracy. For 28-element bundle tests, dryout power prediction is found o be improved significantly by the modified cold-wall effect odel; a RMS reduction of 2.5% in dryout power, from the 6.2% btained by excluding the cold-wall effect to the 3.7% by the recom- ended v3.2 model set, as shown in Table 4. The combined effect f all CHF model changes is a RMS reduction of 8.9%, from the 12.6% btained by no CHF model changes (i.e., the V3R1 CHF model set) o the 3.7% by the v3.2 CHF model set. The sensitivity study also hows that the modified BLA model and quality/gap effect model re superior to their V3R1 counterparts in dryout power prediction. he RMS is 4.3% using the V3R1 BLA model, and 6.1% using the V3R1 uality effect correction model, respectively, as compared to 3.7% y the v3.2 CHF model set. For 37-element tests, the prediction is significantly improved y the subchannel-size correction model, but the effects of other odel changes are insignificant. For the 37-elemnet 1990s tests, he combined effect of all model changes is a RMS reduction of 1.6%, rom the 7.8% obtained by no CHF model changes to the 6.2% by the Fig. 10. Prediction of CHF radial locations for 37-E 2009 tests. 7.3 6.8 6.9 8.4 7.4 6.8 7.7 11.7 v3.2 CHF model set. For the 37-elemnet 2009 tests, the combined effect of all CHF model changes is a RMS reduction of 2.8%, from the 11.0% by no CHF model changes to the 9.2% by the v3.2 CHF model set. For CANFLEX tests, all the CHF model changes improve the pre- diction of dryout power except for the cold-wall effect model. The dryout power prediction is significantly improved by the BLA model (by 3.0% in RMS) and the gap/quality effect model (by 3.2%) for the CANFLEX Phase 1 tests. Similar effects of the two model changes can be observed also for the Phase 2 tests, but with a smaller degree of improvement. The combined effect of all CHF model changes is a reduction in RMS of 9.6%, from the 13.2% by no CHF model changes to the 4.8% by the v3.2 CHF model set for the Phase 1 tests, and a RMS reduction of 7.2%, from 11.7% to 4.5%, for the Phase 2 tests. 5.2. CHF axial location prediction Due to the fact that the V3R1 model is also capable of predict- ing the CHF axial location, the statistics of sensitivity of predicted CHF axial location to each model change is not tabulated in this paper, but the sensitivity results are still briefly reported here. Fig. 11. Prediction of CHF radial locations for 43-E Phase 1 tests. Y.F. R ao et al. / N uclear Engineering and D esign 276 (2014) 216–227 225 Table 5 Sensitivity study: summary of prediction statistics of CHF radial location. Bundle Rod ring Inside/ outside Stern Model 3.2 LUT 1995 No BLA Existing CHF enh. model No subch. size correc. model No cold-wall effect, �CW = 0 Existing quality effect, �x = 0.25 Existing BLA No CHF enhancement Subch. size corr. expon. (n = 1/3) Cold-wall effect, �CW = 1.5 No quality effect, �x = 0.0 28-e Inner Inside Outside Interm. Inside 6 15 18 17 17 17 20 15 18 12 17 16 16 Outside 6 Outer Inside 24 23 33 34 27 186 199 37 24 64 31 19 36 Outside 226 224 211 211 218 59 43 210 220 186 214 227 210 37-e 1990s Inner Inside 54 65 65 65 65 69 65 65 66 12 66 64 65 Outside 7 Interm. Inside 22 11 14 14 11 7 11 16 14 49 10 11 21 Outside Outer Inside 5 12 9 9 12 12 12 7 8 27 12 12 2 Outside 1 37-e 2009 Inner Inside 138 191 186 186 192 206 191 186 198 11 193 191 186 Outside 114 Interm. Inside 100 139 155 161 140 128 139 162 157 250 138 139 169 Outside 4 Outer Inside 5 33 22 16 31 29 33 15 8 102 32 33 8 Outside 2 43-e Phase 1 Inner Inside 4 3 5 5 5 4 8 7 1 4 4 12 Outside 1 7 12 12 12 33 10 8 4 15 6 11 Interm. Inside Outside Outer Inside 109 167 158 164 135 143 167 162 152 178 162 125 158 Outside 71 3 8 29 3 18 2 46 43-e Phase 2 Inner Inside 3 2 2 3 5 Outside 33 35 42 57 101 33 41 33 57 33 49 Interm. Inside Outside Outer Inside 267 270 265 259 246 200 270 262 267 303 246 269 249 Outside 36 1 226 Y.F. Rao et al. / Nuclear Engineering a T L q t c i p t a h p 5 o p m e i c m b “ ( n p t s T o 5 r n o i l r w o Fig. 12. Prediction of CHF radial locations for 43-E Phase 2 tests. he CHF axial location prediction is slightly improved by the CHF UT 2005, the modified CHF enhancement model and the modified uality effect model for 28-element bundle tests. For 37-element ests, the CHF axial location prediction is improved by all the model hanges except for the cold-wall effect model, which has no effect f no CHF is predicted at the outer-ring facing outside toward the ressure tube. For CANFLEX tests, the CHF axial location predic- ion is significantly improved by the modified quality effect model nd BLA model. Similar to 37-element tests, the cold-wall model as no effect on the improvement of CANFLEX CHF axial location rediction. .3. CHF radial location prediction Significant improvement in radial CHF location prediction is one f the major features of the v3.2 model set. Table 5 summarizes the redicted radial CHF locations compared to the Stern Lab experi- ents, in the form of the number of cases of CHF initiation at certain lement rings and facing certain direction (inside or outside). For 28-element bundle tests, the radial location prediction is mproved by all the CHF model changes, but the subchannel-size orrection model and the cold-wall effect model contribute the ost improvement. Without these two model changes, CHF would e predicted at the outer-ring facing “inside”, rather than facing outside” as measured in the experiment, for most of the cases see Table 5). This confirms that neither the subchannel-size effect or the pressure-tube cold-wall effect can be ignored in accurate rediction of CHF radial location. For 37-element tests, the CHF radial location prediction is rela- ively insensitive to most model changes. For CANFLEX bundle tests, the CHF radial location prediction is ignificantly improved by the subchannel-size correction model. he other model changes are found to have minor (mixed) effects n the prediction of CHF radial location. .4. Summary of sensitivity studies The CHF LUT 2005 was found to be superior to the 1995 LUT in educing the code bias and associated standard deviation, but the ew look-up table does not improve significantly the prediction f CHF location. The subchannel-size (diameter) correction model mproves significantly the prediction of dryout power and axial CHF ocation for 37-element bundles. It also improves the prediction of adial CHF location for 28-element and CANFLEX tests. The cold- all effect model was found to improve significantly the prediction f dryout power and CHF location for 28-element bundle tests, in nd Design 276 (2014) 216–227 which CHF occurs at the outer-element rings facing outside toward the PT for most of the cases, but it has no effect on the prediction of 37-element and CANFLEX bundles. The modified CHF enhance- ment model improves significantly the prediction of dryout power for CANFLEX Phase 1 tests, but it has no significant effect for 28- element and 37-element bundle tests when compared with the existing CHF enhancement model. The gap/quality effect model was found to improve significantly the predictions of dryout power for 28-element and CANFLEX tests, and of axial CHF location for CAN- FLEX tests. The BLA model improves significantly the prediction of dryout power for 28-element and CANFLEX tests. It also improves the axial CHF location prediction, except for the 28-element bundle tests. 