New paradigm of plastic deformation theory modelling:Artificial intelligence empowered and data science-driven
Traditional plastic deformation theories are based on the phenomenological approach which mainly rely on the experience of the developers and are obtained by fitting the experimental data,which have severe limitations.With the increasing complexity of microstruc-tures and mechanical properties of materials,it has become extremely difficult to develop more complex phenomenological constitutive models.Non-universal material models,which are difficult to relate to material properties and manufacturing processes,have become one of the main problems limiting plastic forming and processing,and also pose a serious challenge to the development of plasticity theory.The development of artificial intelligence and data science opened up new possibilities in materials and mechanical sciences.With the rapid development of material design and application,the traditional explicit constitutive model is not needed,the theoretical modelling approa-ches to plastic deformation that reflect the relationship between the microstructures and macroscopic properties of materials come into be-ing.Among them,the new paradigm modelling theories represented by mechanistically informed neural networks,efficient multi-scale clustering analysis and model-free data-driven computational mechanics have their distinctive features and significant advantages,which are the most promising simulation methods.Through the summary of these three aspects,the development status and future trend of plastic theoretical modeling and multi-scale simulation technology development driven by artificial intelligence and data science were discussed.
data drivenmulti-scalecomputational mechanicsmachine learningmodel-free simulation
万方数据
维普
