Discussions on the ductile fracture prediction of ship structures under impact loads

2021 ◽  
pp. 236-244
Author(s):  
Y. Lu ◽  
K. Liu ◽  
Z. Wang ◽  
W. Tang ◽  
J. Amdahl
2017 ◽  
Vol 17 (2) ◽  
pp. 321-329 ◽  
Author(s):  
Lei Mu ◽  
Yuan Wang ◽  
Yong Zang ◽  
Pedro Malaquias Araujo Stemler

Author(s):  
Anne M. Fullerton ◽  
Thomas C. Fu ◽  
David E. Hess

Navy fleet problems with damage to hatches and other appendages after operation in high sea states suggest that wave impact loads may be greater than the current design guidelines of 1000 pounds per square foot (48 kilopascal) (Ship Specification Section 100, General Requirements for Hull Structure and Guidance Manual for Temporary Alterations, NAVSEA S9070-AA-MME-010/SSN, SSBN). These large impact forces not only cause damage to ships and ship structures, they can also endanger the ship’s crew. To design robust marine structures, accurate estimates of all encountered loads are necessary, including the wave impact forces, which are complex and involve wave breaking, making them difficult to estimate numerically. An experiment to investigate wave impact loads was performed at the Naval Surface Warfare Center, Carderock Division in 2005. During this experiment, the horizontal and vertical loads of regular, non-breaking waves on a 12 inch (0.305 m) square plate and a 19.75 inch (0.5 m) diameter horizontal cylinder were measured while varying incident wave height, wavelength, wave steepness, plate angle and immersion level of the plate and cylinder. Wave heights of up to 1.5 feet (0.46 m) were tested, with wavelenghs of up to 30 feet (9.1 m). In all cases, the horizontal wave impact force increased with wave steepness. For some angles, the horizontal wave impact force increased with greater submergence. A feed-forward neural network (FFNN) developed by Applied Simulation Technologies was used to predict the horizontal forces measured during the experiment based on the values of wave height, wavelength, wave steepness, plate angle and immersion level of the plate and cyclinder. A FFNN is a computational method used to develop nonlinear equation systems that use input variables to predict output variables. Predictions of forces from the FFNN compare well with the experimental data, and may be useful in future design of ships and ship structures.


Sign in / Sign up

Export Citation Format

Share Document