This research provided a package of experimental and numerical results. The phenomenon of unusual turbulence intensity distributions in closely-spaced rod bundles was first experimentally studied. The structures of fully developed isothermal air turbulent flow through simulated rod bundles were measured with hot-wire anemometers and a Preston tube. The rod bundle was formed by a single rod regularly mounted in a trapezoidal duct, and the Reynolds number range was 2.5 × 10⁴ to 5.5 × 10⁴ The existence of a high turbulence kinetic energy patch was identified in the rod-to-wall gap region when the gap-to-diameter ratio g/d was within the range of 0.10 to 0.03 approximately. Energy density spectrum measurements revealed that cross-gap large-scale eddy motion is the probably mechanism behind this phenomenon and this kind of eddy motion is characterized by a peak frequency which depends on the geometry and flow condition. Secondary velocities on the rod setting with g/d as 0.220 were measured with X-probe;​ the obtained flow pattern coincided well with the bulges of wall shear stress and turbulence kinetic energy distributions. The impact of subchannel symmetry was also experimentally investigated.

Following these experimental studies, numerical predictions featuring the use of anisotropy factor and coordinate system transformation were carried out based on the finite volume method. Simulations were performed on fully-developed turbulent flow through simulated symmetric rod bundle subchannels formed by the rod-trapezoidal duct. With a unique coordinate system transformation from orthogonal cylindrical system to non-orthogonal curvilinear system, the highly irregular flow passage of rod-trapezoidal duct was converted to a regular rectangle. An empirical anisotropic eddy viscosity distribution derived from existing experimental data was used in conjunction with the algebraic stress model to address the influence of coherent large-scale cross-gap eddy motion, whose existence in closely spaced rod bundle subchannels has been substantiated by the extensive hot-wire measurements. Results of the calculation compared favourably with experimental data, with emphasis on secondary flow and turbulence kinetic energy. The credibility of this numerical scheme was established through a series of numerical tests on simple geometry flows.

Further numerical prediction on fully developed turbulent flow through simulated asymmetric rod bundle subchannels was carried out using the finite element method and standard two-equation turbulence model. Preliminary results indicated a large single vortex circulating around the rod which might explain the high mixing factors for asymmetric rod-to-wall gaps.