Educating ANNs means the weighting is over In-depth | LIGHTSHIP WEIGHT


George Sachinoglou, a post graduate student at the School of Marine Science & Technology of Newcastle University, uses artificial neural networks (ANNs) to estimate the lightship mass of container carriers.


that can be carried, so that the displacement is a balanced sum of the payload and the lightship, [1]. Improving the accuracy and efficiency of predicting the lightship weight has a significant impact on structural design as well, [2]. Te context of this study involves using


P


ANNs as a different approach to learn the relationship between principle features of containerships and their lightship weight as a means to provide a useful design tool with respect to lightship weight estimation at the preliminary design stage seeking more accurate results once more information is available. In order to apply the above mentioned,


a massive network interconnecting input variables, being the parameters affecting the overall quantity of lightship weight, and output variable, being the known lightship weight, was constructed. Tis network has the ability to be adaptive, in terms of creating and editing the strength on its own connections (synapses), which in the end appear in the form of coefficients appointed to each of the selected variables.


Data Sources Data selection can be a demanding and intricate task, since a neural network’s performance is directly attached to the data used to train it. Te gathered data, which is used in this


project for testing and training purposes, has been extracted from several issues of Significant Ships, published by RINA, [3]. A range of issues from 1995 to 2007 were available to the author and resulted in 56 container ships, ranging from 94.27m to 334m with corresponding capacities from 478TEU up to 10,050TEU. For the selection of the variables’ quality, from this widely accepted source, a number of discussions involving experienced individuals such as Prof. Ian Buxton and Dr Bryan Barrass of the


22


redicting the lightship weight of a vessel is vital during the design process as it specifies the required deadweight


Fig. 1 General structure of an Artificial Neural Network.


School of Marine Science and Technology at Newcastle University took place and valuable advice was provided on the reliability of this source data.


Selection of Variables In order to develop a model to predict the lightship weight for container ship designs, the breakdown of the lightship weight into machinery, outfit and steel weight is critical and defining. For obtaining a successful input-output mapping, it is best to examine the variables that influence each category separately. However, the three weight sub-groups of steel, outfit and machinery were not given explicitly in the Significant Ships issues and only an explicit value of the total lightship weight was given; thence, they could not be determined separately in this study. Also, understanding the data’s physical meaning and relation is vital for the user for


evaluation of the results, since networks do not provide any physical sense of its findings in relationships. Based on existing knowledge, the selected sixteen input variables available from the issues of Significant Ships are: length between perpendiculars, Lbp , moulded, Bmld design, Td


, draught scantling, Ts


, block coefficient, Cb bulkheads, BHDtransverse


breadth , draught


, depth moulded to main deck, Dmld , number of transverse , number of decks,


proportion in (%) of high tensile steel, HTS% service speed, Vs


, in knots, payload capacity,


in TEUs, number of officers, number of crew members, number of single rooms, SR, number of double rooms, DR, extended cargo handling gear capacity, (CHG)capacity


, in tonnes . . Te


single output variable is the lightship weight, Wlightship


ANN Architecture A typical architecture (structure) of a feed-forward ANN model can be


The Naval Architect November 2009


Page 1  |  Page 2  |  Page 3  |  Page 4  |  Page 5  |  Page 6  |  Page 7  |  Page 8  |  Page 9  |  Page 10  |  Page 11  |  Page 12  |  Page 13  |  Page 14  |  Page 15  |  Page 16  |  Page 17  |  Page 18  |  Page 19  |  Page 20  |  Page 21  |  Page 22  |  Page 23  |  Page 24  |  Page 25  |  Page 26  |  Page 27  |  Page 28  |  Page 29  |  Page 30  |  Page 31  |  Page 32  |  Page 33  |  Page 34  |  Page 35  |  Page 36  |  Page 37  |  Page 38  |  Page 39  |  Page 40  |  Page 41  |  Page 42  |  Page 43  |  Page 44  |  Page 45  |  Page 46  |  Page 47  |  Page 48  |  Page 49  |  Page 50  |  Page 51  |  Page 52  |  Page 53  |  Page 54  |  Page 55  |  Page 56  |  Page 57  |  Page 58  |  Page 59  |  Page 60  |  Page 61  |  Page 62  |  Page 63  |  Page 64  |  Page 65  |  Page 66  |  Page 67  |  Page 68