Hughes et al. [1] introduced the term isogeometric analysis (IGA) in the framework of finite element analysis (FEA). Its main idea is to use the same mathematical description for the geometry as well during the design process in a computer aided design (CAD) environment as in the later analysis phase using FEA. Numerous research papers devoted to IGA have demonstrated beneficial and superior analysis properties, using higher order and higher continuity basis functions compared to standard, low order finite elements. As B-splines and non-uniform rational B-splines (NURBS) are the most widely used geometry descriptions in CAD, NURBS-based finite elements have been developed and implemented into LS-DYNA over the last few years.
IGA
Isogeometric Analysis (IGA) [1] is a rather new approach to Finite Element Analysis (FEA), using spline basis functions known from Computer Aided Design (CAD) for describing both the geometry and the solution field. The main motivation for IGA is the integration of design and analysis. Achieving such a full integration requires a holistic approach with a fundamentally different modeling strategy and development process to exploit the full potential of IGA. Such changes certainly take time and cannot be achieved overnight. Fortunately, IGA with its higher-order and higher-continuity elements also offers several additional advantages such as an accurate geometry description, superior analysis qualities, a larger explicit time step size or smart modeling techniques. Thus, users may benefit from IGA immediately, even without a full paradigm shift.
Engineering workflows are habitually split into a modelling phase and a consecutive analysis phase, which is primarily driven by the finite element method (FEM). However, bridging the gap between design and analysis remains a sophisticated problem and may consume a vast amount of computational as well as manual operations, especially in highly iterative development processes. To avoid this major bottleneck, Isogeometric Analysis (IGA) [1] and later Isogeometric B-Rep Analysis [2] were developed. They rely on the mathematical descriptions of Computer Aided Design (CAD), such as NURBS- and B-Spline-based boundary representation (B-Rep) models. However, classical B-Rep formulations describe a solid only by its boundary faces and do neither provide any physical nor geometrical description of the interior. Therefore, the IGA concept cannot be applied to three-dimensional structures in a straightforward manner.
During the last two years, ANSA, the leading preprocessor for crash analysis, has been heavily involved supporting the IGA community, helping create IGA models and Hybrid FE-IGA full vehicle crash models. In doing so we a set a dual target. First, advance and explore all needed technologies which are mainly new and mostly in academic research phase. Second, bring in, industry expertise in making these technologies robust and well suited for production, both in terms of stability and performance, but also in terms of data interoperability and adaptation to the highly automated process flows that characterize current automotive crash procedures. The cooperation with the ANSYS LST team has been very fruitful and rewarding. These latest developments in the pre-processor side are presented in this presentation.
At present, the Finite Element Analysis (FEA) method is indispensable in the field of simulation technology, as this kind of numerical analysis method can assist engineers to predict results, which are often difficult to obtain from experimental tests. However, there exist some problems in terms of finite element mesh generation time and geometric representation. In this studying, we adopted a new numerical analysis method, Isogeometric Analysis (IGA) to develop static and dynamic analyses on two models, a notched plate and a wind turbine tower model in Ls Dyna software. From the static convergence analysis result, it is shown that IGA is more time-efficient compared with FEA. In terms of fatigue analysis results, IGA can predict the fatigue life corresponding very well to the fatigue life computed by FEA. It can be concluded that IGA is appropriate for the numerical analysis.