Wake distribution prediction on the propeller plane in ship design using artificial intelligence

Abstract

In this study, a new neuro-fuzzy technique is applied to estimate the wake field distribution on propeller plane of ship. The wake distribution data of stern flow fields have been collected systematically by model tests of ship. When a correlation between geometrical hull information and wake distribution of ship is grasped through the collected data, the obtained correlation can be helpful in designing similar kinds of ships. Characteristics of the wake distribution by the change of stern hull form can be figured out easily by using a neuro-fuzzy system with the collected data. The neuro-fuzzy technique is applied to the ship designing, which is well known as one of the prospective methods for prediction, classification and diagnosis of real complicated problems. To prove the proposed method, three-dimensional stern hull forms and nominal wake values from a model test are structured as processing elements of input and output layers, respectively. In order to embody three-dimensional hull forms in a neuro-fuzzy system at the same time to avoid excessive increase in the input data, angle values of certain points of the hull form are used to avoid excessive increase in the input data and simultaneously in order to embody three-dimensional hull form in a neuro-fuzzy system. This method proved satisfactory in obtaining safe predictive results. The results from this study may give advanced information to hull form designers for evaluation and optimization of stern hull form by predicting a ship wake distribution at the initial designing stage.