Use this URL to cite or link to this record in EThOS:
Title: Reliability analysis and reliability-based optimisation design of SWATH ships
Author: Pu, Yongchang
Awarding Body: University of Glasgow
Current Institution: University of Glasgow
Date of Award: 1995
Availability of Full Text:
Access from EThOS:
Access from Institution:
The primary objective of the current research presented in this thesis was to develop a rationally-based structural design procedure for SWATH type vessels by applying reliability analysis and reliability-based optimisation techniques. Firstly, the primary loads for PATRIA, a built SWATH ship, were calculated by a theoretical method developed by Chan (1990,1991). The response amplitude operators of loads in regular waves were calculated by a three-dimensional oscillating source distribution method in association with linearised potential theory. A program, SPEC, was developed to carry out spectral analysis and to calculate extreme design values as well as instantaneous pressure distribution in the submerged parts of the vessel. Furthermore several factors, such as load combination and longitudinal side force distribution, which are important in structural analysis, were discussed. Having determined the primary loads acting on PATRIA, a series of finite element analyses was carried out aimed at increasing the understanding of the structural behaviour of the ship, and establishing a simplified model for system reliability analysis and multiple criteria optimisation. At the component level the existing ultimate strength formulations for plate panels and stiffened plates were discussed and calibrated by using a considerable amount of experimental and numerical data. A new algorithm for stiffened plates was proposed. The reliability of plating and stiffened plates was then evaluated by using AFOSM, SORM and Monte Carlo simulation methods to investigate the accuracy of these methods for these types of limit states equations. It is found that Guedes Soares' formulae and Faulkner's method are the best for plate panels and stiffened plates respectively. The results for failure probability from SORM are much better than those from AFOSM. In these cases the AFOSM always underestimates the failure probability. The largest relative errors of failure probability and reliability index reach -45.1% and 7.4% respectively. Considering the nominal nature of reliability index the difference between the two methods is so small that the values obtained from AFOSM are acceptable in practice. At the system level the conventional beta-unzipping method was extended by introducing a discarding process in searching for significant failure modes of the structural system. The extension could save computational time when the combined load effects are considered in the analysis. The method was then used to analyse a typical frame in PATRJA. In the analysis the combined load effect including buckling was considered. It is found that the most critical part is in the haunch area, and all the critical sections in the identified significant failure modes are in the haunch area. Hence it may be said that more attention should be paid to the haunch area. The buckling has a moderate effect on system reliability in this particular case, and should be considered in the analysis. Finally, the reliability-based optimisation techniques were used to achieve an efficient design. Various reliability-based optimisation formulations and their associated problems were first discussed. An algoritlim, in which the component and system reliability indices could be balanced, was proposed. The proposed strategy was then applied to optimise the one-dimensional model for the transverse cross-deck frame in PATRIA. It is found that: The algorithm works very well. Computational time in the analysis is not a problem because the system reliability calculation is only applied to the optimum structure. The original design is quite close to the optimal one, so the margin for optimisation is small. It is of interest to note that the system reliability index for the original structure is only 3.756, while it is 4.712 for the optimum structure, at the same time the optimum one is 13% lighter than the original one. The haunch area is confirmed as the critical part. From the values of design variables of the optimum structure, it is observed that increasing the thickness of the side shell is the most efficient way to improve the safety in this area.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available