Award Date

5-1-2019

Degree Type

Thesis

Degree Name

Master of Science in Engineering (MSE)

Department

Mechanical Engineering

First Committee Member

Yi-Tung Chen

Second Committee Member

William Culbreth

Third Committee Member

Robert Boehm

Fourth Committee Member

Jichun Li

Number of Pages

66

Abstract

Reactor fuel rod bundles serve as the primary heat source in light water reactors (LWRs), commonly found in the aforementioned PWR plants. The fuel rod bundles’ structure consists of a collection of fuel rods put into a parallel grid configuration. The bundles also include fuel rod spacers, which hold the fuel rods in place, in accordance with the grid. Repeating structures of the fuel bundles create the meta-structure in the reactor. In other words, the grid configuration repeats until it fills the entire space of the reactor. This results in reactor fuel rods suspended in the working fluid domain, oriented parallel length-wise to the flow direction, by the spacer grids.

The generated heat from the fission reactions within the fuel rod elements provide the primary heat source for the power cycle. As the working fluid, light water, in this case, flows through the reactor, the heat generated by the fuel rods’ fission reactions is transferred to the fluid, adding its potential to do work. Taking advantage buoyancy effects of the heated working fluid, reactors usually have the working fluid enter through the bottom, then pumped up vertically through the fuel rod bundles and spacers.

Since the working fluid flows through a flow region inside the reactor, containing fuel rod elements and the spacer grids, the drag force caused by these obstacles requires extra pumping power to overcome. This need for extra pumping power lowers total thermal system efficiency. Fortunately, the spacers have extra geometries in the end called mixing vanes, which agitate the working fluid as it flows through the reactor,

disturbing the hydraulic and thermal boundary layers. As these boundary layers are disturbed, heat transfer rate increases; which, in turn, increases the amount of energy added to the working fluid during the heat addition phase of the thermal power cycle, adding more potential to do work in the work output phase of the cycle.

Focusing only on hydraulic performances, this study attempted to lower hydraulic pressure drop across the spacer grid by numerically simulating spacer grids with different changes to their geometries inside a flow field. Several geometrical variations were chosen due to their ease of manufacturing and minimal tooling changes required at the supplier level; these variations include spacer grid length, mixing vane angle, spacer grid entrance geometry, and mixing vane shape.

This study used the sectional method proven by Conner et al. (2010) and Navarro et al. (2011), to establish its baseline. In order to save the limited resources in computational power, the results from Navarro et al. (2011) were first transferred from CFX to Fluent 18.2. This step required the numerical results from both software to be compared and benchmarked. Fluent’s segregated method of calculating velocity and pressure makes the calculations easier on the limited hardware. In addition, Fluent’s finite volume method with cell-centered scheme also allows solutions to more accurately reflect cases where unstructured meshes are used, such as this study. To justify the results, this study also introduced different fluid domain inlets, additional boundary layers, and finer mesh sizes than the previous studies that provided the baseline.

Mesh independent study was done to find the correct mesh size for a good compromise between resolution and convergence time. The results show that an element count of 4.27 × 107 or greater yielded computational results independent from element counts. However, since going with a higher element count does not significantly lengthen the computational time, the highest element count 6.18 × 107 elements, along with its

respective body-sizing, 3.0 × 10−4 m, were chosen. The resulting y+ values of this study was around 1.75, less than the value used by Navarro et al. (2011). The total pressure drop across the region of interest also closely replicated the previous results found by various studies.

Lastly, the study compared results from each variation. Shortening the spacer grid length decreases the pressure drop across its span. However, with a 25% reduction in spacer grid length, the pressure drop only reduced by 10.2%. This implied that the reduction in spacer grid length does not scale in unity with the reduction in pressure drop. Fortunately, increasing mixing vane angle significantly increase the agitation of the

boundary layers. Specifically, a 20% increase in the mixing vane angle resulted in a 15.6% increase in swirl-factor, with only around 9.28% increase in pressure drop. Meanwhile, adding a 45-degree chamfer to the entrance of the spacer grid, with depth that bisects its thickness, decreased the pressure drop by 47%, without adding significant manufacturing steps to the construction. Then, a curvature was added to the mixing vane’s profile to ease fluid flow’s transition back to the freestream. However, the curvature on the mixing vane had detrimental effects on the overall performance, increasing the pressure drop across the spacer by 31.3%, while decreasing swirl factor by 4.3%.

Keywords

Computational fluid dynamics; Nuclear spacer grid; Numerical simulation

Disciplines

Aerodynamics and Fluid Mechanics | Nuclear Engineering | Thermodynamics

Language

English


Share

COinS