Convective heat transfer within mechanically-ventilated building spaces
A hierarchy of interacting and interdependent approaches have been developed for calculating internal surface convective heat transfer coefficients within mechanically-ventilated rooms. A 'high-level' computer code is developed for non-bucyant and buoyant flow based on the 'elliptic' code of Pun and Spalding (1977), in which 'upwind' finite-difference approximations to the governing partial-differential equations for continuity, momentum and thermal energy are formulated in terms of 'primitive' pressure-velocity variables. Closure of these time-averaged, elliptic equations is obtained via transport equations for both the turbulence kinetic energy and its dissipation rate. The high-level code solves the difference equations for a predetermined size, staggered grid in an iterative 'line-by-line' manner using a guess-and-correct procedure. An 'intermediate-level' computer code (the ROOM-CHT program) has also been developed for the above purpose, which employs 'informed' estimates of the flow and thermal field based on the known mean flow properties of wall-jets. The corresponding heat transfer distribution across the room surface is calculated using wall-jet profile analysis or improved data correlations for bucyancy-driven convection as appropriate. Caqputations are presented for a room into which air is injected through a low or high side wall register. The supply of air governed by both cyclic and modulating control was examined. The intermediate-level code is advocated as being the most appropriate for meeting the requirements of dynamic building thermal models. This code was verified by comparison with the high-level code and with experimental measurements. The oomputed heat transfer coefficients from the intermediate-level code were found to be in good agreement with that of the high-level code. Both indicate significantly higher values than those which would be obtained from established design guides. These high values suggest errors in building thermal models based on guide data, including substantial under-estimation of preheat times.