Grounded on structural mechanics, we are working on innovations in design theory of systems based on advanced mathematical approaches, which cover the entire life cycle of mechanical systems from design and development to manufacturing, operation, and disposal. To contribute to sustainable societal development, we aim to create next-generation mechanical systems and devices by integrating research from diverse fields.
Some structures obtained through topology optimization may have material distribution in very thin regions or discontinuities. These features can lead to stress concentration and pose challenges for actual manufacturing. To address this issue, our laboratory is researching an optimization method based on the MMC approach, which utilizes a set of moving morphable components to represent material distribution. This method allows for the control of minimum material thickness and facilitates the design of practical structures that maintain continuity across boundaries in the design domain.
Structural optimization using discrete structural elements, such as frame elements, is formulated based on the principles of material mechanics and structural mechanics, making its mechanical properties well-defined and the interpretation of optimal solutions relatively straightforward. Moreover, the optimal solutions obtained using these elements are explicitly represented by the placement and dimensional parameters of each element, directly linking them to practical design proposals, which is highly beneficial for designers. In our laboratory, we are advancing the development of structural optimization techniques using discrete structural elements, aiming to establish practical design methodologies.
Structural optimization using frame elements (optimal solution and optimization process)
Fluid machinery, such as pumps and turbines, is responsible for energy transfer between fluids. These machines must meet a variety of requirements, including efficient design and reduction of vibration and noise. In particular, many fluid machines involve rotation, making numerical simulations essential for predicting the complex internal flow. In our laboratory, we are advancing research by combining computational fluid dynamics and topology optimization to propose and enhance innovative design methods for fluid machinery.
Derivation of the optimal shape for rotating fans (optimal solution and optimization process)
Many materials, ranging from natural substances like rubber and metal to synthetic materials, exhibit nonlinear characteristics such as viscosity, plasticity, and damage. However, conventional research has often aimed for structures that deform as little as possible, assuming materials to be linear elastic for optimal design. In our laboratory, we are advancing research on optimal design methods that actively combine theories of material mechanics and topology optimization to create shapes that exploit the unique properties of materials, which are not achievable within existing frameworks.
Structural optimization using multiple elastoplastic materials (small and large deformations)
All-solid-state batteries have received significant attention as a solution to address the fire hazards associated with traditional liquid electrolytes. However, the solid-state nature of the electrolyte poses challenges in achieving sufficient contact area between the reactive particles, which limits their overall performance. In our laboratory, we apply topology optimization methods to the internal structure of the battery, enabling us to derive material distributions that maximize the performance of each component based on mathematical and physical principles.
Multi-scale structural optimization for all-solid-state batteries
One of the challenges in topology optimization is that the optimal solution often results in highly complex structures that are difficult to manufacture. As a result, designers must manually modify the obtained optimal structures. However, adjusting the shape to be manufacturable while maintaining its optimal performance is not straightforward and heavily relies on the designer’s intuition, experience, and trial and error. In our laboratory, we are conducting research on dimension constraint methods that incorporate image processing techniques, such as manufacturability-aware thinning algorithms, during the optimization process.
Topology optimization based on the thinning algorithm
Structural optimization is an important research field in engineering design and is applied in a wide range of fields such as architecture, aerospace, mechanical engineering, and automotive industries. The goal of structural optimization is to maximize design performance by reducing material usage, increasing rigidity, and improving durability. However, in actual design environments, uncertainties exist, such as variations in external forces, manufacturing errors, and fluctuations in material properties. Robust optimization aims to find solutions that do not significantly degrade in performance under such uncertainties and variations. In our laboratory, we develop methods to efficiently represent uncertainties affecting structural performance, thereby reducing computational costs and deriving highly robust optimal solutions.
Optimization of trusses (optimization results without and with consideration of variation)
Phononic crystals are materials with artificially designed periodic structures aimed at controlling the propagation of acoustic and elastic waves. These crystals exhibit the ability to block wave propagation within specific frequency ranges known as "band gaps," enabling applications such as filtering acoustic and elastic waves, controlling propagation directions, and localizing energy. In our laboratory, we are introducing topology optimization into the structural design of phononic crystals, focusing on developing methods to maximize band gaps and achieve weight reduction.
Multi-material topology optimization for phononic crystals
With the miniaturization and enhancement of electronic devices, heat generation has become a significant issue. To address the uneven heat distribution, it is possible to determine the dimensions of individual fins within a heat sink, thereby adjusting the flow field and heat transfer rates to facilitate heat dissipation. However, the design of functionally graded heat sinks requires extensive thermal-fluid analysis, making appropriate designs challenging with existing methods. In our laboratory, we are advancing research on optimization methods for fin dimension distribution by utilizing surrogate models that learn the characteristics of the fins
Optimization of fin dimension distribution in gradient-function heat sinks
Bayesian optimization is a method for black-box optimization that explores points sequentially, focusing on those with a high probability of yielding optimal solutions based on existing information. This approach does not rely on the gradients of the objective function, making it promising for application to complex physical problems. Moreover, topology optimization based on Bayesian optimization is expected to be more efficient compared to traditional non-gradient methods. In our laboratory, we are advancing research on the development of this method and its application to complex physical challenges
Topology optimization of a cantilever beam using Bayesian optimization
Dynamic structural performance is a critical metric for evaluating structural behavior and plays a crucial role in ensuring both machinery performance and safety. As a kind of bio-inspired structure, multiscale coated structures combine the advantages of coated structures (enhanced protection and improved buckling resistance ability) and lattice microstructures (high stiffness-to-weight ratio and effective vibration reduction). In our laboratory, we are advancing the potential of multiscale coated structures (especially with non-homogeneous microstructures) through topology optimization to enhance dynamic performance, specifically focusing on maximizing eigenfrequencies and minimizing dynamic compliance.
