Titanium alloys have become an indispensable material for all walks of life because of their excellent strength and corrosion resistance. However, grinding titanium alloy is exceedingly challenging due to its pronounced material characteristics. Therefore, it is crucial to create a theoretical roughness prediction model, serving to modify the machining parameters in real time. To forecast the surface roughness of titanium alloy grinding, an improved radial basis function neural network model based on particle swarm optimization combined with the grey wolf optimization method (GWO-PSO-RBF) was developed in this study. The results demonstrate that the improved neural network developed in this research outperforms the classical models in terms of all prediction parameters, with a model-fitting R² value of 0.919.

To establish a high-precision prediction model for the material removal rate (MRR) of TC4 titanium alloy material in vertical vibratory finishing equipment, an orthogonal experiment was conducted using TC4 titanium alloy plate as the experimental specimen. We performed variance analysis of the impact of vibration frequency, the phase difference, the mass of upper eccentric block, and the mass of lower eccentric block on the MRR. We then drew the main effect diagram and analyzed the influence of various process parameters on the MRR. Mathematical regression and a neural network were used to construct predictive models for the MRR with respect to process parameters, and a genetic algorithm (GA) was coupled to optimize the neural network to improve the predictive performance of the model. By calculating the R², validating the set sample prediction error, and averaging the absolute percentage error (MAPE) of each model, it was found that the neural network model had better prediction performance than the mathematical regression model, with an accuracy of 82.2%. After coupling with the GA, the prediction accuracy reached 95.5%. The research results indicated that, compared with mathematical regression and the original neural networks, the neural network coupled with the GA had better predictive performance, providing an effective method for predicting the MRR in vertical vibratory finishing.

The surface roughness of blisks during vibratory finishing is a critical evaluation index for their processing effect. Establishing a surface roughness prediction model helps reveal the processing mechanism and guide the optimization of process parameters. Therefore, based on wear theory and the least squares centerline system, a relationship between the surface roughness and material removal depth was established, and a scratch influence factor was introduced to correct the impact of surface scratches on the theoretical model. Interaction parameters between the blisk and granular media were obtained through discrete element simulations and used as input parameters for the model. Machining experiments were conducted to solve the model coefficients and verify the model’s effectiveness. The results show that the average error between the surface roughness model predictions and experimental results is 11.8%. As the machining time increases, the surface roughness exhibits three successive stages: accelerated decrease, decelerated decrease, and stability. The surface roughness decreases most rapidly at 48 min of machining and reaches the machining limit at 198 min. The surface roughness prediction model established in this study effectively reveals the coupling mechanism between the scratch accumulation and material removal during vibratory finishing, providing a basis and methodology for determining the process parameters in blisk vibratory finishing.