Stability Modeling in Aerospace Machining: A Physics-Guided Machine Learning Approach with Adaptive Sampling for Self-Aware Operations

Stability control of the machining process for aerospace parts with complex geometries and tight tolerances can be difficult to maintain in operational environments due to uncertainty about optimal parameter selection. On the shop floor, machining parameters are typically selected based on manufacturer recommendations or, in most cases, operator experience. The fusion of physics-based and data-driven approaches in physics-guided machine learning (PGML) represents a transformative breakthrough for optimal parameter selection that helps to assure simultaneous achievement of part accuracy, high surface finish and productivity. PGML leverages measurement data generated during the machining process while incorporating decades of theoretical process modeling efforts and domain knowledge. Acquiring the large amounts of data needed for purely data-driven models can be costly. On the other hand, many physical processes are not completely understood by domain experts so physics-based models must make simplifying assumptions that compromise optimal parameter selection. Further, machining parameters that yield stable dynamics depend not only on the time-varying dynamics of the specific machine-tool-workpiece, but also on specific conditions in the ambient operational environment. As a result, selection of optimal parameters to avoid instabilities (c.f. chatter) and maximize productivity requires continuous monitoring and updating of the SLD to assure that the process remains within a stable regime. In this paper, the PGML approach combined with an adaptive, data-driven sampling methodology for collecting measurement data is shown to offer an efficient method to determining the true underlying stability model for optimal parameter selection. Knowledge of the true stability model enables operators—and ultimately the machines themselves—to make in-process optimum parameter adjustments during production. As machines become more intelligent, self-knowledge of the true stability model is the foundation for self-aware operations and, as required, machine self-control to perform parametric adjustments that maintain process stability.