Design optimization of a shell-and-tube heat exchanger with disc-and-doughnut baffles for aero-engine using one hybrid method of NSGA II and MOPSO

Abstract

A performance prediction model is built utilizing Slipcevie method and the experimental verification results show that the average errors of Q, ΔPt, ΔPs are 4.4%, 5.3%, 6.0%, respectively. Through parametric study of Ddo_i, Ddb, Nb by Sobol’ method based on the model, it could be found that the parameters’ effects are unlinear and conflicted in most cases, and whether the interaction between factors is considered or not, the influences of Ddb, Ddo_i, and Nb on Q/ΔPt, Q/ΔPs, and G are the strongest, respectively. One configuration optimization method by NSGA II-MOPSO is proposed to increase Q/ΔPt, Q/ΔPs and decrease G, simultaneously, and the comparison results between four chosen optimal solutions and the original indicate that Q/ΔPt increases by 1.50% on average, and Q/ΔPs increases by 32.35% on average, while G averagely decreases by 29.18%. Besides, optimization comparison among NSGA II-MOPSO, NSGA II, and MOPSO is conducted and it could be concluded that NSGA II-MOPSO is the best, followed by MOPSO and then NSGA II, which illustrates that this method is feasible and valid, and could provide significant guidance for design.

Keywords

• Performance prediction
• Parametric study
• Configuration optimization
• Slipcevie
• Sobol’
• NSGA II-MOPSO