In this work it has been carried out a regression analysis of the experimental data of displacement catamarans tested at the towing tank of the Dipartimento di Ingegneria Navale di Napoli (DIN). The recorded data have been obtained testing 3 different hull forms in the speed range corresponding to Froude number up to 0.7, combining 6 clearances and 3 displacements, for a total of 42 catamaran configurations. The regression model has been carried out studying the relationship, for assigned Fn values, between a set of dependent variables (CR×103) and a set of independent variables (L/▽1/3, LCB/LCF, b/T, C B, IE, AT/AM, S/L). The data show independent variables strongly correlated, missing values of the residuary resistance at higher velocity for some catamaran configurations and a limited number of observations, so the Ordinary Least Squares (OLS) cannot be applied. The most proper statistical methodology to analyse the data with these characteristics is the Partial Least Squares (PLS) regression. The regression model has been also validated by the PRediction Error Sum of Squares (PRESS).
Resistance of Displacement Catamarans through Regression Analysis
CAMMINATIELLO, Ida;
2008
Abstract
In this work it has been carried out a regression analysis of the experimental data of displacement catamarans tested at the towing tank of the Dipartimento di Ingegneria Navale di Napoli (DIN). The recorded data have been obtained testing 3 different hull forms in the speed range corresponding to Froude number up to 0.7, combining 6 clearances and 3 displacements, for a total of 42 catamaran configurations. The regression model has been carried out studying the relationship, for assigned Fn values, between a set of dependent variables (CR×103) and a set of independent variables (L/▽1/3, LCB/LCF, b/T, C B, IE, AT/AM, S/L). The data show independent variables strongly correlated, missing values of the residuary resistance at higher velocity for some catamaran configurations and a limited number of observations, so the Ordinary Least Squares (OLS) cannot be applied. The most proper statistical methodology to analyse the data with these characteristics is the Partial Least Squares (PLS) regression. The regression model has been also validated by the PRediction Error Sum of Squares (PRESS).I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
