Embedded Detail Microscopic Models of Rooms within Macroscopic Models of Whole Building Systems

摘要

Established methods of computational fluid dynamics (CFD) have been applied to predict the details of airflow, contaminant dispersal and thermal transport within isolated zones, yet zone transport processes do not occur in isolation. They result from and interact with the bulk airflows from the larger whole-building systems in which they are embedded. As multi-zone models can reliably predict these bulk airflows, there is a growing interest in embedding detailed CFD models of specific zones within multi-zone models of the enclosing whole-building system to more faithfully account for these interactions and thus the details within the zone(s) of interest. This paper presents an analysis of an embedded CFD model and outlines some of the associated problems. A formulation is presented using unambiguous mathematical definitions of the coupling relations between the governing CFD and multi-zone equations made possible by using a port-plane approach to the multi-zone model. It is thereby shown that while the micro-to-macro coupling is straightforward, macro-to-micro coupling must remain indeterminate due to the fact that multi-zone models do not account for non-normal airflows to zone port-planes or for turbulent characteristics of the airflow that may be used by the embedded CFD model. The formulation of the embedded detail problem considered herein leads to the direct mathematical coupling of the semi-discrete Finite Element form of the CFD models used in the nonlinear algebraic enclosing multi-zone models. To investigate the limitations of embedded detail analysis the approach taken was applied to a hypothetical test problem configured to be sensitive to one obvious shortcoming of multi-zone models - their inability to account for non-normal inflow velocity components - and one less obvious shortcoming of CFD models - their tendency to model flow resistance differently and thus incompatibly with multi-zone models. The results indicated that these two shortcomings alone may critically limit the value of embedded detail analysis. Specifically, it appears that embedded detail analysis can not, in general, be expected to faithfully model flow detail in the CFD embedded zone nor model the larger macroscopic bulk flow structure correctly - even though these objectives may well be realized in special cases. Additional research is clearly needed to identify the special cases when embedded detailed analysis may be expected to be reliable. As such this paper should be of interest to not only the specialist in this emerging field, but to those seeking to employ or to fund research using these methods.