Search

This paper proposes a multi-objective optimization method using an adaptive weight determination scheme and tunneling method. To find the optimum designs that are evenly distributed on the Pareto front, the weights of different objective functions are adaptively determined by using the concept of a hyperplane so that new solutions can be gradually found in the objective space. In addition, to avoid locally optimized solutions that are dominated by other Pareto optimum designs, a tunneling method is employed in the optimization process. To confirm the effectiveness of the proposed method, topology optimization of the magnetic actuator is performed. The magnetic force and volume of the actuator are considered as two conflicting objective functions in the optimization problem, and Pareto optimum solutions that are evenly distributed in the objective space could be obtained. [ABSTRACT FROM AUTHOR]

In this paper, a new method is presented to concurrently determine the structural layout and joint interface during the multi-material topology optimization (MMTO) process. Although the development of additive manufacturing techniques allows the fabrication of multi-material structures for soft materials with graded properties, joint materials for joining metals or composites are still needed. This paper proposes a novel material interpolation scheme for defining the joint area between dissimilar materials using a two-step filtering process. The general MMTO process is performed in the first filtering step. Furthermore, the filtered variables generate the joint area, and the filter radius controls the thickness of the area in the second filtering step. A modified discrete material optimization (DMO) approach is developed to control the different materials independently and to expand the applicability of the method to cases with more than two design materials. To demonstrate the performance of the proposed method, a compliance minimization problem is formulated for various volume constraints, joint thicknesses, material properties, mesh types, and number of materials. To show the scalability, 3-D design and compliant mechanism design examples are adopted. Based on numerical examples, it was confirmed that the proposed method performs well in various cases; moreover, the results demonstrate that the concurrent designing of the structural layout and joint interface leads to better performance than when joint stiffness is not considered. • A new multi-material topology optimization for considering joint stiffness between dissimilar materials is proposed. • A modified discrete material optimization (DMO) approach for defining joint area is developed. • The two-step filtering process for controlling the structural feature size and the joint thickness are adopted. • Via numerical examples with various cases: joint thickness, material property, and mesh type, the proposed method is validated. [ABSTRACT FROM AUTHOR]

In this paper, we propose the use of deep learning-based image super-resolution to accelerate structural topology optimization. Topology optimization suffers from iterative computation and a time cost that increases with the number of elements. Recently, there have been attempts to accelerate topology optimization using deep learning-based models, but they often do not address a wide range of physical conditions or require extensive training data. This highlights the need for an approach that can effectively solve different problem conditions with less training data. Our approach first starts topology optimization at a low resolution to quickly obtain an optimized structure. It is converted to the structure at a higher resolution by using an image super-resolution model. Then, the final structure is obtained by performing topology optimization at this high resolution by using the converted structure as the starting configuration. The super-resolution model learns how to transform low-resolution structural features into high-resolution ones regardless of optimization conditions. As a result, it can be applied to any problems different from the problem used to generate the data for training the model. The proposed approach's excellent acceleration and generalization performance is demonstrated for three representative problems in structural topology optimization with a small number of training datasets. Furthermore, it has been shown to be effective in the deck arch bridge problem and the natural frequency maximization problem. Finally, the proposed approach is also found to be effective on the three-dimensional cantilever beam problem. [ABSTRACT FROM AUTHOR]