A prediction methodology of trimaran form factor (1+k) based on neural network
Year: 2026
Type: article
Abstract: This study proposes a calculation framework of the form factor (1+k) for predicting the total resistance of trimarans and examines its variation with forward speed and outrigger layout. A total of 30 scenarios are analyzed, comprising six outrigger layouts and five forward speeds. Computational Fluid Dynamics (CFD) simulations using OpenFOAM are employed to obtain the total resistance across these configurations. The numerical results are validated against experimental data, demonstrating the accuracy and efficiency of the CFD approach. Additionally, potential flow theory is applied to determine the wave-making resistance, enabling the derivation of the form factor for each scenario. The study systematically analyzes how the form factor varies with trimaran forward speed and outrigger layout. The results reveal distinct trends in the total resistance of trimarans concerning longitudinal distance ( a ), transverse distance ( p ), and forward speed. Furthermore, a neural network is developed to predict the form factor, enabling an efficient estimation of total resistance across various outrigger layouts and forward speeds. The neural network is trained through two strategies using a dataset derived from the numerical simulations. The validation results show that the trained neural network using both strategies can accurately predict the trimaran total resistance, achieving an average error below 5 % and a maximum error not exceeding 10 %. • Development of a framework for determining the form factor of trimarans. • Comparative analysis of the effects of outrigger layout and forward speed on the total resistance of a trimaran. • Development of a neural network-based method for predicting the form factor of trimarans.
Source: Ocean Engineering
Institution: University of Stavanger
Language: English
Subfield: Ocean Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: hybrid
Estimated APC paid: $4,150
Award: G65762986
OpenAlex ID: W7127134267