Main objectives of the present paper are to describe the development of two alternative Simulation Based Design (SBD) environments based on complex CFD analysis, to demonstrate their capability by applying them to a real ship design problem and to give account of the experimental campaign carried out to validate the numerical results and assess the success of the optimization. The fundamental elements of a SBD framework are analyzed and alternative components are described. A derivative-free and a gradient-based local optimizers are illustrated with emphasis on advanced strategies for the use of CPU time expensive CFD solvers in the optimization process, like the variable-fidelity/trust region algorithm and a portable, multi-level parallel framework. Alternatives in the geometry and mesh manipulation techniques, to automatically adapt the volume grid to the evolving shapes, are also illustrated. As analysis tools for the evaluation of objective function and functional constraints, two up-to-date free surface fitting RANS solvers are used. With these elements, two different SBD versions are realized and demonstrated on a real design problem, the optimization of the bow and sonar dome of the DTMB 5415, used as a typical example of a highly complex ship redesign. Realistic functional and geometrical constraints for preventing unfeasible results and to get a final meaningful design are discussed and enforced. As expected, being the two versions based on a local strategy and constructed with different basic components, their use resulted into two different final designs, which however, beyond local differences, clearly display common geometrical trends. Finally, dedicated experimental campaigns for both the optimized models have been carried out to validate the computations and establish the success of the optimization processes. The two optimized models demonstrated improved characteristics beyond the experimental uncertainty, confirming the validity of both the SBD versions.
Comparison and Validation of CFD Based Local Optimization Methods for Surface Combatant Bow
2004
Abstract
Main objectives of the present paper are to describe the development of two alternative Simulation Based Design (SBD) environments based on complex CFD analysis, to demonstrate their capability by applying them to a real ship design problem and to give account of the experimental campaign carried out to validate the numerical results and assess the success of the optimization. The fundamental elements of a SBD framework are analyzed and alternative components are described. A derivative-free and a gradient-based local optimizers are illustrated with emphasis on advanced strategies for the use of CPU time expensive CFD solvers in the optimization process, like the variable-fidelity/trust region algorithm and a portable, multi-level parallel framework. Alternatives in the geometry and mesh manipulation techniques, to automatically adapt the volume grid to the evolving shapes, are also illustrated. As analysis tools for the evaluation of objective function and functional constraints, two up-to-date free surface fitting RANS solvers are used. With these elements, two different SBD versions are realized and demonstrated on a real design problem, the optimization of the bow and sonar dome of the DTMB 5415, used as a typical example of a highly complex ship redesign. Realistic functional and geometrical constraints for preventing unfeasible results and to get a final meaningful design are discussed and enforced. As expected, being the two versions based on a local strategy and constructed with different basic components, their use resulted into two different final designs, which however, beyond local differences, clearly display common geometrical trends. Finally, dedicated experimental campaigns for both the optimized models have been carried out to validate the computations and establish the success of the optimization processes. The two optimized models demonstrated improved characteristics beyond the experimental uncertainty, confirming the validity of both the SBD versions.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
