Date of Award

Spring 5-2017

Embargo Period

5-25-2017

Degree Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Mechanical Engineering

Advisor(s)

Shi-Chune Yao

Abstract

Pressurized water reactor nuclear plants, currently under construction, have been designed with passive containment cooling systems. Turbulent, natural-convective condensation, with high non-condensable mass fraction, on the walls of the containment vessel is a primary heat transfer mechanism in these new plant designs. A number of studies have been completed over the past two decades to justify use of the heat and mass transfer analogy for this scenario. A majority of these studies are founded upon natural-convective heat transfer correlations and apply a diffusion layer model to couple heat and mass transfer. Reasonable success in predicting experimental trends for vertical surfaces has been achieved when correction factors are applied. The corrections are attributed to mass transfer suction, film waviness or mist formation, even though little experimental evidence exists to justify these claims. This work examines the influence of film waves and mass transfer suction on the turbulent, natural-convective condensing flow with non-condensable gas present. Testing was conducted using 0.457 m x 2.13 m and a 0.914 m x 2.13 m condensing surfaces suspended in a large pressure vessel. The test surfaces could be rotated from vertical to horizontal to examine the inclination angle effect. The test facility implements relatively high accuracy calorimetric and condensate mass flow measurements to validate the measured heat and mass transfer rates. Test results show that application of the Bayley (1955) and Al-Arabi and Sakr (1988) heat transfer correlations using the heat and mass transfer analogy is appropriate for conditions in which the liquid film remains laminar. For transitional and wavy film flows, a clear augmentation in heat transfer was observed due to disruption of the gas layer by film waves. This result has implications for the scalability of existing correlations. A new correlation is proposed and results compared to several other datasets.

Share

COinS