ART3-CFD Simulation and Analysis of Reactor Integral Hydraulic Tests

of 12
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  CFD simulation and analysis of reactor integral hydraulic tests Shuo Chen, Hanyang Gu ⇑ School of Nuclear Science and Engineering, Shanghai Jiao Tong University, Shanghai 200240, China a r t i c l e i n f o  Article history: Received 3 May 2019Received in revised form 12 June 2019Accepted 27 July 2019 Keywords: CFDFlow distributionScaling velocityCore resistance distributionLateral resistance a b s t r a c t Scaled-down reactor models are widely used to investigate the reactor integral hydraulic characteristics,due to the feasibility and economy. However, no definite conclusion can be drawn from previousresearches for the effects of different scaling methods on simulation results. In this paper, in order to ana-lyze the effects of simulation core design parameters and flow flux on core inlet flow distribution, a CFDsimulation is conductedbased on thehydraulic testsof the1/4 scale modelof 600 MWreactor inQinshanphase II. The effects of different scaling velocities, core resistance distributions and lateral resistance onflow distribution are investigated. Calculation results reveal that the uniformity of flow distributionincreases with test velocity. Ignoring the resistance of fuel assembly would reduce the uniformity of flowdistribution. The flow distributions under conditions of uniformly and segmented distributed core resis-tance are basically the same, indicating that the detailed simulation of axial pressure drop distribution inassemblies is unnecessary for the design of hydraulic simulator of assembly. The lateral resistance couldsignificantly affect the flow distribution and excessive lateral resistance would reduce the uniformity of distribution.   2019 Elsevier Ltd. All rights reserved. 1. Introduction Thermal hydraulic characteristics of reactor core are directlyaffected by the flow distribution at the core inlet, which wouldinfluence the operating limits of nuclear power plants ( Jeonget al., 2005). Thus, verification tests of core inlet flow distributionmust be carried out in the process of safety review for the designof core of new-structure reactor. In consideration of the feasibilityand economy of tests, the scaled-down reactor models are widelyused to investigate the reactor integral hydraulic characteristics.Takingvariousinfluencingfactorsinto account,differentscalingmethods were adopted in different research institutions aroundthe world. In terms of the scaling velocity, linear scaling methodol-ogywasusedinthe1/4scalemodelofQinshanphaseIIreactorandthe scaling velocity was 0.6 times of that of the prototype (Yanget al., 2003). Modified linear scaling methodology was applied inthe 1/5 scale model of APR + reactor and  ffiffiffi  5 p   = 5 times of prototypevelocity was selected as the scaling velocity (Euh et al., 2011;Euh et al., 2012a). Unlike two scaling methods mentioned above,researchers kept the scaling velocity the same as that of the proto-type in the 1/5 scale model of SMART reactor (Euh et al., 2012b).For the simulation of reactor core, there were different ways todeal with the core resistance according to previous studies. In Qin-shan phase-II reactor model, the core was divided into several seg-ments, of which the resistance was the same as that of theprototype respectively (Wang and Zong, 2000). Remarkably, theresistances were different for each segment. In APR + reactormodel, the core pressure drop was scaled with some scaling lawsand four perforated plates were regularly installed in the core sim-ulator to have a uniform resistance distribution (Bae et al., 2011).However, the core regions were empty in US-APWR reactor model.The resistance of fuel assembly was ignored in the tests(Watanabe, 2008).In present literatures, no definite conclusion is made for theeffects of different scaling methods of velocity and core resistanceon flow distribution. In this paper, the 1/4 scale model of 600 MWreactor in Qinshan phase II is taken for research objects. The effectsof scaling velocity, resistance distribution and lateral resistance onthe core inlet flow distribution are numerically investigated bycomputational fluid dynamics (CFD). The simulation results arevalidated with test data. The analysis can provide reference forthe reactor integral hydraulic tests. 