An object-oriented MATLAB toolbox for automotive body conceptual design using distributed parallel optimization. (April 2017)

• Record Type: Journal Article
• Authors: Qin, Huan; Liu, Zijian; Liu, Yu; Zhong, Haolong
• Is Part Of: Advances in engineering software. Volume 106(2017)
• Page Start: 19
• Page End: 32
• Publication Date: 2017-04
• Subjects: BIW structure -- Conceptual design -- Transfer stiffness matrix method -- Size optimization -- Object-oriented MATLAB toolbox -- Distributed parallel computing; Computer-aided engineering -- Periodicals; Engineering -- Computer programs -- Periodicals; Engineering -- Software -- Periodicals; Periodicals; 620.0028553
• DOI: 10.1016/j.advengsoft.2017.01.003
• Languages: English
• ISSNs: 0965-9978

Abstract

Highlights: An exact transfer stiffness matrix method is proposed for both static and dynamic analyses of three-dimensional semi-rigid framed structures. The matrix storage and equation solution techniques for large-scale engineering structures are considered. The size optimization mathematical model with static and dynamic constraints for BIW lightweight design is formulated and solved by genetic algorithm. Distributed parallel computing technique is adopted to speed up the optimization up to 5.46 times with 8 cores in two desktops. An object-oriented MATLAB toolbox is developed to promote the conceptual design of BIW structure. Abstract: Appropriate structural analysis and optimization methods are of great significance for the conceptual design of automotive body-in-white (BIW) structure. This paper simplifies BIW structure as a spatial semi-rigid framed structure to provide early-stage predictions. Then a novel exact transfer stiffness matrix method (TSMM) is proposed for both static and dynamic analyses of three-dimensional semi-rigid framed structures. The matrix storage and equation solution techniques for large-scale engineering structures are also considered. Additionally, a size optimization mathematical model for BIW conceptual structure is formulated and solved by genetic algorithm (GA). Afterwards, to promote the conceptual design of BIW structure, an object-oriented MATLAB toolbox, based on TSMM, is developed. The Unified Modeling Language (UML) and strategy design pattern are employed to facilitate the development of the toolbox. Distributed parallel computing technique is adopted to speed up the former sequential optimization algorithm with simple modifications. Lastly, the validity of this easy-used toolbox is demonstrated by a benchmarking auto-body.