화학공학소재연구정보센터
Atomization and Sprays, Vol.21, No.2, 121-126, 2011
PROCESS FOR COUPLING OPTIMIZATION AND PROBABILITY
Often, after a spray has been produced as a method to increase the surface area upon which a reaction can take place, the byproducts of the reaction in the liquid form need to be removed from the gas phase. This paper presents a rapid automated design technique suitable for designing highly efficient spray separators with minimal pressure drop. This numerical approach is especially useful to industrial applications as it utilizes commercially available computational fluid dynamics (CFD) with a MATLAB-based optimization routine. The optimization routine manages the CFD code by importing any number of variables within ranges and/or distributions set by the user. The geometry and conditions presented as an example neglected splashing effects, assuming a dilute spray with highly constrained linear geometry and extremely high gas velocities. As a result, the separator consisted of streamlined airfoils to maximize particle trapping, while maintaining a low-pressure drop. The optimum spacing, chord length, airfoil thickness, and placement were evaluated as design parameters. The results of the optimization increased the number of particles trapped while maintaining a low-pressure drop for 2.5-mu m-size particles. Further, the automated process was used to determine the sensitivity of individual parameters on performance, providing input on where to hold tolerances during manufacturing. In this example, 180 designs were evaluated in 8 h on four processors.