H. Weiler

Methods for lightweight design of mechanical components in humanoid robots

With the development of humanoid robots, lightweight construction and energy efficiency play an important role as these mobile, dynamic systems have to work self-sufficiently. The application of computer-aided (CAE) methods in the development process is one possibility to achieve the required weight reduction. On the basis of a classical topology optimization carried out on the robot ARMAR III, an extended system-based method for dynamic, mechatronic systems is presented. Different analysis domains, namely hybrid multibody system dynamics (MBS), finite element analysis (FEA), control system simulation and topology optimization are integrated into a straightforward, automatic way. For the use of fiber reinforced composite materials in such parts, a FE-based method for the determination of ply orientations and thickness relations is presented.

Integrated Topology and Fiber Optimization for 3-Dimensional Composites

The importance of computer aided engineering in product development processes and research has been increasing throughout the past years. As e.g. energy efficiency and therefore mechanical lightweight structures of new products plays a large role, optimization tools gained more and more importance. Weight reduction can be achieved by a change of the component’s design and by selection of adapted materials. Such an improved utilization of material can be implemented only if there is an accurate knowledge of the loads and the conditions on the material. As modern composite materials can make a clear weight reduction possible, appropriate tools and methods are necessary within the design process. Even for isotropic materials, the design of complex parts is not trivial. For the design of composites, additional parameters have to be considered, such as number and thickness of the plies and the orientation of fibers. Hence, design by intuition leads only in few cases to optimal parts. For the determination of the basic layout of a new design topology optimization can be used. It involves the determination of features such as the number, location and shape of holes and the connectivity of the domain. Today topology optimization is very well theoretically studied and also a very common tool in the industrial design process but is limited to isotropic materials. Several approaches for the determination of optimal fiber orientation have been presented in the past e.g. placing the fibers in the direction of the first main stress. Based on a finite element analysis, a method is presented that uses the orientation of main stresses to determine optimal orientations and thickness relations of plies. It is now applicable to complex 3D geometries. The result is a design proposal for the laminate structure (orientation and thicknesses of plies), taking multi-axial load cases into account. To determine a design proposal for complex 3D laminate structures, the application of both methods, topology and fiber optimization, is appropriate. Regarding an independent serial application of topology and fiber optimization it makes sense carrying out topology optimization in a first step and the determination of fiber orientations in a second step. An integrated approach might show even better results in certain cases. For that, we combined topology and fiber optimization in a two-level approach by optimizing laminate structure within each iteration of topology optimization process. In this paper topology and fiber orientation optimization are integrated into a straightforward, automatic way.Copyright

A new Approach for Optimization of Sheet Metal Components

Beads are a widespread technology for reinforcing sheet metal structures, because they can be applied without any additional manufacturing effort and without significant weight increase. The two main applications of bead technology are to increase the stiffness for static loading conditions and to reduce the noise and vibrations for dynamic loadings. However, it is difficult to design the bead patterns of sheet metal structures due to the direction-controlled reinforcement effect of the beads. A wrong bead pattern layout can even weaken the properties of the structure. In the past, the designs were predominantly determined empirically or by the use of so called bead catalogues. Recently, different optimization approaches for bead patterns were developed, which are based upon classical mathematical programming optimization algorithms together with automatically generated shape basis vectors. However, these approaches usually provide only vague suggestions for the designs. One of the most severe difficulty with these approaches is to transfer the optimized results into manufacturable designs. Furthermore, another severe difficulty is that the optimization problem is non-convex, which frequently leads the mathematical programming algorithms into a local optima and thus to sub-optimal solutions. The investigations in this article show an optimization method, which within a few iterations leads to bead structures with excellent reinforcement effects using optimality criteria based approach. Generally, the results can be transferred without large effort into a final design. The new optimization method calculates the distribution of the bending stress tensor and the principal bending stresses based upon the results of a finite element analysis. The bead orientations are calculated by the trajectories of the principal bending stress with the largest magnitude. The beads are projected on to the mesh of the component using geometric form functions of the desired bead cross section. A local bead ratio of 50% (defined as average area of the beads in relation to total area of the sheet) is used by the algorithm to determine the maximum moment of inertia. The proposed algorithm is numerical implemented in the optimization system TOSCA and available for being applied with the following finite element solvers: ABAQUS, ANSYS, I-DEAS, NX Nastran, MSC.Nastran, MSC.Marc and PERMAS. The optimization algorithm is successfully applied to static and dynamic real world problems like car body parts, oil pans and exhaust mufflers. In the present work several academic and industrial examples are presented.