Neural Network Prediction of an Optimum Ship Screw Propeller (CROSBI ID 549194)
Prilog sa skupa u zborniku | izvorni znanstveni rad | međunarodna recenzija
Podaci o odgovornosti
Matulja, Dunja ; Dejhalla, Roko
engleski
Neural Network Prediction of an Optimum Ship Screw Propeller
A neural network (NN) has been trained to enable the choice of an optimum ship screw propeller geometry. The idea was to avoid the mistakes that can occur in the use of open water diagrams, as well as to save time required to run various computer programs. The network provides the optimum diameter, pitch ratio and thrust for a given case of delivered power, propeller revolution, advance velocity, blade number and expanded area ratio, within the treated design range. The necessary training data were obtained using the polynomials representing the Wageningen B-screw series. The NN was used as a black box, and the data were approximated to a multiple variable function. After the training, the NN was tested with satisfactory results, especially for diameter and thrust. In addition to that, the processing speed is not affected by the number of cases to compute. This makes the neural network a reliable tool, fit to predict the principal geometric features of an optimum propeller.
ship screw propeller; optimization; neural network
nije evidentirano
nije evidentirano
nije evidentirano
nije evidentirano
nije evidentirano
nije evidentirano
Podaci o prilogu
829-830.
2008.
objavljeno
Podaci o matičnoj publikaciji
Annals of DAAAM for 2008 & Proceedings of the 19th International DAAAM Symposium
Katalinić, Branko
Beč: DAAAM International Vienna
978-3-901509-68-1
1726-9679
Podaci o skupu
19th International DAAAM Symposium
predavanje
22.10.2008-25.10.2008
Trnava, Slovačka