In this paper the problem of the optimization of heat transfer enhancing surfaces is faced. The surfaces are parameterized usingBezier curves and solved with a finite volumes CFD code. The optimization is based on three different algorithms used in cascade:MOGT, MOGA-II and simplex for non-linear optimization. The Reynolds number is based on the average velocity and hydraulicdiameter of the channels and is set to 1000, the Prandtl number is 0:744. The same optimization process is carried out twice usingtwo different turbulence models: k-e RNG with enhanced wall treatment and k-w SST with low Reynolds number correction to theturbulent viscosity. It is found that the differences in the optimum channel shapes are significant and even that the same channelcan give quite different results being solved with a turbulence model or another. Though, from the results of the two optimizationprocesses, the rules of thumb we can infer on which elements a channel needs in order to have a certain performance, are almostthe same.
Turbulence Models comparison in heat exchanger enhanced surfaces optimization
CAVAZZUTI, Marco;CORTICELLI, Mauro Alessandro
;BAROZZI, Giovanni Sebastiano
2008
Abstract
In this paper the problem of the optimization of heat transfer enhancing surfaces is faced. The surfaces are parameterized usingBezier curves and solved with a finite volumes CFD code. The optimization is based on three different algorithms used in cascade:MOGT, MOGA-II and simplex for non-linear optimization. The Reynolds number is based on the average velocity and hydraulicdiameter of the channels and is set to 1000, the Prandtl number is 0:744. The same optimization process is carried out twice usingtwo different turbulence models: k-e RNG with enhanced wall treatment and k-w SST with low Reynolds number correction to theturbulent viscosity. It is found that the differences in the optimum channel shapes are significant and even that the same channelcan give quite different results being solved with a turbulence model or another. Though, from the results of the two optimizationprocesses, the rules of thumb we can infer on which elements a channel needs in order to have a certain performance, are almostthe same.I documenti in SFERA sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.