Loading...

Blade shape optimization of marine propeller via genetic algorithm for efficiency improvement

Taheri, R ; Sharif University of Technology | 2012

642 Viewed
  1. Type of Document: Article
  2. DOI: 10.1115/GT2012-68183
  3. Publisher: 2012
  4. Abstract:
  5. In this paper, a numerical optimization method has been carried out to optimize the shape and efficiency of a propeller. For analysis of the hydrodynamic performance parameters, an extended vortex lattice method was used by implementing an open-source code which is called OpenProp. The method of optimization is a non-gradient based algorithm. After a trade-off between a few gradient-based and non-gradient based algorithms, it is found that the problem of being trapped in local optimum solutions can be easily solved by choosing nongradient based ones. Hence, modified Genetic algorithm is used to implement the so-called hydrodynamic performance analyzer code. The objective function is to maximize efficiency by considering the design variables as non-dimensional blade's chord and thickness distribution along the blade. For initial guess data of the DTRC 4119 propeller which are radially distributed along the blade is used. The hydrodynamic performance analyzer code is modified by a higher order QuasiNewton scheme. Also hybrid function is used to accurate the convergence. Finally, parallel processing implementation on the codes has been done successfully. To improve the computation speed, the algorithm is improved to be extended on a parallel processing system. The process of parallelizing has been done simplicity by Matlab M-code and the number of cores has been chosen as 4. The final results verify both fast convergence in comparison with common methods and nearly 10% improvement in propeller efficiency (mechanical efficiency of the system) which is significant for these kinds of problems. Therefore, the algorithm starts with geometry arrived at by other researchers and improves it to a more efficient propeller
  6. Keywords:
  7. Efficiency improvement ; Hydrodynamic performance ; Local optimum solution ; Mechanical efficiency ; Modified genetic algorithms ; Numerical optimizations ; Thickness distributions ; Vortex lattice method ; Algorithms ; Efficiency ; Exhibitions ; Fluid dynamics ; Optimization ; Parallel architectures ; Parallel processing systems ; Gas turbines
  8. Source: Proceedings of the ASME Turbo Expo ; Volume 5 , 2012 , Pages 235-242 ; 9780791844717 (ISBN)
  9. URL: http://proceedings.asmedigitalcollection.asme.org/proceeding.aspx?articleid=1694586