6. Conclusions This paper presents an assessment of the subchannel code ASSERT-PV 3.2 for prediction of dryout power and CHF location in CANDU bundles. Code predictions were compared with measure- ments from Stern Laboratories’ full-scale 28-element, 37-element and CANFLEX bundle tests conducted in 1990s and in 2009, which encompass the bundle geometries and range of flow conditions for the intended ASSERT-PV applications for existing CANDU reactors. The prediction statistics using the recommended model set of ASSERT-PV 3.2 were compared with those from the previous code version V3R1. Overall, the assessment demonstrated significant improvement in prediction of channel dryout power and axial and radial CHF locations in horizontal fuel channels containing CANDU bundles. Separate-effects sensitivity studies were performed for each CHF model change or enhancement, including the new CHF look-up table and a number of newly developed or modified CHF correction models accounting for the effects of various subchannel geometries and flow conditions that are different from those in a vertical, uni- formly heated tube of 8 mm inner diameter. It was demonstrated that overall the CHF model changes reduce the ASSERT code bias, associated standard deviation and RMS difference in prediction of dryout power, and improve significantly the prediction of CHF radial location as well. The ASSERT-PV 3.2 CHF model set is expected to do well for other applications involving small change in bundle geometries or flow conditions, such as applications for OPG’s modified 37- element bundles where the bundle geometry change is small and the test conditions are within the ASSERT intended application range. However, it cannot be expected to do similarly well for other applications with flow conditions far beyond the ASSERT applica- tion ranges, such as under significantly lower pressures or mass fluxes. References Carver, M.B., Tahir, A., Kiteley, J.C., Banas, A.O., Rowe, D.S., Midvidy, W.I., 1990. Sim- ulation of flow and phase distribution in vertical and horizontal bundles using the ASSERT-4 subchannel code. Nucl. Eng. Des. 122, 413–424. Carlucci, L.N., Hammouda, N., Rowe, D.S., 2004. Two-phase turbulent mixing and buoyancy drift in rod bundles. Nucl. Eng. Des. 227, 65–84. Cheng, Z., Rao, Y.F., Waddington, G.M., 2014. Assessment of subchannel code ASSERT-PV for prediction of post-dryout heat transfer in CANDU bundles. In: International Congress on Advances in Nuclear Power Plant, Charlotte, NC, April 6–9, Paper 14351 (an extended version has been submitted to Nucl. Eng. Des.). Dimmick, G.R., Inch, W.W.R., Jun, J.S., Suk, H.C., Hadaller, G.I., Fortman, R.A., Hayes, R.C., 1999. Full scale water CHF testing of the CANFLEX bundle. In: 6th Interna- tional Conference on CANDU Fuel, Canadian Nuclear Society, September. Doerffer, S., Groeneveld, D.C., Rudzinski, K.F., Martin, J.W., 2000. Some aspects of critical-heat-flux enhancement in tubes. In: Paper Abstract for ASME IMECE Paper No. 2-13-5-4. Fortman, R.A., 2010. PDO and CHF tests on CANDU 28-element fuel. In: Stern Labo- ratories Report (Internal Report) SL-195. http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0005 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0005 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0005 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0005 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0005 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0005 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0005 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0005 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0005 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0005 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0005 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0005 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0005 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0005 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0005 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0005 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0005 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0005 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0005 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0005 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0005 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0005 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0005 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0005 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0010 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0010 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0010 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0010 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0010 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0010 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0010 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0010 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0010 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0010 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0010 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0010 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0010 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0010 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0010 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0010 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0030 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0030 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0030 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0030 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0030 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0030 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0030 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0030 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0030 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0030 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0030 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0030 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0030 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0030 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0030 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0030 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0030 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0030 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0030 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0030 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0030 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0030 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0030 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0030 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0030 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0030 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0030 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0030 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0030 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0030 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0030 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0030 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0030 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0030 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0030 