An experimental investigation on turbulent natural convection boundary layers has been conducted with water on a vertical plate of constant heat flux. Local heat transfer data are presented for laminar, transition, and turbulent natural convection, with the emphasis on the turbulent regime. The data extend to a modified Rayleigh number of 1016 for a threefold range in Prandtl number. The results indicate that natural transition occurs in the range 1012 < Ra* < 1014; i.e., fully developed turbulent flow occurs by Ra* = 104. This latter value can be as low as 2 × 1013 with the use of a trip rod. The physical structure of the turbulent boundary-layer flow was studied using the combined time-streak marker hydrogen bubble method. Temperature data and temperature corrected velocity data obtained by hot-film sensors are presented for Ra* values between 8.7 × 1013 and 7.1 × 1014. For the range of variables investigated, the major conclusions are (a) the local heat transfer coefficient exhibits a slight decrease with length, (b) confirmation that the vortex street layer in the transition region decays into a longitudinal-vortex-type structure, and (c) the outer portion of the thermal and velocity fields can be approximated by power profiles that fit almost all the data available to date.

Experimental local heat transfer data are presented for natural convection on constant-heat-flux inclined surfaces using water and air. The data extend to Grz* Pr = 1016, cover angles from the vertical to 30 deg with the horizontal, and include the laminar, transition, and turbulent regimes. In the laminar regime the data correlate well with vertical plate theory when the gravitational component parallel to the surface is used. Transition is strongly affected by inclination, the transition Grz* Pr decreasing from near 1013 for vertical surfaces to approximately 108 for a surface at 30 deg to the horizontal. The turbulent local heat transfer data correlate using the actual gravity rather than the parallel component, and indicates a change in the Grz* Pr exponent from near 0 22 for a vertical surface to approximately 1/4 as the inclination decreases. The turbulent data can be correlated quite well by Nuz = 0.30(Grz* Pr)0.24.

Local heat transfer data were obtained for turbulent natural convection on vertical and inclined upward and downward facing surfaces. The test surface consisted of a 1.83 m (6 ft) wide × 7.32 m (24 ft) high plate with a constant heat flux obtained by electrical resistive heating of a metal foil on the surface. The tests were conducted in air for modified Grashof numbers up to 1015. Measurements were made of the local surface temperature for this constant heat flux condition, for the plate inclined at angles from 30 deg to the vertical (upward facing, unstable) through the vertical to 80 deg to the vertical (downward facing, stable). The results show the location of the transition to be a function of the plate angle. For the unstable case, the transition length decreases as the plate angle increases from the vertical while for the stable case the position of transition increases with the angle from the vertical. The laminar data for both orientations are correlated as: Nux=0.55(Grx*Pr)0.20 in which the gravity is the component along the surface, g cos θ. The turbulent natural convection data are correlated quite well by the relation: Nux=0.17(Grx*Pr)0.25 In the turbulent case the correlation is independent of angle for the unstable case, whereas for the stable case the data correlate best when the gravity is modified by cos2 θ, where θ is measured from the vertical. Thus, there is a significant influence of angle on the convective heat transfer for the stable turbulent region.

This i nves t i gation i nvol ved the devel opment of a numerical model f o r the t r a n s i e n t simulation of the double-effect, water-1 ithium bromide absorption cooling machine, and the use of the model to determine the e f f e c t of the various design and input variables on the absorption u n i t performance. The performance parameters considered were c o e f f i c i e n t of performance and cool i ng capacity. formed by selecting a "nominal condition" and determining performance sensi t i vi ty f o r each variabl e w i t h others he1 d constant.

No abstract