EXPERIMENTAL STUDY OF AXIAL FLOW IN A FINITE ARRAY OF RODS AND THE APPLICATION OF FINITE ELEMENT TECHNIQUES TO FLOW IN DUCTS AND ROD BUNDLES

Experimental measurements of the wall shear stress, axial velocity, axial turbulence intensity and turbulence power spectra were made for axial turbulent flow along a rectangular 3 x 6 array of rods with a pitch to diameter ratio = 4/3.; Wall shear stress measurements were performed on six rods and on two side walls using hot film probes. Measurements of the axial velocity, axial turbulence intensity and the turbulence spectra were performed in the corner, wall and central subchannels of the array using hot wire probes.; The results show that the local shear stress maxima occur near the largest flow areas. The ratio of the maximum to the minimum shear stress on an individual rod is largest for the corner rod. Side wall maximum local shear stress occurs in the first wall-subchannel. Overall friction factors calculated from the wall shear stress measurements agree with those calculated from pressure drop data.; Measured axial velocity and axial turbulence intensity distributions agree with available experimental and analytical results. Secondary flow effects on these distributions and on the wall shear stress distribution are more pronounced in the corner and the adjacent subchannels than in the central subchannels.; An analytical study of turbulent flow in rod bundle subchannels was carried out using the finite element method. The computational scheme used in the analysis employed the Galerkin weighted residual method and the model of turbulence proposed by Roco and Zarea (1976).; The applicability of the computational scheme to flow in subchannels of square and triangular rod arrays was investigated at different values of pitch to diameter ratio and Reynolds number. The results indicate the correct qualitative trends of the axial velocity and shear stress distributions. Differences between the calculations and the available experimental results were caused by secondary flows, which were neglected in the analysis.; The present computational scheme has displayed a great range of applicability to flows in complex geometries. However, detailed information on some of the parameters used in the turbulence model is necessary to obtain better agreement between the predictions and reliable experimental results.