Automatic procedures to design ship hull geometries yielding minimal wave resistance and wave breaking are an attractive opportunity from both the economical and practical standpoints. Estimating the cost function gradient according to sensitivity equations and adjoint equation method instead of using standard finite difference approximations offers margins of cost reduction which this paper aims to assessing. Speed-up factors up to 3.3 have been obtained in the gradient evaluation and of about 1.6 in the full optimization procedure. The main reasons allowing the alternative method to perform better than finite differences are the reduction of flow solutions needed to compute the cost function gradient and the possibility of using the same LU factored matrix both for the flow solver and the SEM or AM equations.
Efficient strategies to design optimal ship hulls
2000
Abstract
Automatic procedures to design ship hull geometries yielding minimal wave resistance and wave breaking are an attractive opportunity from both the economical and practical standpoints. Estimating the cost function gradient according to sensitivity equations and adjoint equation method instead of using standard finite difference approximations offers margins of cost reduction which this paper aims to assessing. Speed-up factors up to 3.3 have been obtained in the gradient evaluation and of about 1.6 in the full optimization procedure. The main reasons allowing the alternative method to perform better than finite differences are the reduction of flow solutions needed to compute the cost function gradient and the possibility of using the same LU factored matrix both for the flow solver and the SEM or AM equations.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.


