An enhanced elastic prediction scheme for thermo-elastoplastic materials with temperature-dependent properties and its intrinsic characters

•Enhanced elastic prediction scheme is developed to overcome the inevitable ‘over-reached plasticity’ problem.•The enhanced scheme possesses higher accuracy, reliability and stability, achieving smooth transition from thermo-elasticity to thermo-elastoplastic state.•Accuracy, stability, existence and complementary constraints are derived.•Accurate result via the scheme ensures KT complementary condition and intrinsic stability automatically, reducing residual of the FB smoothing function.•Optimization with reasonable threshold is designed to promote its performance. Elastic prediction scheme is essential in thermo-elastoplastic simulation; however, current methods do not incorporate nonlinear effects comprehensively and cause the inevitable 'over-reached plasticity' problem: the stress state after prediction is actually plastic, which contradicts to the theoretical foundation of the elastic prediction. An enhanced elastic prediction scheme (EEPS) is developed within framework of the general thermo-elastoplastic constitutive theory. EEPS incorporates both direct and induced nonlinearity due to temperature-dependent properties and overcomes the problem, since it discriminately minimizes the relative yield function in elasticity. The existence of accordant stationary point achieving the minimization is theoretically proved. Several intrinsic characters are derived, revealing that accurate result via EEPS ensures the Kuhn-Tucker complementary condition and grants the Drucker's stability postulates automatically. As proved by theoretically and verified by examples, EEPS exhibits higher accuracy, reliability and robustness, achieving smooth elastic-plastic transition in thermal environment and synchronized evolution of thermo-elastoplastic variables, simultaneously. Furthermore, optimization using reasonable threshold is designed and validated.