Experience-based process optimization cannot accurately and automatically adapt to the time-varying cutting condition of thin-walled parts because of the large amount of removed material. Thus, machining chatter and error are easily generated if the process parameters are not adjusted in time. To address this problem, a novel intelligent milling methodology is proposed for thin-walled parts, which has three innovations: data-driven time-varying information model construction, adaptive process optimization, and service-oriented automatic process execution. Using this methodology, a time-varying information model of a machining system was first constructed to characterize the actual cutting condition, and this model can be used as the digital twin model for machining thin-walled parts. In addition, the in-process physical and geometric data in the current process step were collected to construct the information model. Subsequently, the process parameters of the next process step were adaptively optimized based on the information model. Finally, a service-oriented process execution stream was built to automate the entire milling process, including data collection, process optimization, and process execution. To verify the proposed method, intelligent milling was performed on thin-walled parts. The experimental results demonstrated that this method can effectively and automatically control the machining chatter and thickness error according to the actual cutting condition. Compared with nominal milling, the machining surface roughness improved from Ra 2.4 to Ra 1.6, and the thickness accuracy improved from ±0.03 to ±0.02 mm.
Read full abstract