Optimization method for increasing the failure loads of bolt joints in fiber-reinforced structures
Fiber-reinforced composites have enormous lightweight potential and a wide range of possible applications. However, their anisotropy and inhomogeneity complicate the strength analysis and sizing of fiber-reinforced structures. The many degrees of freedom in the designing process (Material selection, number, thickness, orientation and sequence of layers) also make it difficult to achieve optimum material utilization. A structured and objective-oriented sizing can be achieved with the help of numerical structural optimization. This process is part of the present thesis. The optimization of bolt joints, which are used for releasable connections of fiber composite structures, is treated. For this purpose, a method is developed, which uses an existing commercial software (HyperWorks from Altair) for the composite optimization. Additionally, the process includes the topology optimization of a modified initial model, the transfer of the generated result into a laminate model and the dimensioning based on it. Optionally, contact modeling and analysis can be applied before and after this process to provide an approximation of the contact load for detailed mapping of the load application in a load model and to enable an accurate verification calculation of the optimization result. The capabilities of the developed method are demonstrated by two examples. With a cantilever disk optimization, the plausibility and the quality of the results are proven using comparison models and a reference process. In the application example, the bolt joint is treated in detail. In addition to considering the contact situation and the approximation of the contact load, an initial design is dimensioned and modeled. As an optimization objective, this bolt joint is reinforced with the smallest possible added volume so far that the double failure load can be carried at maximum stiffness. The optimum reinforcement is achieved with the developed method and represents an improvement with respect to comparison models. Also, the generalized initial design within the method simplifies the finding of a feasible optimal design compared to a reference process.
