【正文】 dels. However, tests carried out in a model should be made with dynamic and thermal similarity if they are to be directly applied to the full scale. This normally requires the equality of the Reynolds number, Re, and the Archimedes number, Ar, [5, 6] which is not possible to achieve in the model concurrently. The other problem which is often encountered in air distribution design is the interference to the jet from rough surfaces and surfacemounted obstacles such as structural beams, light fittings etc. Previous studies [7, 8] have shown that surfacemounted obstacles cause a faster decay of the jet velocity and when the distance of an obstacle from the air supply is less than a certain value called “the critical distance”, a deflection of the jet into the occupied zone takes place. This phenomenon renders the air distribution in the room ineffective in removing the heat load and, as a result, the thermal fort in the occupied zone deteriorates. Here again there is a scarcity of design data, particularly for nonisothermal air jets. Air distribution problems, such as those discussed here, are most suitable for numerical solutions which, by their nature, are good design optimization tools. Since most air distribution methods are unique to a particular building a rule of thumb approach is not often a good design practice. For this reason, a mockup evaluation has so far been the safest design procedure. Therefore, numerical solutions are most suitable for air distribution system design as results can b readily obtained and modifications can be made as required within a short space of time. Because of the plexity of the air flow and heat transfer processes in a room, the numerical solutions to these flow problems use iterative procedures that require large puting time and memory. Therefore, rigorous validation of these solutions is needed before they can be applied to wide ranging air distribution problems. In this paper a review is given of published work on numerical solutions as applied to room ventilation. The finite volume solution procedure which has been widely used in the past is briefly described and the equations used in the kε turbulence model are presented. Numerical solutions are given for two and threedimensional flows and, where possible, parison is made with experimental data. The boundary conditions used in these solutions are also described. 如何寫論文的展開部分 (Approach), 結果和討論 (Results and Discussion) 材料和 方法部分 對于以實驗為主的研究論文，該部分往往位于論文展開部分的前面。 對于實驗，描述應盡可能詳細。詳細的程度應使別的研究者可以重復你的實驗，對難以重復的實驗可評價你的實驗。 這一部分經(jīng)常采用小標題，如： subjects, apparatus, experimental design, and chemical synthesis。 在這一部分，你應當說明： (1) 你所用的材料和化學藥品的名稱； (2)實驗條件； (3)實驗儀器； (4)實驗方法和步驟。 9 原理和理論模型部分 對于理論分析和數(shù)值計算為 主的研究論文，該部分往往位于論文展開部分的前面。 一般首先用數(shù)學方法描述所討論的問題，如列出控制方程、邊界條件和初始條件。為簡化問題并突出問題本質(zhì)，常需對問題進行合理假設。這部分會引入一些方程、格式、邊界條件和初始條件，下面通過一些例子說明其經(jīng)常采用的表達方式。 例 1[5] DEVELOPMENT OF MODEL The model assumes that VOCs are emitted out of a single uniform layer of material slab with VOCimpermeable backing material, and a schematic of the idealized building material slab placed in atmosphere is shown in . The governing equation describing the transient diffusion through the slab is 22 ),(),( x txCDt txC ????? （ 1） where C(x,t) is the concentration of the contaminant in the building material slab, t is time, and x is the linear distance. For given contaminant, the mass diffusion coefficient D is assumed to be constant. The initial condition assumes that the pound of interest is uniformly distributed throughout the building material slab, ., 0C ( x , t ) C fo r 0 x L? ? ?, t=0 （ 2） where L is the thickness of the slab, and C0 is the initial contaminant concentration. Since the slab is resting on a VOCimpermeable surface, the boundary condition of the lower surface of the slab is 0,00),( , ????? xtt txC (3) A third boundary condition is imposed on the upper surface of the slab () .,0))()((),( , LxttCtChx txCD sm ??????? ? (4) where hm is the convective mass transfer coefficient, m/s； Cs(t) is the concentration of VOC in the air adjacent to the interface。 mg m3。 C?(t) is the VOC concentration in atmosphere, mg m3. It should be mentioned that almost all the physically based models in the literature assumed Cs(t)= C?(t), . implied that hm is infinite, (Dunn, 1987。 Clausen et al., 1991。 Little et al., 1994). Obviously, the case assumed is a special case of equation (4). Besides, equilibrium exists between the contaminant concentrations in the surface layer of the slab and the ambient air, or (Little et al., 1994) .Lx,0t)t(KC)t,x(C ,s ??? (5) where K is the socalled partition coefficient. C(L,t) Cs(t) L C?(t) x C(x,t) building material air interface )(tm?10 Schematic shown of a building material slab in atmosphere. The solutions to equations (1)(5) derived by us are as follows (6) where ,KDhH m? β m (m=1,2,… ) are the positive roots of (7) Equation (6) gives the contaminant concentration in the building material slab as a function of distance from the base of the slab, and also of time. Thus, VOC emission rate per unit area at instant t )t(m? and VOC mass emitted from per unit area of the building material slab before instant t m(t) can be respectively expressed as follows ????? ??? ????????? 1 22 222 )( )(2)(s i n),()( m m mmLx HHL HLDx txCDtm ? ?? ? ????? ??? t tDtD KdCeeKCC mm 0 )(0 )]())0([( 22 ???? (8) ? ?? ??? ??? ???????? t m m mmt Lx HHL HLDdtx txCDtm 0 1 22 2220 )( )(2)(s i n),()( ?