![Institutional Repository [SANDBOX]](images/dias.logo.png "Institutional Repository [SANDBOX]"){kind=logo}
Institutional Repository [SANDBOX]
Technical University of Crete
Technical University of Crete
EN
|
EL
| URI: | http://purl.tuc.gr/dl/dias/9EEEE118-645C-4A06-B5FE-C0E3764D99DD | ||
| Year | 2007 | ||
| Type of Item | Peer-Reviewed Journal Publication | ||
| License |
|
||
| Bibliographic Citation | I. M. Valakos, M. S. Ntipteni, I. K. Nikolos, "Structural Optimization of a Centrifugal Impeller Using Differential Evolution in CATIATM Environment", Operational Research an International Journal, vol. 7, no. 2, pp. 185-211, May 2007. DOI: 10.1007/BF02942387 https://doi.org/10.1007/BF02942387 | ||
| Appears in Collections |
A Differential Evolution (DE) algorithm is used to optimize the backface geometry of a centrifugal impeller, with respect to the calculated maximum stress, in order to extend its overspeed limits. A detailed fully parametric 3D model of the impeller was initially constructed using CATIA V5; the backface geometry is defined using a Bezier curve with its parameters used as design variables for the present optimization procedure. The stress analysis is performed using the Generative Structural Analysis workbench of CATIA. Two different versions of the same DE algorithm are utilized in this work. The first one was developed in order to cooperate with different analysis software, in the form of executables or batch files, which are automatically called to evaluate each candidate solution; for the problem at hand, CATIA software is used to analyze each solution in a batch optimization procedure. The values of the independent design variables of each solution are provided to CATIA by using specific macro commands in batch mode, thus automatically updating the geometry, along with the corresponding mesh and the following stress calculation, in an automated manner. As an alternative approach to the same optimization problem, a DE plug-in, fully compatible with CATIA has been developed, utilizing exposed objects of CATIA (open CAA V5 automation architecture), which provide all the necessary properties and methods to interact directly with a part or analysis document through the VBA programming language. This plug-in also features a friendly and ‘easy to use’ graphical interface, which enables the user to manipulate the part’s design and analysis parameters, as well as the objective function, in order to specify the problem as suited. Optimization results are presented and compared with the results provided by the Simulated Annealing (SA) optimizer embedded in CATIA. The advantages of the proposed procedures are discussed with respect to the alternative approach.
Copyright © DIAS 2013
