Signal

CNC machines are used in production industries for batch production. In CNC machining, a minor issue can cause production downtime, reducing productivity and profit for the Industry. In CNC machines, drilling machine maintenance is crucial because of the complexity of the drill tools. Drill tools have complex shapes and geometries, making tool wear prediction particularly challenging. Tool wear in CNC drilling severely hinders performance and affects the dimensional accuracy and surface finish obtained. This paper presents a machine-learning-based approach to drill wear detection using the Hilbert–Huang Transform for feature extraction from airborne Acoustic Emission (AE) signal and the CatBoost algorithm for classification. For controlled drilling operations, AE signals from four wear-condition samples representing Healthy Tool (HT), Low Wear (LW), Medium Wear (MW), and Severe Wear (SW) were recorded. Wear levels of 0.3mm,0.6mm, and 0.9mm for the drill bits of 3.0 mm, 3.2 mm, 3.4 mm, 3.6 mm, and 3.8 mm diameters were created using Electrochemical Machining in the Lab. Using AE sensors, the signals were collected and converted into the required format with the support of signal conditioning and a data acquisition system. LabVIEW software was used to display the signal, and it was then decomposed using the Hilbert-Huang Transform (HHT) to obtain the required Intrinsic Mode Functions (IMFs). Features needed for classification, such as magnitude, entropy, and instantaneous frequency, were selected in the time-frequency domain. These features were used as input to a classifier (CatBoost), which was trained and evaluated using 10-fold cross-validation. HHT-CatBoost achieved 99.1% accuracy, indicating a promising sign for the proposed algorithm in real-time maintenance for small- to medium-sized datasets.