2. Numerical approach  2.1. Geometry Fig. 1 shows the geometry of calculation regions, including twoinlets, downcomer, lower plenum, core inlet dome, core, upper   2019 Elsevier Ltd. All rights reserved. ⇑ Corresponding author. E-mail address: (H. Gu).Annals of Nuclear Energy 135 (2020) 106962 Contents lists available at ScienceDirect Annals of Nuclear Energy journal homepage:  plenum and two outlets. Since the test geometry is relatively com-plex, some simplification is utilized in several parts of calculationregions. The heat shielding boards in small scale and thin thicknessare simplified, due to the little effect on the flow distribution(Zhang et al., 2010a,b). Before the simplification, 484 holes arelocated in the core inlet dome and every four of them arecorresponding to each of the 121 fuel assemblies in the core.According to the area equivalent principle, 121 simplified holes Fig. 1.  Schematic of calculation regions. 0 2 4 6 8 10 12 Column G Mesh-A Mesh-B Mesh-C     M  a  s  s   F   l  o  w   R  a   t  e   [   k  g   /  s   ]  Assembly Number  Fig. 2.  Mesh independence.  Table 1 Calculation settings of all cases. Case U inlet (m/s)U inlet /U p  MinimumReynoldsnumberCoreResistanceDistributionC R2l  = C R2a 1 4.35 1/4 47,960 Uniform Distribution 102 8.70 1 =  ffiffiffi  4 p   95,920 Uniform Distribution 103 10.44 0.6 115,100 Uniform Distribution 104 17.40 1 191,830 Uniform Distribution 105 10.44 0.6 115,100 No Resistance —6 10.44 0.6 115,100 Segmented Distribution 107 10.44 0.6 115,100 Uniform Distribution 28 10.44 0.6 115,100 Uniform Distribution 20 Nomenclature C  R1 linear resistance coefficient, kg/(m 3  s) C  R2 quadratic resistance coefficient, kg/m 4 K  loss  empirical loss coefficient, 1/m K   perm  permeability, m 2 n  number of assemblies q i  normalized mass flux of assembly  iQ  i  mass flux of assembly  i , kg/s U  coreinlet   core inlet velocity, m/s U  inlet   inlet velocity, m/s U   p  prototype velocity, m/s V   physical volume of the porous medium, m 3 V’   volume available to flow, m 3 D  p core  core pressure drop, Pa Greek symbols c  volume porosity l  dynamic viscosity, Pa  s q  density, kg/m 3 f  local resistance coefficient, 1/m Subscripts a axiall lateral Fig. 3.  Adopted mesh of the reactor.2  S. Chen, H. Gu/Annals of Nuclear Energy 135 (2020) 106962  Fig. 4.  Streamlines in reactor model. Fig. 5.  Calculation results of normalized mass flux. S. Chen, H. Gu/Annals of Nuclear Energy 135 (2020) 106962  3  are substituted for the 484 holes to reduce meshing difficulties andcomputing time. What’s more, porous media is introduced to sim-ulate the complicated core region which is filled with lots of fuelassemblies. The upper plenum imposes little influence on the flowdistribution, due to the great distancesfrom thecore inlet. Thus, anempty cylinder is selected to modify the upper plenum. Unlike themodified regions mentioned above, the lower plenum composedwith combined supporting frames and some other components isnot simplified resulting from their significant effects on the flow.In the whole geometry structure, only the core region presents agood symmetry.  2.2. Mesh generation and independency The mesh generations in all computational regions are dis-played in Fig. 2. Combining the structured and the unstructuredgrids, different types of mesh are utilized in different calculationregions. Due to the irregularity of shape or complexity in structure,the tetrahedral mesh is utilized in the lower plenum and flowregions near the inlet and outlet. In other parts of regular shape,the hexahedral mesh is utilized for reducing the calculationresource. As reported by Yan et al. (2012), General Grid Interface(GGI) could provide satisfied connection between different typesof mesh elements and allow non-uniformity of node location onthe interface. Hence, the GGI connection is adopted in the connec-tion of various mesh types in the meshing progress. The core inletflow distribution is greatly influenced by the severe mixing of flowfluidin thelowerplenum.As a resultof connectingthelowerfluidswith the core, supporting columns also have a significant effect onflow distribution. Therefore, mesh refinements are utilized in theseregions.To guarantee the independence of the calculated results withthe mesh numbers, three kinds of mesh are utilized in the sensitiv-ity study. The total numbers of mesh elements are 12 million(Mesh-A), 34 million (Mesh-B) and 65 million (Mesh-C) respec-tively. Calculations are performed under the same conditions givenincase 3 in Table 1. As illustrated inFig. 2, the calculatedmassflow rateinColumnG (Thelocationis giveninFig.5)isselectedto makethe comparison. It is found that the difference between results of Mesh-B and Mesh-C is considerably less. However, the simulationresult of Mesh-A is not in good consistent with the others shownin Fig. 6.  Deviations between calculation and test values of normalized mass flux. -16-14-12-10-8-6-4-20246810121416051015202530354045    F  r  e  q  u  e  n  c  y % Difference (Calculation to Test) Fig. 7.  Statistic results of calculated deviations.4  S. Chen, H. Gu/Annals of Nuclear Energy 135 (2020) 106962
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!