Analysis of ship collision damage by combining Monte Carlo simulation and the artificial neural network approach

作者：Sun, Liping; Zhang, Qi; Ma, Gang^*; Zhang, Tianyue

来源：Ships and Offshore Structures, 2017, 12(sup1): S21-S30.

DOI：10.1080/17445302.2016.1258759

• Ship collision
• damage assessment
• ANN
• MCS

Ship collision is a highly hazardous accident with potentially serious consequences. Thus, evaluating the ship collision risk with quantitative analysis of damage, which is characterised by uncertainties, is necessary. Monte Carlo simulation (MCS) is a flexible and efficient method for this purpose. However, large computing resources are required for MCS-based risk assessment, especially when numerous random variables are involved. In addition, ship collision damage models are complicated and are usually resolved by the finite element method (FEM), which is computationally expensive. To address this problem, a coupled method is proposed in this paper: the artificial neural network (ANN) model was combined with MCS to assess collision damage quantitatively. A dynamic simulation in the time domain was used to predict the structural response resulting from a ship collision. Then the ANN was trained based on the collision data and used as a substitute for the iterative FEM runs within the MCS procedure. A benchmark collision case is used to verify the practicability of the proposed method, and a qualitative uncertainty analysis is also conducted in this quantitative damage assessment.

• 出版日期2017
• 单位哈尔滨工程大学