Abstract:<p>In this work, we propose a topology optimization (TO) method using fully adaptive truncated hierarchical B-splines (ATHB-TO), where the adaptive isogeometric analysis mesh and design mesh of ATHB-TO are locally refined and coarsened simultaneously. To meet the demands of full-mesh adaptivity of ATHB-TO, a marking strategy is established based on density variances between adjacent active elements. Then, an adaptive formulation for the mark strategy is proposed to avoid the overrefinement phenomenon that occurred in the adaptive TO method using a fixed mark strategy. Moreover, a density inheritance strategy is devised to project original design variables into design variables concerning a fully adaptive hierarchical mesh. Afterward, ATHB-TO with an adaptive mark strategy is successfully applied to two-dimensional and three-dimensional benchmarks. Numerical results show that the ATHB-TO method can obtain identical optimized designs at a lower computational burden, and improve the convergence rate compared with THB-TO with only hierarchical local refinement considered. Adopting the adaptive mark strategy releases the full potential of ATHB-TO in improving the computational efficiency compared with the fixed mark strategy. Additionally, the adaptive TO schemes outperform TO performed on uniformly refined meshes due to the enhanced numerical efficiency and improved structural performance with a largely accelerated convergence rate. Therefore, ATHB-TO is an effective way of solving structural TO problems.</p>
mechanics,engineering, multidisciplinary,mathematics, interdisciplinary applications
What problem does this paper attempt to address?
The problem that this paper attempts to solve is how to balance computational efficiency and accuracy in structural topology optimization. Specifically, the author proposes a topology optimization method using fully adaptive truncated hierarchical B - splines (ATHB - TO) to overcome the contradiction between accuracy and efficiency in traditional methods when optimizing on a fixed dense grid. By simultaneously locally refining and coarsening the analysis grid and the design grid, the ATHB - TO method can obtain the same optimized design while reducing the computational burden and improve the convergence speed.
### Main problems
1. **Contradiction between accuracy and efficiency**:
- Traditional topology optimization methods are usually carried out on a fixed dense grid, which leads to a contradiction between accuracy and efficiency. The use of high - order elements further exacerbates this contradiction.
- In order to alleviate this contradiction, adaptive grid techniques or improving the accuracy of the analysis model can be adopted.
2. **Limitations of adaptive grid techniques**:
- Early adaptive topology optimization methods mainly focused on local refinement of the grid, but lacked consideration of grid coarsening, which limited their potential.
- Fixed marking strategies may lead to excessive refinement near the structural boundary, thereby increasing the computational degrees of freedom.
3. **Geometric discretization error**:
- Using linear Lagrange elements for meshing will result in low numerical accuracy, poor stability, and geometric discretization errors.
### Solutions
- **Fully adaptive truncated hierarchical B - splines (ATHB - TO)**:
- By introducing a fully adaptive marking strategy, the phenomenon of excessive refinement near the structural boundary is avoided.
- Simultaneously considering the local refinement and coarsening of the design grid and the analysis grid improves computational efficiency.
- Using truncated hierarchical B - splines (THB) restores the partition - of - unity property of hierarchical B - splines and improves numerical accuracy.
### Technical details
1. **Adaptive marking strategy**:
- It includes a triggering strategy, a local refinement strategy, and a local activation strategy, which are used to determine when to perform full - grid adaptation, which areas need local refinement, and which areas need local coarsening.
- By adjusting the marking parameter \(\theta\), the excessive refinement phenomenon caused by the fixed marking strategy is avoided.
2. **Density inheritance strategy**:
- When the grid is locally refined, the new design variable is initialized as the density of the parent element.
- When the grid is locally coarsened, the new design variable is initialized as the average of the densities of the sub - elements.
3. **Numerical implementation**:
- The optimization criterion (OC) is used as an optimizer.
- The effectiveness of the method is verified through multiple two - dimensional and three - dimensional benchmark tests.
### Conclusions
- The ATHB - TO method can obtain the same optimized design as traditional methods while reducing the computational burden and significantly improve the convergence speed.
- Through the fully adaptive marking strategy, the ATHB - TO method performs excellently in improving computational efficiency, outperforming methods with fixed marking strategies and uniformly refined grids.
In conclusion, this paper effectively solves the contradiction between accuracy and efficiency in traditional methods by proposing a new adaptive topology optimization method, providing a new solution for structural optimization.