Two procedures are described for solving the Navier-Stokes equations for steady, fully three-dimensional flows: both are extensions of earlier methods devised for three-dimensional boundary layers, and have the following common features: (i) the main dependent variables are the velocities and pressure;(ii) the latter are computed on a number of staggered, interlacing grids, each of which is associated with a particular variable;(iii) a hybrid central-upwind difference scheme is employed; and (iv) the solution algorithms are sufficiently implicit to obviate the need to approach the steady state via the time evolution of the flow, as is required by wholly explicit methods.The procedures differ in their manner'of solving the difference equations. The SIVA (for Simultaneous Variable Adjustment) procedure, which is fully-implicit, uses a combination of algebraic elimination and point-successive substitution, wherein simultaneous adjustments are made to a point pressure, and the six surrounding velocities, such that the equations for mass and (linearised) momentum are locally satisfied.The SIMPLE (for Semi-Implicit Method for Pressure-Linked Equations) method proceeds in a successive guess-and-correct fashion.Each cycle of iteration entails firstly the calculation of an intermediate velocity field which satisfies the linearised momentum equations for a guessed pressure distribution: then the mass conservation principle is invoked to adjust the velocities and pressures, such that all of the equations are in balance.By way of an illustration of the capabilities of the methods, results are given of the calculation of the flow of wind around a building, and the simultaneous dispersal of the effluent from a chimney located upstream.

No abstract

No abstract

161°C, 46.5 mmol of Co,(CO)d., P-to-Co = 2.5, Pco = PH* = 32-35 atm. The absorption rates of synthesis gas were measured under carefully controlled gas-liquid interfacial area, thus operating the autoclave as a stirred (or Lewis) cell whereby only the solution was mixed without disturbing the interfacial area (for the theory of gas-liquid absorption combined with chemical reaction, see Danckwertz, 1970).It was found that the enhancement factor F was approximately 1 at 120' and 2.23 mmolh. of carbonyl while a value of 17 was measured at 161'C and 46.5 mmolh. of carbonyl. An F value of 1 indicates a very slow chemical reaction in the solution compared to the gas-liquid mass transfer rate. The CO concentration in the solution is thus equal, or almost equal, to the saturation value. This was confirmed by the ir analysis and further by the observation that the overall octene conversion rate was independent of the impeller speed over 1250 rpm. Thus under the prevailing conditions, the degree of PBu3 liganding is only determined by the thermodynamics of the carbonyl-PBu3 system. On the other hand, a value of 17 for F, found at 161'C, Literature CitedThe reduction of NO with NH3 on alumina-supported vanadium oxide and iron oxide-chromium oxide catalysts has been studied in flow reactors between 200 and 515'C. The rate of reduction of NO on both catalysts reaches a maximum at about 400'C and is considerably enhanced by the presence of 02. With 1000 ppm of NO and either stoichiometric or excess "3, acceleration of NO reduction is a strong function of 0 2 concentration up to 1 % 02. In simulated flue gas containing 14% C02, 5 % H20, 3% 0 2 and with NO varying from 250 to 1000 ppm, reduction of NO is a function of NH3/NO ratio from 0.5 to 1.4. The presence of SO2 in the gas mixture does not appreciably affect the reduction of NO on either catalyst at 400'C.

No abstract