저속선의 수학 모형 식별을 위한 최적 입력의 설계

Abstract

In this study, identifiability of low-speed manoeuvring mathematical models presented by Takashina (1986), Yoshimura (1988), and Yoon & Kim (2005) was investigated and characteristics of each model were analyzed. Identifiability of each model was inspected by assessing the correlation matrix among hydrodynamic parameters of each model through numerical simulations of 35 degree port turning and 20/20 zigzag test. Although the correlation values are very high in all the three mathematical models, the model presented by Yoon and Kim was found to have better identifiability. Since the simulation data of sea trials in low-speed situation were found not to be sufficient to estimate all the hydrodynamic coefficients, designing an input scenario to improve the identifiability of mathematical models was required. Based on the idea that ships in harbor or channels can experience large drift motion such as tug boat pulling, excitation by side thrusters was considered through numerical simulation. Side thruster scenarios were designed by comparing A and D-optimality criteria through exhaustive search. With the designed scenario, it was found that excitation by side thrusters can improve the identifiability of mathematical models for low-speed manoeuvring considerably.