Zhang Xun
School of Management, Guangzhou University, Guang Zhou, 510006, GuangDong, China.
Heliyon. 2024 Feb 27;10(5):e27179. doi: 10.1016/j.heliyon.2024.e27179. eCollection 2024 Mar 15.
Seismic design principles advocate for simple and regular structures to minimize earthquake damage. However, this frequently does not lead to unique and aesthetically pleasing designs, leading some engineers to select irregular structures despite the potential risks. The primary aim of this investigation is to achieve the optimal design of torsional irregularity coefficients for planar irregular reinforced concrete (RC) frames under static and dynamic loads, utilizing a 3D 6-layer model. Structural ground vibration analysis was conducted using the ETABS software. By imposing limits on the torsional irregularity coefficients for each layer of the frame layout, we subsequently applied the combination of artificial neural networks (ANN) with the particle swarm optimization (PSO) algorithm, namely ANN-PSO, to address the size distribution issue across the structure. The design variables included the dimensions of the columns located in each layer of the layout. The results demonstrate that the ANN-PSO algorithm optimizes the cross-sectional area of columns with significant variations. The coefficients of the torsion inequality rule in the optimized solution closely approach the minimum values. The dimensions and orientations of the optimized columns slightly differ from the pre-optimized scheme. In the optimized scheme, the coefficients of the torsional irregularity in the Y-direction meet the requirements, preventing any torsional irregularities from occurring. The research presented an effective method, including an innovative combination of ANN-PSO and the finite element method (FEM), for designing RC structures. The findings of the research provided a practical solution to fulfill torsional regularity criteria, indicating the proposed approach is an effective method for the economical and safe design of RC structures in earthquake-prone areas. The outcomes of the present study highlighted the innovative framework to achieve optimal and safe designs for irregular RC structures while minimizing torsional damage during earthquakes.
抗震设计原则主张采用简单规则的结构,以尽量减少地震破坏。然而,这往往不会产生独特且美观的设计,导致一些工程师尽管存在潜在风险仍选择不规则结构。本研究的主要目的是利用三维六层模型,实现平面不规则钢筋混凝土(RC)框架在静态和动态荷载作用下扭转不规则系数的优化设计。使用ETABS软件进行结构地面振动分析。通过对框架布局各层的扭转不规则系数施加限制,随后我们应用人工神经网络(ANN)与粒子群优化(PSO)算法相结合,即ANN - PSO,来解决结构的尺寸分布问题。设计变量包括布局各层中柱的尺寸。结果表明,ANN - PSO算法优化了柱的截面积,且变化显著。优化解中的扭转不等式规则系数非常接近最小值。优化后的柱尺寸和方向与预优化方案略有不同。在优化方案中,Y方向的扭转不规则系数满足要求,可防止任何扭转不规则情况发生。该研究提出了一种有效的方法,包括ANN - PSO与有限元方法(FEM)的创新组合,用于RC结构设计。研究结果提供了满足扭转规则标准的实用解决方案,表明所提出的方法是在地震多发地区经济安全地设计RC结构的有效方法。本研究结果突出了创新框架,以实现不规则RC结构的最优和安全设计,同时在地震期间将扭转破坏降至最低。