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0030 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0030 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0030 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0030 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0030 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0030 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0065 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0065 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0065 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0065 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0065 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0065 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0065 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0065 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0065 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0065 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0065 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0065 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0065 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0065 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0065 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0065 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0065 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0065 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0065 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0065 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0065 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0040 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0040 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0040 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0040 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0040 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0040 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0040 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0040 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0040 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0040 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0040 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0040 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0040 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0040 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0040 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0040 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0050 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0050 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0050 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0050 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0050 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0050 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0050 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0050 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0050 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0050 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0050 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0050 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0050 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0050 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0050 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0050 ring an F F G G Y.F. Rao et al. / Nuclear Enginee ortman, R.A., 2011. Critical heat flux and post-dryout tests using the reference 37- element fuel simulation in water. In: Stern Laboratories Report (Internal Report) SL-212 January. ortman, R.A., Hadaller, G.I., Hayes, R.C., Stern, F., 1997. Heat transfer studies with CANDU fuel simulators. In: 5th International Conference on Nuclear Engineering (ICONE5), Nice, France, May. roeneveld, D., Joober, J., Leung, L., et al., 1992. The effect of fuel subchannel geom- etry on CHF. In: Proceedings of NURETH-5, Salt Lake City. roeneveld, D.C., Shan, J.Q., Vasic, A.Z., Leung, L.K.H., Durmayaz, A., Yang, J., Cheng, S.C., Tanase, A., 2005. The 2005 CHF look-up table. In: 11th International Topical Meeting on Nuclear Reactor Thermal-Hydraulics (NURETH-11), Avignon, France, October 2–6, Paper No. 166. d Design 276 (2014) 216–227 227 Leung, L.K.H., Jun, J.S., Bullock, D.E., Inch, W.W.R., Suk, H.C., 2001. Dryout power in a CANFLEX bundle string with raised bearing pads. In: 7th International Confer- ence on CANDU Fuel Canadian Nuclear Society, September. Nava-Dominguez, A., Rao, Y.F., Waddington, G.M., 2014. Assessment of subchan- nel code ASSERT-PV for flow-distribution predictions. Nucl. Eng. Des. 275, 122–132. Rao, Y.F., Hammouda, N., 2003. Recent development in ASSERT-PV code for sub- channel thermalhydraulics. In: Proc. 8th CNS Int. Conf. on CANDU Fuel, Honey Harbor, ON. Rao, Y.F., Cheng, Z., Waddington, G.M., Nava-Dominguez, A., 2014. ASSERT-PV 3.2: advanced subchannel thermalhydraulics code for CANDU fuel bundles. Nucl. Eng. Des. 275, 69–79. http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0060 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0060 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0060 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0060 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0060 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0060 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0060 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0060 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0060 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0060 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0060 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0060 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0060 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0060 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0060 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0060 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0060 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0060 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0060 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0060 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0060 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0060 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0060 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0055 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0055 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0055 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0055 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0055 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0055 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0055 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0055 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0055 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0055 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0055 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0055 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0055 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0055 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0055 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0055 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0055 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0055 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0045 