高强钢变模量随动强化本构模型匹配与解耦标定策略研究

高强钢通过微观组织调控获得高强度,但不同牌号高强钢微观的不均匀变形和微观诱导塑性机理不同,使得高强钢卸载及反向加载行为更加复杂,并且牌号间差异增大,为此给出模型自适应匹配及参数解耦匹配的系统化策略,实现了高强钢回弹精确预测。首先建立幂函数和指数函数混合硬化模型,基于混合模型给出自由弯曲加载的弯矩平衡方程和曲率约束方程,基于变模量模型构建截面弹性弯矩的积分方程,基于加载和卸载解析模型建立逆向识别卸载参数的子优化模型。确定变模量线性随动强化、变模量非线性随动强化模型和含边界面的变模量随动强化模型的匹配策略。基于自由弯曲、单向拉伸和拉压试验数据,确定相应本构的子优化模型参数的优化次序,最终形成本构匹配及其参数解耦标定的系统化策略,并基于Fortran语言开发标定程序库。建立U形弯曲件和弧形弯曲件预测模型,分别对DP980和DH980两种高强钢不同应变水平下的识别结果及回弹预测结果进行对比分析,验证了解耦标定策略不仅提高了不同牌号数据的相关度,而且大幅度提升了同一牌号下的模型精度和稳定性,为基于数据的材料性能统一自辨识方法研究奠定了基础。

Research on Matching of Variable Modulus Kinematic Hardening Constitutive Models and Decoupling Calibration Strategy for High-strength Steel

Higher strength of high-strength steel can be obtained by controlling microstructure, but the microscopic uneven deformation and micro-induced plasticity mechanism of different grades of high-strength steel are different, which makes the unloading and reverse loading behavior of high-strength steel more complicated and increases the difference between grades. For this reason, a systematic strategy of model adaptive matching and parameter decoupling matching to achieve accurate prediction of high-strength steel springback is given. Firstly, a hybrid hardening model of power function and exponential function is proposed.Based on the hybrid hardening model, the bending moment balance equation for free bending load and curvature constraint equation are raised. In view of the variable modulus model, the integral equation of the section elastic bending moment is established. A sub-optimization model is established to reversely identify the unloading parameters basing on the loading and unloading analytical models. The matching strategy including variable modulus linear and nonlinear kinematic hardening models and variable modulus kinematic hardening models with boundary surface is determined. Based on the experimental data of tension and compression, free bending and uniaxial tension, the optimization sequence of the corresponding constitutive sub-optimization model parameters is determined, and a systematic strategy of constitutive matching and its parameter decoupling calibration is formed finally, and calibration library is developed on the basis of the Fortran language. A prediction model for U-shaped bending parts and arc-shaped bending parts is established, and the recognition results and springback prediction results of DP980 and DH980 high-strength steels at different strain levels are compared and analyzed. The decoupling calibration strategy is verified. Not only the correlation of data of different grades of high-strength steel but also the accuracy and stability of the model under the same grade is greatly increased. It lays the foundation for the research on the unified self-identification method of material properties in view of data.