Concrete columns are the most important load-bearing components in civil structures. The potential damage in reinforced concrete (RC) columns could be categorized into three different failure modes: flexural shear (FS) failure, flexural-failure (FF), and shear failure (SF). The corresponding hysteresis loops for each mode differ significantly. Therefore, a multi-parameter hysteretic restoring force model is needed to describe the hysteretic energy dissipation phenomenon and behavior. Identification of the optimal parameter values of a multi-parameter hysteresis model of RC columns under different failure modes is essential in the evaluation of structural inelastic seismic performance. In this paper, a multi-objective optimization algorithm called NSGA-II is employed to identify the parameters of Bouc–Wen–Baber–Noori model (BWBN) hysteresis model, this model has been used for describing the response and modelling restoring force behavior in several structural and mechanical engineering systems, that can fully describe the hysteretic restoring force characteristics of RC columns. An objective function for the restoring force is proposed to identify the parameters of BWBN model. In order to ensure the accuracy of identification, based on the sensitivity analysis, the parameters distribution law of RC columns in different failure modes is obtained. Furthermore, the reference values under different failure modes are proposed. The results presented in this paper will significantly reduce the calculation of subsequent identification. Twelve groups of experimental data are randomly selected to verify the feasibility of the above algorithm. It is demonstrated that using the multi-objective optimization algorithm leads to better identification accuracy with minimum prior experience. Performance of the algorithm is verified using simulated and experimental data. The experimental data of the RC columns were collected from the database of the Pacific Earthquake Engineering Research Center.

混凝土柱是土木结构中最重要的承重构件。钢筋混凝土 (RC) 柱的潜在损坏可分为三种不同的破坏模式：弯曲剪切 (FS) 破坏、弯曲破坏 (FF) 和剪切破坏 (SF)。每种模式对应的磁滞回线差别很大。因此，需要一个多参数的滞后恢复力模型来描述滞后能量耗散现象和行为。确定不同破坏模式下 RC 柱多参数滞后模型的最佳参数值对于结构非弹性抗震性能评估至关重要。在本文中，采用一种称为 NSGA-II 的多目标优化算法来识别 Bouc-Wen-Baber-Noori 模型 (BWBN) 滞后模型的参数，该模型已用于描述多个结构和机械工程系统中的响应和建模恢复力行为，可以充分描述 RC 柱的滞回恢复力特性。提出了恢复力的目标函数来识别 BWBN 模型的参数。为保证辨识的准确性，在灵敏度分析的基础上，得到了钢筋混凝土柱在不同破坏模式下的参数分布规律。此外，还提出了不同故障模式下的参考值。本文提出的结果将大大减少后续识别的计算量。随机选取12组实验数据来验证上述算法的可行性。结果表明，使用多目标优化算法可以以最少的先验经验获得更好的识别精度。使用模拟和实验数据验证了算法的性能。RC柱的实验数据来自太平洋地震工程研究中心的数据库。