A detected-data-enhanced FEM for residual stress reconstruction and machining deformation prediction

Abstract

Deformation is one of the bottleneck problems for complex mechanical components during their multi-process machining. The traditional FEM method generally predict the deformation by the simulation stresses, however, this model is too idealistic to get an accurate result. Therefore, a detected-data-enhanced finite element method (DDEFEM) is proposed. First, some scattered stress data are detected by practical tests. Then, the data are input into the FE model to reconstruct the stress by cycle compute the static equilibrium equations, until the calculated stress is approach to the measured. Thus, DDEFEM realize the accurate prediction of multi-process machining deformation based on stress, that revised by the measured data. To verify the method, three different models are carried out: the FE model without data input, the FE model with simulation data input, and the DDEFEM model. The results of the three method is 34.88 %, 26.80 %, and 15.84 % respectively, which verifies that the proposed method has a higher accuracy. Moreover, some valuable conclusions are obtained by using DDEFEM, such as the machining deformation increases with the material removal rate, and the nonlinear coupling effect between IRS (Initial residual stress) and MIRS (machining-induced residual stress).

Keywords

• Residual stress
• Machining deformation
• Stress reconstruction
• FEM and data fusion
• DDEFEM