Computational & Technology Resources
an online resource for computational,
engineering & technology publications
PROCEEDINGS OF THE TENTH INTERNATIONAL CONFERENCE ON COMPUTATIONAL STRUCTURES TECHNOLOGY
Stochastic Analysis in Parametric Finite Element Modelling
J.G. Sanchez, K. Kirchner and T. Vietor
Institute for Engineering Design, Technische Universität Braunschweig, Germany
J.G. Sanchez, K. Kirchner, T. Vietor, "Stochastic Analysis in Parametric Finite Element Modelling", in , (Editors), "Proceedings of the Tenth International Conference on Computational Structures Technology", Civil-Comp Press, Stirlingshire, UK, Paper 201, 2010. doi:10.4203/ccp.93.201
Keywords: finite element method, stochastic analysis, parametric modelling, structural optimization, vehicle engineering, body design, evolutionary algorithms.
In the past decade, the automotive sector has seen an increasing demand for individualized products at a competitive price. Currently, mass production vehicles are being externally modified in an attempt to fulfil individualized customer needs. Nevertheless, the core of the vehicle is only slightly modified to keep competitive prices through economies of scale. The vehicle underbody is shared by different models on a so called "platform". Today, complex and non standardized alternative fuel systems have pushed platform sharing to the limit.
In trying to create flexible vehicle parts that can easily adapt to external and internal changes, parametric CAD modellers such as FlexBody are developed . FlexBody allows the generation of a parametric CAD model for the vehicle frame body as a tool set; producing parametric beams and joints that are brought together by overlapping connections. To reduce cost and weight, the overlapping area should be minimized. The performance of the connection not only depends on the overlapping area, but also on the production variability of the parts.
In CAD applications, such as FlexBody, safety factors are applied to reckon production variability issues. When a safety factor is applied to fulfil the structural performance target, it may be in conflict with the target of reducing vehicle weight, hence the need for optimization methods that incorporate production variability into the CAD applications.
An optimization methodology named StocFEM is under development, and is the main topic of this paper. In StocFEM the production variability of the different components is used as independent parameters to create a stochastic analysis  of a CAD Model. First, the analysis highlights the key parameters and how they interact . Then, StocFEM will find the most suitable combination of parameters that fulfils the conflicting targets using finite element analysis.
Further development of this methodology could result in customized systems for ANSYS or other finite element applications where engineers can import their CAD models to plug and play key design parameters. These parameters will be predefined in libraries of the finite element application and classified for different parts of the vehicle.
purchase the full-text of this paper (price £20)