Date of Award

12-1-2004

Document Type

Thesis

Degree Name

Master of Science (MS)

Department

Mechanical Engineering

Abstract

Pin fin arrays are one of the more common internal cooling features in many turbine vanes and blades. Pin fins increase turbulence of passage flow and internal heat transfer surface area. Most pin fin array designs use empirical correlations to predict heat transfer rates and pressure drop. Previous research suggests computational models for heat transfer and velocity distribution predictions can be improved if the characteristics of turbulence and its response near surfaces such as pins and endwalls are known. Detailed heat transfer and turbulence data are needed to investigate why current turbulence models fail. Improvements can be made by introducing more accurate physics into these models.

A cause and effect between local heat transfer and the local velocity and turbulence distributions in pin fin arrays is needed to advance understanding and improve predictive modeling. In this research a comprehensive set of data including surface static pressure, velocity, turbulent components, and turbulent spectral information was acquired in a staggered pin fin array. Data for local heat transfer were previously recorded. Hot wire anemometry data were taken for Reynolds numbers of 3,000, 10,000, and 30,000.

Hot wire measurements were acquired off the pin and off the endwall to show a specific cause and effect between local heat transfer and turbulent transport at various locations within the array. Detailed measurements were taken in rows 2, 3, and 4 of the array, as turbulence generation from pins upstream appeared to have the most profound effect on the physics of the flow in these locations. Turbulence measurements and energy spectra were acquired using first the single-wire technique and then using x-wire techniques. These measurements have been taken in regions without separation. Comprehensive velocity profiles off the pin and endwall with single-wire and x-wire surveys across the pin spaces were compared with computational fluid dynamic (CFD) predictions. Additionally, v’ spectra were taken at locations off the endwall where corresponding heat transfer measurements were available.

Share

COinS