A constitutive model is proposed for both the cyclic hardening and softening of structural steel. A shrinking rule is incorporated into the memory surface for cyclic softening. The theoretical formulation of the proposed model is derived under cyclic loading with a gradually decreasing strain amplitude. The material constants are identified by numerically minimising the difference between the theoretical result and a single material test result using a genetic algorithm. The proposed model is validated by comparing numerical and material test results under various cyclic loadings for structural steel, where the model exhibits good ability to qualitatively and quantitatively evaluate the stress amplitude variation. The model is implemented in Abaqus FEA software, and the convergence rate of the consistent tangent operator is evaluated. FEA with the consistent tangent operator yields a quadratic convergence.

提出了结构钢的循环硬化和软化的本构模型。收缩规则被纳入记忆表面以进行循环软化。所提出模型的理论公式是在应变幅度逐渐减小的循环载荷下推导出来的。材料常数是通过使用遗传算法在数值上最小化理论结果和单一材料测试结果之间的差异来确定的。通过比较结构钢在各种循环载荷下的数值和材料试验结果，对所提出的模型进行了验证，该模型具有良好的定性和定量评估应力幅变化的能力。该模型在 Abaqus FEA 软件中实现，并评估一致切线算子的收敛速度。