Prediction of main particulars of a chemical tanker at preliminary ship design using artificial neural network

Abstract

Preliminary ship design is an important part of the ship design and a reliable design tool is needed for this stage. The aim of this study was to develop an artificial neural network (ANN) model to predict main particulars of a chemical tanker at preliminary design stage. Deadweight and vessel speed were used as the input layer; and length overall, length between perpendiculars, breadth, draught and freeboard were used as the output layer. The back-propagation learning algorithm with two different variants was used in the network. After training the ANN, the average of mean absolute percentage error value was obtained 4.552%. It is also observed that the correlation coefficients obtained were 0.99921, 0.99775, 0.99537 and 0.9984 for training, validation, test and all data-sets, respectively. The results showed that initial main particulars of chemical tankers are determined within high accuracy levels as compared to the sample ship data.