Experimental study of MHD effects on turbulent flow of Flibe simulant fluid in circular pipeIntroductionExperimentsExperimental apparatusMagnetPIV systemThermophysical properties of KOH solutionResults and discussionMean velocityTurbulence intensitiesReynolds shear stressConclusionsAcknowledgementsReferencesFusion Engineering and Design 83 (2008) 1082–1086Contents lists available at ScienceDirectFusion Engineering and Designjournal homepage: www.elsevier.com/locate/fusengdesExperimental study of MHD effects on turbulent flow of Flibe simulant fluid incircular pipeJunichi Takeuchia,∗, Shin-ichi Satakeb, Neil B. Morleya, Tomoaki Kunugic,Takehiko Yokomined, Mohamed A. AbdouaaMechanical and Aerospace Engineering Department, Univeristy of California, Los Angeles, 420 Westwood Plaza, 44-114 Engineering IV, Los Angeles, CA 90095, USAbDepartment of Applied Electronics, Tokyo University of Science, 2641 Yamazaki, Noda, Chiba 278-8510, JapancDepartment of Nuclear Engineering, Kyoto University, Yoshida, Sakyo, Kyoto 606-8501, JapandInterdisciplinary Graduate School of Engineering Science, Kyushu University, 6-1 Kasuga-koen, Kasuga, Fukuoka 816-8580, Japanarticle infoArticle history:Available online 5 November 2008Keywords:MHDFlibeTurbulent pipe flowPIVabstractAn investigation of MHD effects on a Flibe (Li2BeF4) simulant fluid has been conducted under theU.S.–Japan JUPITER-II collaboration program using the “FLIHY” pipe flow facility at UCLA. The presentpaper reports experimental results on turbulent pipe flow of an aqueous potassium hydroxide solutionunder magnetic field using particle image velocimetry (PIV) technique. The modification of turbulencewas investigated by comparison of the experimental results with a direct numerical simulation (DNS)data base. The PIV measurements at Re = 11,300 were performed with variable Hartmann numbers, andthe modification of the mean flow velocity as well as turbulence reduction was observed. A flat velocityprofile in the pipe center and a steep velocity gradient in the near-wall region exhibit typical character-istics of wall-bounded MHD flows. The DNS was performed approximately the same conditions and thecomparison of turbulence statistics between PIV and DNS shows good agreement for up to Ha = 10.© 2008 Elsevier B.V. All rights reserved.1. IntroductionThe design of tritium breeding blankets and plasma facing com-ponents is an important area of R&D activities toward a viablecommercial nuclear fusion reactor. In recent research, a molten saltcoolant, Flibe (Li2BeF4), has attracted attention. Moriyama et al. [1]surveyed various design concepts using Flibe and suggested its usein reactor designs where high temperature stability and low MHDpressure drop were special concerns. Among the design conceptsutilizing Flibe are HYLIFE-II [2], the APEX thick/thin liquid walls[3], FFHR [4], and a solid first wall design based on advanced nano-composite ferritic steel [5]. Although Flibe has attractive features ascoolant and tritium breeding material, there are some issues mak-ing Flibe-based blanket design challenging [5]. The main issuesinclude (1) thermal conductivity of Flibe (1 W/mK) is low com-pared to other lithium-containing metal alloys, Pb–17Li (15 W/mK)and Li (50 W/mK), (2) kinematic viscosity of Flibe is high, espe-cially at temperatures close to the melting point (11.5 × 10−6m2/sat 500◦C), and (3) the high melting point of Flibe requires structuralmaterial with temperature range over 650◦C.∗Corresponding author. Tel.: +1 310 794 4452; fax: +1 310 825 2599.E-mail address: [email protected] (J. Takeuchi).The limited operating temperature window of Flibe coolantrequires good heat transfer from heated surface to the bulk flow.However, the high viscosity and low thermal conductivity put Flibein the class of high Prandtl number fluids. In order to obtain suffi-ciently large heat transfer using high Prandtl number fluid coolant,strong turbulence is required. On the other hand, Wong et al. [5]suggested that the parameter Ha/Re would exceed the critical valueof 0.008 given by Branover [6], especially in large channels, whichindicated the suppression of turbulence might be significant.In this paper, the Reynolds number is defined as Re = UbD/,and the Hartmann number as Ha = BR / where Ub, D, , B,R, and  are mean velocity, pipe diameter, kinematic viscos-ity, magnetic flux density, pipe radius, electrical conductivity andfluid density, respectively. Comparisons of typical non-dimensionalparameters between the current experiments and the design pro-posed by Wong et al. [5] are shown in Table 1.The MHD effects on turbulent flows have been investigatedextensively; however, most of the experimental efforts were con-ducted using liquid metals as working fluids [7,8]. Liquid metalsare generally classifie d as low Prandtl number fluids, and the heattransfer characteristics of low Prandtl number fluids are conduc-tion dominant. However, the MHD effects on high Prandtl number,low conductivity fluids are yet to be understood. Thus it is impor-tant to investigate the effect of magnetic fields on the turbulent0920-3796/$ – see front matter © 2008 Elsevier B.V. All rights reserved.doi:10.1016/j.fusengdes.2008.08.050J. Takeuchi et al. / Fusion Engineering and Design 83 (2008) 1082–1086 1083Table 1Comparison of typical dimensionless parameters for the present experiment withmolten salt blanket [5].Ha Re Ha/Re N(=Ha2/Re)Experiment 20 11,300 1.8 × 10−33.5 × 10−2Self-cooled fusion blanket 20 12,000 1.7 × 10−33.3 × 10−2flow and heat transfer characteristics of the high Prandtl numberfluids.To understand the underlying phenomena controlling the fluidmechanics and heat transfer of Flibe, a series of experiments as apart of the U.S.–Japan JUPITER-II collaboration has been conducted.The approach includes flow and heat transfer measurements usinga Flibe simulant fluid along with numerical simulation and model-ing. In the preceding research, free surface flows were investigatedthrough experiments and modeling [9,10]. In the present phase, aflow facility utilizing water and aqueous electrolytes as Flibe sim-ulants has been constructed, and flows in a close channel are beinginvestigated. Turbulent flow field measurements using PIV [11] andheat transfer measurements [12] have been carried out withoutmagnetic field to establish the experimental techniques and verifythe performance of