Marine propeller design using artificial neural networks
C.C. Neocleous, Christos N. Schizas · 2003
The present work deals with the task of propeller design using techniques from the field of computational intelligence. An important requirement of the task is to help a designer to reach an acceptably good design in a fast and simple manner in which the most readily available propeller data are used as raw inputs. A neural network system has been developed that can help a naval architectural designer to select a suitable marine propeller that satisfies desired propulsion requirements. Different neural network architectures and learning parameters were tested, aiming at establishing a near optimum setup. To achieve this, a large number of experimental data was used. The end result in the network output is a set of suitable dimensional characteristics and a desired performance.