Inverse analysis of residual stress in orthogonal cutting

• Published in: Journal of manufacturing processes 2019-02, Vol.38, p.462-471
• Main Authors: Mirkoohi, Elham, Bocchini, Peter, Liang, Steven Y.
• Format: Article
• Language: English
• Subjects: Cutting process parameters; Inverse modeling; Process parameters planning; Residual stress prediction
• ISSN: 1526-6125; 2212-4616
• DOI: 10.1016/j.jmapro.2019.01.033

•The proposed inverse analysis method is able to estimate the process parameters and find the optimal desired residual stress.•The maximum error between the predicted residual stress and experimental measurements is less than 2.95%, so the proposed approach is reliable.•The solution is obtained after three to four iterations which indicates efficiency of the proposed model.•The proposed approach, not only does not require the finite element analysis --A process that is computationally expensive--, but also it eliminates the costly experiments. An inverse analysis is proposed to obtain the orthogonal cutting process parameters for desirable residual stresses for the first time. The residual stress has substantial influence on fatigue life, corrosion resistance, part distortion, and crack propagation. Hence, having a desired residual stress can help to control the abovementioned concerns. The proposed model uses an analytical model to predict the residual stress directly. The residual stress is induced by severe thermo-mechanical loading. The heat generated during the machining process is obtained using an imaginary heat source model. An elastic-plastic stress analysis approach is used to predict the residual stress by incorporating a hybrid function as a limiting case for very large and very small plastic strain range. Next, an iterative gradient search is used to adaptively approach the specific residual stress by the optimization of the process parameters. The inverse analysis identifies the cutting process parameters such as depth of cut and cutting speed and predicts the optimal solution for the residual stress. Experimental measurements validated the effectiveness of the proposed model since the maximum calculated error is less than 4.5% between the predicted residual stress and experimental measurements.