Page 88 - Thermal Hydraulics Aspects of Liquid Metal Cooled Nuclear Reactors
P. 88

Rod bundle and pool-type experiments in water serving liquid metal reactors  63


            Table 3.1.2 Some RIM solids with a refractive index close to water

                                                             Absorption coefficient
                                  n                               21
            Material                                         α (cm  )
            Fluorinated poly-     1.33 (Hassan and Dominguez-  1.2 (Mahmood, 2011)
            ethylene propylene    Ontiveros, 2008)
            Teflon AF2400         1.29 (Yang et al., 2008)
            Teflon AF1601         1.31 (Yang et al., 2008)
            Teflon AF1300         1.31 (Yang et al., 2008)
            Cytop                 1.34 (Cyt, 2017)           0.08 (Wu and Knoesen,
                                                             2001)

            Notes: Values for light of wave length λ ¼ 532 nm. Absorption coefficients are 10-based.

           Hassan (2009) used FEP for the determination of the flow field in a 5   5 rod bundle
           geometry (in conjunction with PTV) and Mahmood et al. (2011) performed similar
           experiments, thereby using LDA. The effect FEP has (or better: has not) on the refrac-
           tion of light when submerged in water is shown in Fig. 3.1.5. There are, however, two
           issues that require attention when using FEP, being

           l  compared with the very transparent PMMA, FEP has a relatively large absorption coefficient
              for visible light; and
           l  due to the slight opacity, thin foils have to be used, making the construction less stiff and
              subject to pressure variations in the flow.


           Absorption coefficient
           Although the refractive index matches the one of water, the absorption coefficient is
           relatively large. A plate of FEP with a thickness of 0.5 cm, for example, attenuates by

                  Ið0:5Þ    1:2 0:5
               1        ¼ 10      ¼ 75%:
                   Ið0Þ
           We therefore need to make the internal walls (being the rods in the rod bundle) as thin
           as possible. In Dominguez-Ontiveros and Hassan (2009), for example, a thickness of
           0.125 mm has been used, giving an attenuation of about 4%. Mahmood et al. (2011)
           used a thickness of 0.25 mm, giving an attenuation of 7%. There are, however, two
           additional difficulties that need to be taken into account:

              For optical techniques, light is sent into the facility and has to be detected after being
           l
              scattered (in the form of, e.g., recorded digital images or scattered LDA laser beams). Hence,
              light requires at least to cross FEP walls twice. But in most practical cases, a single rod is
              passed two times, resulting in light transmitted through an FEP wall four times. For a wall
              thickness of 0.25 mm, this would mean that the attenuation already becomes 25%.
           l  Usually light crosses the solid walls slantwise and not in a perpendicular way. Especially
              for circular geometries this may result in a large effective thickness, as is shown in
              Fig. 3.1.5.
   83   84   85   86   87   88   89   90   91   92   93