Distortion-minimized de-homogenization for optimization of cell-size distribution in TPMS structures
Hiroki Kawabe, Kaito Ohtani, Yusibo Yang, Musaddiq Al Ali, Kentaro Yaji · 2026
Read Distortion-minimized de-homogenization for optimization of cell-size distribution in TPMS structures by Hiroki Kawabe, Kaito Ohtani, Yusibo Yang, Musaddiq Al Ali, Kentaro Yaji (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-06011.txt.
Introduction
Hiroki Kawabe and four co-authors improve how optimized lattice like structures are turned into buildable designs. Their setting is topology optimization of TPMS structures, where cell size varies across a part. A step called de-homogenization converts the optimized pattern into a real structure, but the standard method distorts it badly.
Their new technique directly shrinks the mismatch between target and actual cell wavenumbers, solving the task as a Poisson equation. In a stiffness test, the strain energy of their design differed from the ideal model by just 0.8 percent, against 63.6 percent for the standard method. Physical tests confirmed the gain, with effective stiffness well above the uniform design.
Catalog record
| CREATOR | Hiroki Kawabe, Kaito Ohtani, Yusibo Yang, Musaddiq Al Ali, Kentaro Yaji |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 88 KB |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.06011
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Distortion-minimized de-homogenization for optimization of cell-size distribution in TPMS structures Hiroki Kawabe, Kaito Ohtani, Yusibo Yang, Musaddiq Al Ali, Kentaro Yaji Abstract This paper presents a homogenized topology optimization (TO) method for spatially optimizing cell-size distribution of triply-periodic minimal surface (TPMS) structures, with high accuracy in the optimized structural response after de-homogenization. To achieve this, we introduce a novel de-homogenization technique that directly minimizes the difference between the wavenumbers obtained from the target and actual size distributions. This minimization problem is efficiently solved as a typical Poisson’s equation utilizing the discrete cosine transform (DCT). We first verify the proposed de-homogenization method through numerical examples, showcasing its capability in significantly reducing the known distortion of the de-homogenized TPMS structures from the conventional periodic modulation (PM) method. Then, we apply the proposed method to a stiffness maximization problem, to demonstrate its effectiveness in improving the structural response compared to the PM method, through numerical examples and experimental validation. The proposed method successfully reduced the distortion of the de-homogenized structures compared to the PM method, leading to 0.8% difference in the strain energy compared to the homogenized model, as opposed to 63.6% difference in the PM method. The optimized structure from the proposed method shows a significant improvement in the strain energy by 50.1% compared to the uniform case in the FE analysis on the de-homogenized models, while the PM method results in a significant decrease of 45.8%. The experimental validation shows that the effective stiffness of the optimized structure from the proposed method is 54.2% higher than that of the uniform case, while the PM method results in a significant decrease by 77.3%. These results exhibit the proposed method effectively increases the accuracy of the de-homogenization, thereby maximizing the potential of the homogenized TO for the spatial cell-size optimization of TPMS structures. 1 Introduction Porous structures [ 5 ] , such as truss lattice structures [ 31 ] and honeycomb structures [ 32 ] , are well known for their lightweight nature combined with high stiffness, high strength, and high surface area. Owing to these advantageous properties, these structures have been widely used in various fields, including structural components [ 18 ] , medical applications [ 50 ] , and thermal systems [ 36 ] . These traditional porous structures can be easily manufactured using conventional manufacturing methods, such as casting and machining. However, they inherently possess sharp corners and edges, which can lead to stress concentration and potential failure under mechanical loading [ 52 ] . With the advancement of additive manufacturing (AM), the fabrication of complex structures has become increasingly feasible, enabling the development of highly efficient porous architectures with enhanced performance [ 27 , 37 ] . Recently, triply-periodic minimal surface (TPMS) structures [ 1 ] have gained much attention as a promising alternative to traditional porous structures, due to their smooth surfaces without sharp corners and high surface area to volume ratio, leading to high specific stiffness and strength [ 11 ] . TPMS structures are represented by implicit functions, which allow for easy control of the thickness and size of the TPMS structure, full connectivity, and quasi-self-supporting nature. Utilizing these capabilities of TPMS structures, various design methods have been proposed to generate functionally graded TPMS structures with spatially varying thickness distributions [ 33 , 17 ] and size distributions [ 47 , 16 ] , or combining multiple types of TPMS structures [ 48 , 23 ] . Due to its simplicity, the thickness variation method has been widely used in various applications, such as the design of bone scaffolds [ 9 ] , tissue scaffolds [ 24 ] , heat exchangers [ 8 ] , and structural components [ 49 ] . Despite its effectiveness in enhancing mechanical performance of the TPMS structures, the enhancement is limited because their shape is not drastically changed. The multiple type method can offer more drastic shape changes, by mixing multiple types of TPMS structures [ 2 , 43 , 46 , 29 ] . However, the multiple type method requires a transition region with special treatment like thickening, which may compromise the performance of the structure [ 35 ] . In contrast, the size variation method can offer more gradual changes in the TPMS structure without the need for a distinct transition region, potentially preserving the mechanical performance of the structure. The method has been used in various applications, such as the design of medical implants [ 21 ] , heat exchangers [ 13 ] , and structural components [ 2 , 7 ] . However, the design of the size distribution has been mostly based on heuristic approaches, with the number of design variables often limited to a few. To overcome these limitations, topology optimization (TO) [ 4 ] can be promising. TO creates innovative and high-performance designs based on the mathematical formulation of the optimization problem and numerical optimization algorithms, which enables the efficient exploration of a vast design space without the need for human intuition. However, it is inherently challenging to utilize TO for designing complex geometries such as TPMS structures, due to the high-resolution discretization required to accurately capture their complex geometries, leading to extremely high computational cost. To address this issue, Hu et al. [ 15 ] adopted a super-element technique [ 28 ] , which divides the design domain into coarse subdomains called super-elements, while each super-element is divided into fine elements called background elements. This method can reduce the computational cost by performing most of the finite element analysis (FEA) processes on the super-element level, while the design variables such as parameters for controlling the thickness and size of the TPMS structures are defined on the background element level. The super-element-based TO proved to be effective in designing high-stiffness functionally graded TPMS structures, later extended to the heat dissipation problem [ 44 ] . Ning et al. [ 30 ] extended the super-element-based TO to the design of hybrid TPMS structures by introducing a data-driven surrogate model. These methods, however, require the integral computation of the stiffness matrix for each super-element at each iteration, which can be computationally prohibitive especially when each super-element needs high-resolution to capture small cell sizes or drastic changes in geometry. Another approach to tackle the computational cost issue is to use a homogenization method [ 3 ] . This homogenization-based TO method assumes that the effective properties of the TPMS structure can be estimated by analytical or numerical homogenization methods, utilizing its periodic nature. By optimizing the distribution of the local structural features in the homogenized model and reconstructing the actual geometry of the TPMS structure from the optimized distribution of structural features, the method can efficiently design functionally graded TPMS structures with a large number of design variables. This reconstruction process is called de-homogenization, which is a critical step to ensure that the optimized design from the homogenized model can be accurately reflected in the reconstructed TPMS structure to achieve the optimized performance. Li et al. [ 19 ] proposed an asymptotic homogenization-based TO method for the stiffness problem and the thermal conductivity problem, optimizing the relative density distribution of the TPMS structures. Strömberg [ 38 ] utilized a numerical homogenization method using three repeated cells