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0045 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0045 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0045 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0045 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0045 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0045 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0045 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0045 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0045 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0045 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0045 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0045 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0045 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0045 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0045 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0035 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0035 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0035 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0035 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0035 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0035 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0035 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0035 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0035 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0035 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0035 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0035 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0035 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0035 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0035 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0035 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0035 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0035 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0035 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0035 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0035 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0035 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0035 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0035 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0070 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0070 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0070 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0070 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0070 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0070 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0070 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0070 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0070 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0070 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0070 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0070 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0070 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0070 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0070 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0070 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0070 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0070 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0070 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0070 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0070 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0070 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0070 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0025 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0025 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0025 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0025 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0025 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0025 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0025 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0025 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0025 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0025 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0025 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0025 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0025 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0025 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0025 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0025 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0015 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0015 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0015 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0015 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0015 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0015 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0015 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0015 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0015 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0015 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0015 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0015 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0015 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0015 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0015 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0015 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0015 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0015 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0015 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0015 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0015 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0020 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0020 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0020 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0020 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0020 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0020 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0020 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0020 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0020 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0020 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0020 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0020 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0020 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0020 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0020 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0020 http://refhub.elsevier.com/S0029-5493(14)00342-2/sbref0020 Assessment of ASSERT-PV for prediction of critical heat flux in CANDU bundles 1 Introduction 2 Code assessment guidelines and methodology 2.1 Code assessment guidelines 2.2 Code assessment methodology 3 Experiment data and ASSERT idealization 3.1 Stern Lab bundle tests 3.2 ASSERT idealization 3.2.1 Subchannel geometry 3.2.2 Axial nodalization 3.2.3 Axial pressure-tube creep profiles 3.2.4 Boundary conditions 3.2.5 Fluid properties 3.2.6 Flow distribution modeling 3.2.7 CHF modeling 3.2.8 Calculation control 4 ASSERT-PV results and accuracy 5 Model sensitivity studies 5.1 Dryout power prediction 5.2 CHF axial location prediction 5.3 CHF radial location prediction 5.4 Summary of sensitivity studies 6 Conclusions References