Research On Fatigue Damage Characterization Based On Maximum Entropy Principle

Fatigue damage of material is caused by cyclic loading.In microelectronics packaging microscopic field,auto industry,aerospace or other fields,their main components and parts mostly work under cyclic load,and fatigue fracture is the main failure mode.Fatigue damage model is the key to describe the fatigue damage.So it's important to study and improve the fatigue damage model.In order to predict the fatigue behavior of microscopic and macroscopic specimen under cyclic loading,improvement measures were proposed based on the maximum entropy fracture model.And digital image correlation method was also used to characterize the fatigue damage evolution of the sample.Satisfactory results were obtained.The main research work of this article includes the following aspects:(1)A single parameter maximum-entropy fracture model is developed in this paper.The nonlinear kinematic hardening model is used to describe the hardening behavior of the material and the maximum entropy fracture model was used as the fatigue damage model to describe the cyclic degradation and cyclic softening behavior of the material.Then,the non-local integral averaging method was used to solve the problem that the simulation results in the fatigue simulation process are mesh-dependent.(2)The application of the maximum entropy fracture model in the field of microelectronics packaging was studied.By calculating the accumulated plastic dissipation energy in the cyclic curve,the accumulated damage parameter was finally obtained.However,the fitting value was based on the assumption of stress equalization,so it cannot be directly put into the model.Considering that model parameter is difficult to determine,the inverse analysis method based on Kalman filter algorithm was proposed to obtain the accumulated parameter of damage.By constructing the response interpolation function,the rate of inversion is greatly improved.(3)A three-dimensional digital image correlation system based on Bayesian model is proposed to obtain the deformation of specimens accurately.The initial parallax was calculated based on the Bayesian model and iteratively optimized by Newton-Raphson method.At the same time,a highly efficient backward Gauss Newton method was used to perform nonlinear iterative optimization of the initial displacement value to achieve fast and highly accurate timing matching.In the consideration of that initial value is hard to find in the iterative optimization,the seed point method was used.And it not only provides a reliable initial value for the optimization process,but also can effectively avoid transmission of error value.(4)A fatigue damage evolution characterization method was proposed based on the combination of digital image correlation method and finite element method.The matching relationship between the results of the digital image correlation method and the finite element model was also established.It can be validated that the maximum entropy fracture model has a good applicability in macro model,and it also can describe the damage process of the macro sample more intuitively.