ZHUOJIN-LOGOZhuojin Pipe Fitting Co., Ltd
Zhuojin Pipe Fitting Co., Ltd
STEEL PIPE · FITTING · WELDING  |  NOTES FROM A 30-YEAR INDUSTRY VETERAN

Topology Optimization of Square Concrete-Filled Steel Tube Truss Structures Using a Pseudo-Full-Stress Genetic Algorithm

Core Technical Content

The paper by Xie Jun, Guo Fei, Zhu Shouqin, Liang Jinxiao, Yan Jie, and Dong Jie from Hebei University of Architecture presents a novel method for topology optimization of square concrete-filled steel tube (CFST) truss structures. The authors address the discrete variable topology optimization problem by combining a genetic algorithm with a pseudo-full-stress algorithm, hence the name "pseudo-full-stress genetic algorithm."

The methodology involves several key steps. First, a genetic algorithm is used to randomly generate initial topological configurations. Then, a heuristic algorithm is employed to check and correct these configurations, ensuring that they are structurally valid. The optimization variables include the steel tube cross-section type, topological variables, and concrete strength grade. The objective is to minimize the total structural cost. The proposed method integrates the pseudo-full-stress algorithm into the genetic algorithm framework, with some individuals in the initial population generated from the pseudo-full-stress algorithm's optimized solutions. Additionally, the penalty function of the traditional genetic algorithm is improved to enhance computational efficiency.

The effectiveness of the proposed method is demonstrated through a case study involving a 12-bar truss structure. The results show that the proposed method achieves a lower total cost compared to both the pseudo-full-stress algorithm and the traditional genetic algorithm when applied to square CFST truss structure topology optimization. The optimized structure has fewer members, and all members are fully utilized in terms of their load-carrying capacity, indicating a significant improvement in the topology optimization outcome.

Engineering Practice Implications

The topology optimization of truss structures is a critical task in structural engineering, as it directly affects the material consumption, structural weight, and overall cost of the structure. For square CFST trusses, the optimization problem is particularly challenging due to the discrete nature of the design variables (steel tube cross-sections are available in standard sizes) and the interaction between the steel tube and the concrete core.

From my perspective in the steel pipe industry, the use of square CFST members in truss structures offers several advantages. The square cross-section provides excellent torsional rigidity and is well-suited for connections where multiple members meet at a node. The concrete core enhances the compressive capacity of the steel tube and prevents local buckling, allowing the use of thinner wall thicknesses. However, the concrete core also adds weight to the structure, which can be a disadvantage in certain applications, particularly where the structure is subjected to significant lateral loads or where weight is a critical design parameter.

The proposed method's ability to simultaneously optimize the topology, steel tube cross-sections, and concrete strength grade is a significant advancement. In practice, this allows engineers to achieve a more efficient design that balances material costs and structural performance. However, I would caution that the optimization results are only as good as the structural analysis model used to evaluate the performance of each candidate design. The use of a simplified analysis model, such as a linear elastic truss model, may not capture the nonlinear behavior of CFST members, particularly the effects of concrete cracking and steel yielding.

Lessons and Reflections

The paper's contribution lies in the integration of the pseudo-full-stress algorithm with a genetic algorithm, which addresses two key challenges in topology optimization: the computational efficiency of the search process and the need to ensure that all members in the optimized structure are fully utilized. The improved penalty function is a valuable addition, as it helps the genetic algorithm converge more quickly to the optimal solution.

However, I would like to point out that the proposed method has some limitations. First, the case study involves a relatively simple 12-bar truss structure, and the performance of the method on larger, more complex structures is not demonstrated. Second, the optimization is based on a linear elastic analysis, which may not be conservative for structures subjected to seismic loads or other dynamic effects. Third, the method does not explicitly consider the effects of member buckling, which is a critical failure mode for truss structures with slender members.

In my experience, the most common mistakes in topology optimization of truss structures include: (1) neglecting the effects of secondary moments at the connections, which can significantly increase the stresses in the members; (2) using an overly simplified load model that does not represent the actual loading conditions; and (3) failing to consider the practical constraints of fabrication and erection, such as the availability of standard steel tube sizes and the need for adequate connection details.

Reference Value and Outlook

This paper presents a novel and practical approach to the topology optimization of square CFST truss structures. The proposed method is computationally efficient and produces designs that are both cost-effective and structurally efficient. The integration of the pseudo-full-stress algorithm with a genetic algorithm is a clever solution that addresses the challenges of discrete variable optimization.

Looking forward, I believe that the next steps in this field should include: (1) extending the method to consider nonlinear structural analysis, including the effects of concrete cracking and steel yielding; (2) incorporating the effects of member buckling and connection flexibility; (3) applying the method to larger, more realistic structures; and (4) developing design guidelines that can be used by practicing engineers.

For engineers in the field, I recommend using this paper as a reference for understanding the application of optimization algorithms to CFST truss structures. However, I would emphasize that optimization tools should always be used in conjunction with sound engineering judgment and a thorough understanding of the structural behavior. The optimized design should be verified through detailed analysis and, where possible, experimental testing before it is implemented in practice.