Multi-state online monitoring based on smart wireless vibration milling toolholder system

Abstract

The advancement of wireless transmission technology has enabled wireless communication for tool condition monitoring, yet current wireless rotating toolholders still exhibit some limitations, including inadequate stiffness, interference with internal cooling channels, limited tool mounting space, and constrained sensor bandwidth and measurement range. To address these issues, this study introduces a smart wireless vibration measurement toolholder featuring symmetrically positioned sensors, with the transmission circuit integrated onto a thermal mounting base and vibration sensors secured via micro-slots. Dynamic and static evaluations confirm the high accuracy and stability of the self-developed smart wireless toolholder, demonstrating a maximum error of 3.37%. Given the dynamic and intricate operating conditions encountered during machining, which necessitate enhanced multi-process condition monitoring capabilities, a hybrid strategy is designed for tool condition monitoring in this work. Specifically, a threshold-based method is applied to detect tool collision and breakage, whereas tool wear state is identified through a gray wolf optimization-enhanced backpropagation neural network (GWO-BPNN). The full life cycle milling experiments were carried out on SiO₂ ceramic matrix composites. The results showed that the threshold judgment method achieves 100% accuracy in detecting tool collision and breakage. Meanwhile, the GWO-BPNN model demonstrates tool wear recognition accuracies of 99.24% on the slice expansion and 97.22% on the small-sample test sets, fully highlighting the high robustness of the proposed smart wireless vibration milling toolholder system (SWVMTS).