Characteristics of welding induced initial deflections in welded aluminum plates
Introduction
During the last decade, high strength aluminum alloys have been increasingly applied for the design and construction of high speed vessels; and over the same time the size of high speed vessels has grown and their operation has moved into increasingly harsher environments. Subsequently, the design and building process to ensure the structural safety of aluminum high speed vessels has become more complex in terms of strength or reliability analysis and fabrication quality control [1].
In terms of ultimate strength analysis for welded aluminum plate structures, fabrication-related initial deflection of plates is one of primary parameters of influence. In contrast to welded steel plate structures where the necessary information is plentiful [2], [3], lack of information on fabrication related initial deflections of aluminum plates can in particular make the design and building process for aluminum vessels relatively more uncertain [4], [5].
The aim of the present paper is to obtain the statistical database of welded aluminum plate initial deflections that can occur during welding fabrication of aluminum stiffened plate structures for marine applications.
The studies related to initial deflection surveys for welded aluminum structures have mostly been performed in conjunction with mechanical collapse test programs. In the early 1980s, a series of collapse tests were carried out on aluminum unstiffened plates made of 5083 and 6082 alloys, by Mofflin [6] and Mofflin and Dwight [7] at the University of Cambridge, UK; and these are regarded as perhaps one of the largest and most relevant test programs for the collapse strength of aluminum plating until now. After TIG (tungsten inert gas) welding in the longitudinal direction and MIG (metal inert gas) welding in the transverse direction, weld-induced initial deflections were measured.
In the late 1980s, Clarke and Swan [8] and Clarke [9] at the Admiralty Research Establishment (ARE), UK carried out the buckling collapse testing on a total of five aluminum stiffened plate structures made of 5083 alloy. This was one of the earliest collapse test programs to use ship-shaped aluminum stiffened plate structures using full scale prototype models of all-welded construction with multiple frame bays.
Over a decade after the ARE tests, several collapse test programs on aluminum stiffened plate structures constructed by welding were carried out together with various surveys of weld induced initial imperfections [10], [11], [12], [13], [14], [15], [16], [17]. The material of most test structures was 5083 aluminum alloy for plating and 6082 aluminum alloy for stiffeners. Except perhaps for those by Tanaka and Matsuoka [12] and Matsuoka et al. [13], which were full scale prototype models with multiple frame bays, most of these test structures were small scale models with single bay.
Through the present study, it is hoped that some more database is developed on fabrication-related initial deflections of welded aluminum plates which can be used for reliability analysis or code calibrations in terms of ultimate limit state assessment of welded aluminum plate structures. It is evident that there are various types of initial imperfections in welded aluminum stiffened plate structures such as initial distortions of plating and stiffeners, residual stresses and softening in the heat-affected zone [18], but the present paper is focused on the initial deflections of plating between stiffeners.
A total of 78 ship-shaped full scale prototype aluminum structures which are equivalent to sub-structures of a 80 m long all aluminum high speed vessel are constructed by MIG (metal inert gas) welding. The material of plating and stiffeners is varied among 5083-H116 (rolled), 5383-H116 (rolled), 5383-H112 (extruded) and 6082-T6 (extruded) aluminum alloys which are today the most popular for marine applications.
The statistics of weld-induced initial deflections in plating between stiffeners are obtained by direct measurements of the prototype structures. A statistical analysis of measured database is performed to determine mean and coefficient of variation (COV) of the plate initial deflection. Three (slight, average and severe) levels of the plate initial deflection are then obtained.
Section snippets
Design and construction of full scale prototype structures
A total of 78 prototype aluminum structures which are full scale equivalent to sub-structures of an 80 m long all aluminum high speed vessel are considered. They are designed in terms of single and multi-bay stiffened plate structures as those shown in Fig. 1. While various methods for fabricating aluminum ship structures are today relevant, the present test program adopts the MIG welding technique, which is now one of the most popular methods of welding in aluminum ship construction.
Table 1
Measurements of weld-induced initial deflections
The maximum initial deflections of plates between stiffeners are now measured for the prototype structures. Fig. 3(a) shows a schematic of plate initial deflections in the longitudinal (long) direction which has a multi-wave shape, and Fig. 3(b) shows a schematic of plate initial deflections in the transverse (short) direction which has a half wave shape.
Some recognized classification societies [19] suggest the maximum initial deflection of welded plates for the purpose of structural design and
Statistical analysis of initial deflection measurements
The statistical analysis of the extensive initial deflection measurement data obtained is now performed. It is certain that the statistical characteristics of initial deflections can not be reflected by the normal distribution function. After a confirmation that the use of the Weibull probability density function is suitable together with its flexibility in fitting varied sets of data, therefore, mean value and standard deviation of random initial imperfection parameters was calculated by the
Concluding remarks
During the last decade, the application of aluminum alloys to marine structures such as high-speed vessels and littoral surface crafts has been rapidly increasing. To operate in increasingly harsher environments, the size of high-speed vessels has also grown. Subsequently, the structural design and building process to ensure the structural safety has become more complex in terms of limit state strength assessment and fabrication quality control among others.
During welding fabrication of
Acknowledgments
The present study was undertaken at the Ship and Offshore Structural Mechanics Laboratory (SSML), Pusan National University, Korea, which is a National Research Laboratory funded by the Korea Ministry of Science and Technology (Grant no. ROA-2006-000-10239-0). The authors are pleased to acknowledge the support of Alcan Marine, France, Ship Structure Committee, USA, and Hanjin Heavy Industries & Construction Company, Korea.
References (19)
- et al.
Ultimate strength of aluminum alloy outstands in compression: experiments and simplified analyses
Thin-Walled Structures
(1999) - et al.
Ultimate strength of stiffened aluminum panels with predominantly torsional failure modes
Thin-Walled Struct
(2001) - et al.
Stiffened aluminum panels subjected to axial compression
Thin-Walled Struct
(2001) - DNV, Rules for ships/high speed, light craft and naval surface craft. Det Norske Veritas, Oslo, Norway;...
- et al.
Ultimate limit state design of steel-plated structures
(2003) - et al.
Ship-shaped offshore installations: Design, building, and operation
(2007) Aluminum alloy structures
(1985)Aluminum structures—Advances, design and construction
(1987)- Mofflin DS. Plate buckling in steel and aluminum. PhD thesis, University of Cambridge, UK;...
Cited by (20)
Experimental and numerical study on the penetration of the inclined stiffened plate
2022, Ocean EngineeringStatistical analysis of initial deflection of aluminium plating between stiffeners
2021, Thin-Walled StructuresCitation Excerpt :For all panels, the probability distribution of the maximum deflection is shown in Fig. 8. In Fig. 9, a comparison is made between the measured and estimated values proposed by Paik [10] and Paik et al. [11]. According to this comparison, it can be concluded that the measured initial deflections in this paper is usually between “average” and “severe” levels, as suggested by Paik et al. [10,11].