遗传算法在Y型偏心支撑组合框架抗震性能优化中的应用研究

THE APPLICATION OF GENETIC ALGORITHM IN SEISMIC PERFORMANCE OPTIMIZATION OF Y-SHAPE ECCENTRICALLY BRACED COMPOSITE FRAME

  • 摘要: 该文针对Y型偏心支撑组合框架抗震性能优化展开研究,共选择5层、10层和15层高的三类典型钢-混凝土组合框架作为研究对象,采用考虑楼板空间组合效应的高效全杆系纤维模型对研究对象进行数值模拟。提出了适用于Y型偏心支撑组合框架的遗传算法程序,包括染色体编码方法、适应度计算方法、进化终止条件以及在遗传过程中的选择、交叉与变异法则,对偏心支撑的布置位置和力学参数分别进行了优化设计。结果表明:遗传算法程序能够迅速找到Y型偏心支撑组合框架抗震性能优化问题的最优解,收敛性良好,相比于传统枚举遍历算法,在求解中层、高层结构偏心支撑布置位置的优化问题时,计算成本明显降低。根据不同参数下遗传算法计算成本的分析,遗传算法中种群规模、淘汰比例和变异概率的最优取值随着楼层数目的增加而略有增加,对于布置位置优化问题,种群规模宜取4~12,淘汰比例和变异概率宜取10%~20%;对于力学参数优化问题,种群规模宜取4~8,淘汰比例宜取20%~30%,变异概率宜取10%~20%。

     

    Abstract: A research on seismic performance optimization of Y-shape eccentrically braced composite frame is conducted in this paper. A total of three typical structures of five-story, ten-story and fifteen-story are selected as the research objects, and the efficient fiber beam model considering the spatial composite effect of slabs is adopted to simulate the research objects. A genetic algorithm is developed for the Y-shape eccentrically braced frame, and the chromosome coding rules, fitness function, termination condition and the rules of selection, crossover and mutation are specified. This algorithm is then applied to the optimization of the distribution strategies and mechanical parameters of eccentric braces, respectively. The results indicate that the proposed genetic algorithm can rapidly find the optimum solution for the seismic performance optimization problem of Y-shape eccentrically braced composite frame, and shows good convergence. Compared with traditional traversal algorithm, the genetic algorithm can significantly reduce the calculation cost of optimizing the distribution of the eccentric braces when applied to mid-rise and high-rise structures. According to the parametric analysis on the cost of the genetic algorithm, the species population, elimination ratio and mutation ratio are slightly growing with the increase of story numbers. For the distribution optimization problem, the species population is suggested to be 4~12, the elimination ratio and mutation ratio are both suggested to be 10%~20%. For the mechanical parameter optimization problem, the species population is suggested to be 4~8, the elimination ratio is suggested to be 20%~30%, and the mutation ratio is suggested to be 10%~20%.

     

/

返回文章
返回