Finishing milling parameter optimization based on Isight

The surface quality of products, such as the servo power supply bracket from solid rocket engines machined by integral milling, often exhibits instability, necessitating the optimization of milling finishing parameters. To tackle this issue, th is research utilizes the optimal Latin hypercube sampling method within the multidisciplinary optimization software, Isight. Initially, finite element simulation experiments are conducted to ascertain the coefficient of variation (CV) of the milling force of the tool and the surface roughness (Ra) of the material under various milling parameter combinations. Subsequently, radial basis function neural network (RBF) and Kriging methods are employed to construct surrogate models for the milling parameters, as well as for the milling force CV and material surface roughness Ra. Finally, the study utilizes the archival micro-genetic algorithm (AMGA) for multi-objective optimization of the surrogate models, facilitating the iterative calculation and identification of optimal milling finishing parameters.