Steady natural convective flow fields were numerically and experimentally characterized for 0.7 Prandtl number fluids having constant, uniformly distributed, internal heat sources. The bounding isothermal walls containing the fluid were considered to be either a sphere or a cylinder of finite height. An instrumented cylinder containing radioactive tritium gas was used to demonstrate experimental and analytical agreement for local temperatures over a range of Grashof numbers. For the spherical geometry, a generalized correlation was obtained for the surface-averaged Nusselt number as a function of a modified Grashof number.

A gas, discharging from its container through an orifice to the atmosphere, is shown to experience a temperature history that is strongly dependent upon the heat transfer from the container wall. Using a quasi-steady free-convection model for the instantaneous conductance between the wall and gas, a satisfactory correlation results between experimental and analytical mean gas-temperature response during discharge. For the special case of choked orifice flow and constant wall temperature, the mean temporal temperature of the gas remaining within the vessel is shown to depend upon just two parameters. The onset of the oscillating component of gas temperature is shown to occur at times which correspond to the dramatic growth of the diffusional thermal boundary layer on the wall, and this finding is in agreement with the Malkus theory of turbulence. The relative size of the temperature oscillations during discharge is shown to reach a maximum at the time when the scale of dimensionless temperature, a parameter essentially dependent on the instantaneous mean gas temperature, is a maximum.

No abstract

An experimental technique is presented to identify fusion boundary (liquid/solid interface) energy transport mechanisms during welding procedures. The gas-tungsten-arc spot-welding procedure, using a low melting point specimen material (lead), was chosen to demonstrate the methods. Vaporization energy losses were found to be important during the growth of the fusion boundary. Significant thermal convection was absent within the weld pool for applied currents less than about 100 A, and for such cases the location of the fusion boundary was found to be governed primarily by heat conduction. At the current levels of almost 300 A, significant weld pool convection was found to exist, especially at the (inner) stagnation point, causing a deeper penetration of the fusion boundary there.

'lhiadocurnent w=prepard aaanaccountofwork qmwored byanagenry of the United States Govemummt. Neither the United Statae~V emmentnor theuniversity ofcMornia noranyoftheir employees, makes any warranty, axpraae or impk$ or iwannee any legal Hty m responsibility for * accwacy, Complatenaea, or wefubsaofanyhfcmna~apparatus, prod~orprocesaõ r qmaenta that ita uaa would not infringe privately owned rights. Refamca harehtoanyapecific commedal prod~proc=, or service by trade name, trademark, mm&Xmer, or OtkWiSE!, doee not Wxmarily corwtitute or imply ite mdoraamant, recommertdati~or favoring by the United Statea (%Vamment or the hk'lity of California.~views and O@I1.iOIM of authora expresd herein do not mcesmdly state or reflect those of the United States Gov ernment or the university of Californk andskllnotbeuse!d klradvertk4ngor pKlductendorseInent pwpoaee. Ihisreport haabeenreproduced directly from the beet available copy. Available to DOEand DOEcontradDrsfrarn the Office of Scientific and Tecimkal Information P.o.Box62,m Ridge, m 37831 Pricaa available from (615) 576S401, PTS6W6401