| Heat transfer across an air/water interface is of particular importance to limnology, oceanography and some industrial applications. The relationship between the statistics of the air/water interfacial temperature field and the interfacial heat flux is poorly understood, particularly for the mixed convection condition, which is a common heat transfer mechanism for small inland lakes. The few studies that have been conducted under mixed convection conditions have been limited to an uncontrolled surfactant condition (tap water). Therefore, in this dissertation research two sets of experiments for wind speeds from 0 to 4 m/s were conducted: controlled surfactant contaminated conditions (with oleyl alcohol) and clean water surface conditions. The air/water interfacial heat flux and the surface temperature field statistics (root mean square (RMS) sigma and skewness gamma) were computed to study the relationship between them, and the results under different surface conditions were presented and compared. It was found that, for a given wind speed and surface condition, the RMS of the interfacial temperature field increased linearly with heat flux, and the RMS of clean surfaces was greater than that of a surfactant-covered surface. The surface skewness gamma for clean surfaces was found to be more negatively skewed than that under surfactant-covered surfaces. There was almost no wind speed effect on the surfactant-covered surface skewness. However, the clean surface skewness became less skewed when the wind speed increased. The RMS was scaled by DeltaT = Tb -- Ts, which is the maximum possible value of sigma. The scaled RMS sigma/(Tb -- Ts) decreased by a factor of two in the presence of a surfactant monolayer. A parameterization study was also carried out to find the relationship between sigma/( Tb -- Ts) and the Rayleigh number Raq, the Reynolds number Re* and the Prandtl number Pr. It was found that sigma/Delta T = 1.11 x 10-3 Ra0.37q Re*-0.81Pr -1/3 for the surfactant-covered case and sigma/Delta T = 2.99 x 10-3 Ra0.38q Re*-0.90Pr -1/3 for clean surfaces. In both cases, if the exponents of Raq and Re* are rounded to the nearest fraction, the correlations changed to sigma/DeltaT = A · Ra1/3q Re*-4/5Pr -1/3, where A = 2.56 x 10-3 for the surfactant case and A = 5.66 x 10 -3 for the clean case, respectively. The effect of the surfactant monolayer on the relationship between sigma/DeltaT and (Raq, Re*, Pr) was parameterized by introducing a new dimensionless group Dp = Ehrn2 . The probability density functions (PDFs) of the surface temperature fields were also determined. The wind speed, heat flux, and surface condition were all found to affect the temperature PDFs. Finally, the presence of longitudinal vortices, which are an air-side phenomenon, oriented in the wind direction, were observed under certain wind speeds and air/water temperature differences. Experiments were conducted to investigate their onset instability mechanism. The streak spacing and the onset position varied with the Reynolds number and the Grashof number. This research provides an improved understanding of turbulence using an experimental model that is more relevant to lakes than is the case for Rayleigh-Benard convection, which is often used as a model of lakes and oceans. This research also finds application in small lake thermal modeling, atmospheric modeling, volcanic lake modeling, treaty verification, the prediction of ice formation, gas/mass transfer studies, and metal surface solidification. |
