Improved milling stability analysis for chatter-free machining parameters planning using a multi-fidelity surrogate model and transfer learning with limited experimental data

Decision-making on chatter-free machining parameters is essential for process planning since chatter significantly affects production quality and efficiency. Stability lobe diagram (SLD) is commonly used for selecting chatter-free machining parameters, but its analytical prediction often has poor accuracy and experiment-based prediction is time-consuming. This paper proposes a multi-fidelity (MF) surrogate model and transfer learning-based method to improve the milling stability analysis. Firstly, an analytical stability model is constructed to predict low-fidelity (LF) SLDs for key combinations of radial cutting width (ae) and feed rate per tooth (ft). A few spindle speeds (ns) are selected from each key LF SLD to detect high-fidelity (HF) stability limits (aplim) through milling experiments. Subsequently, sufficient LF and limited HF combinations of ns, ae, ft, and aplim are taken to construct additive scaling function-based MF stability models. Predicted MF combinations of ns, ae, ft, and aplim are combined with limited HF combinations to construct more accurate stability models through transfer learning. Then, a neural network is ultimately trained to predict aplim values for arbitrary combinations of ns, ae, and ft. A detailed experimental validation indicates that the proposed method can provide more accurate lobe boundaries for machining parameters selection by introducing fewer experimental samples.