Accurately predicting tool wear during the machining process not only saves machining time and improves efficiency but also ensures the production of good-quality parts and automation. This paper proposes a combined variational mode decomposition (VMD) and back propagation (BP) neural network model (VMD-BP), which maps spindle power to tool wear. The model is trained using both historical and real-time data. To improve accuracy, the internal power data from the machine tool are used to calibrate the model’s input data. Data collected from milling experiments are used to test the model, with sensor-collected power being compared to the model’s predicted power. The average error was 1.1256%, which confirms the reliability of the model. In practical applications, the model enables the real-time monitoring of spindle power, helping prevent excessive tool wear during machining. This offers significant guidance for actual production processes.

Thin-walled aluminum alloy parts are widely used in the aerospace field because of their favorable characteristics that cater to various applications. However, they are easily deformed during milling, leading to a low pass rate of workpieces. On the basis of on-machine measurement (OMM) and surrogate stiffness models (SSMs), we developed an iterative optimization compensation method in this study to overcome the machining deformation of thin-walled parts. In the error compensation process, the time-varying factors of workpiece stiffness and the impact of prediction model errors were considered. First, we performed machining deformation simulation and information extraction on the key nodes of the machined surface, and an SSM containing the stiffness information of discrete nodes of each cutting layer was established. Subsequently, the machining errors were monitored through intermittent OMM to suppress the adverse impact of prediction model errors. Further, an interlayer correction coefficient was introduced in the compensation process to iteratively correct the prediction model error of each node in the SSM along the depth direction, and a correction coefficient between parts was introduced to realize the iterative correction of the prediction model for the same node position between different parts. Finally, the feasibility of the proposed method was verified through multiple sets of actual machining experiments on thin-walled parts with added pads.

The 7075-T7451 aluminum alloy, widely used in aerospace, aviation, and automotive fields for critical load-bearing components due to its excellent mechanical properties, suffers from residual stresses induced by thermo-mechanical coupling during milling, which deteriorate workpiece performance. This study explores how key milling parameters—spindle speed *n_c*, feed per tooth *f_z*, cutting depth *a_p*, and cutting width *a_e*—affect surface residual stress and cutting force via orthogonal experiments and finite element analysis (FEA). Results show *a_e* is critical for X-direction residual stresses, while *f_z* dominates Y-direction ones. Cutting force increases with *f_z*, *a_p*, and *a_e* but decreases with higher *n_c*. Multivariate regression-based prediction models for residual stress and cutting force were established, which effectively characterize parameter–response relationships with maximum prediction errors of 18.69% (residual stress) and 12.27% (cutting force), showing good engineering applicability. The findings provide theoretical and experimental foundations for multi-parameter optimization in aluminum alloy milling and residual stress/cutting force control, with satisfactory practical effectiveness.

In the aerospace field, Ti–Al alloy thin-walled parts, such as blades, generally undergo a large amount of material removal and have a low processing efficiency. Scheduling the feed rate during machining can significantly improve machining efficiency. However, existing feed-rate scheduling methods rarely consider the influence of machining deformation factors and cannot be applied in the finishing stages of thin-walled parts. This study proposes an offline feed-rate scheduling method based on a local stiffness estimation model that can be used to reduce machining errors and improve efficiency in the finishing stage of thin-walled parts. In the proposed method, a predictive model that can rapidly calculate the local stiffness at each cutter location point and a cutting-force prediction model that considers the effect of cutting angle are established. Based on the above model, an offline feed-rate scheduling method that considers machining deformation error constraints is introduced. Finally, an experiment is performed by taking the finishing of actual blade parts as an example. The experimental results demonstrate that the proposed feed-rate scheduling method can improve the machining efficiency of parts while ensuring machining accuracy. The proposed method can also be conveniently applied to feed-rate scheduling in the finishing stage of other thin-walled parts without being limited by machine tools.