The sensitivity of the fitness function comprising of weight coefficients assigned to performance variables in a genetic algorithm for meanline design of a transonic compressor is studied. The sum of the weight coefficients is unity. Six performance variables considered are the pressure ratio, efficiency, De-Haller numbers (for rotor and stator), and diffusion factors (for rotor and stator). Based on prior trials, the optimum weight coefficients for pressure ratio and efficiency were considered 0.3 each in the fitness function. Hence the sum of the weight coefficients for the two De-Haller Numbers and two Diffusion Factors considered is 0.4. The values of assigned weights have a significant impact on optimization outcome. Optimized design trials of weight coefficients with higher weightage to DFR resulted in higher efficiency with lower pressure ratio. Optimized design trials with higher weightages to DEHR and DEHS yielded into higher pressure ratio but lower efficiency. The data generated provides a guideline to choose combinations of weight coefficients for fitness functions for several performance requirements of a similar class of compressors for various applications.
Sensitivity Analysis of Weight Coefficients Used in Multiobjective Optimization in Genetic Algorithm Method for Axial Flow Compressor Design
Lect.Notes Mechanical Engineering
2020-08-01
12 pages
Article/Chapter (Book)
Electronic Resource
English
Compressor design optimization , Genetic algorithm , Fitness function , Weighing functions , Meanline design Engineering , Aerospace Technology and Astronautics , Engineering Fluid Dynamics , Engine Technology , Engineering Thermodynamics, Heat and Mass Transfer , Vibration, Dynamical Systems, Control
AIAA-2006-1908 Characterizing the Design Space in Multiobjective Axial Compressor Blade Optimization
British Library Conference Proceedings | 2006
|Multiobjective Optimization of Circumferential Casing Grooves for a Transonic Axial Compressor
Online Contents | 2011
|ENGINEERING NOTES - Speed-Sensitivity Analysis by a Genetic Multiobjective Optimization Technique
Online Contents | 2002
|