An optimization method of processing parameters for large-bore diesel engine cylinder heads to minimize residual stress
摘要
Suboptimal selection of processing parameters for marine diesel engine cylinder heads can result in significant residual stress in critical structural components, thereby increasing the risk of fatigue failure. This study introduces “birth–death element” technology for the comprehensive simulation of the multi-operation machining process of the large-bore cylinder head (LBCH) from the SCE450 medium-speed diesel engine, with the objective of optimizing process parameters to mitigate residual stress. Initially, the Johnson–Cook (J–C) constitutive model of cylinder head material was experimentally determined. A full-process machining simulation using the birth–death element technique revealed that in SCE450 cylinder heads, the maximum equivalent stress was 139.5 MPa, the maximum principal stress was 117.4 MPa, and the minimum principal stress was − 88.1 MPa. Subsequently, blind-hole method measurements of residual stresses on the fire deck surface demonstrated good agreement with simulation results in fatigue critical regions. The maximum principal stress errors ranged from 5.3 to 13.9%, and the minimum principal stress deviations ranged from 4.3 to 18.7%. Finally, a robust nonlinear mapping between machining parameters and residual stress responses was established through the development of a Back-Propagation (BP) neural network. This mapping achieved a root-mean-square error (RMSE) of 6.2 in stress prediction. Through this high-precision model, multi-process parameter optimization reduced the maximum residual stress in the optimal solution from 205.46 to 186.28 MPa. Results confirmed that residual stress discrepancies between measurements and simulations remained within 20%, verifying the reliability of the simulation model. This integrated simulation–optimization framework provides a systematic methodology for machining process design in heavy-duty engine components.