Acoustic emission and machine learning based classification of wear generated using a pin-on-disc tribometer equipped with a digital holographic microscope

The efficiency of processes involving frictional contacts between surfaces is often characterized by wear rates or friction coefficients. However, the classification and forecasting in related a real industrial challenge that unsolved today. Hence, an on-line monitoring system able to classify rate can be crucial for many industries as it could help preventing catastrophic failures. Applications include lifetime assessment various components where range failures occur such scuffing (a typical sudden failure mechanism). These tribological now sensorized, corresponding sensor signatures modelled monitored using state-of-the-art Machine learning (ML) algorithms. In this study, we use Acoustic Emission (AE) ML frameworks different categories simulated with customized pin-on-disc tribometer. A real-time investigation track necessary find out origins scar visible at surface. To achieve objective, experiments were conducted on tribometer equipped Digital Holographic Microscope (DHM). Experiments carried alumina steel balls against discs room temperature. Real-time DHM images surface recorded each lap same position. An acoustic emission AE signals during complete duration experiments. signatures, combination images, correlated input ground truth labels algorithm. Several compared; they are support vector machine, logistic regression, XGBoost, random forest, neural networks, k-Nearest Neighbor, quadratic discriminant analysis Naive Bayes. classifier was trained differentiate features rates. Most algorithms had average accuracy above 80%, highest obtained machine (84.7%). improved combining two neighboring limited differences terms rate. proposed method has significant potential in-situ quality since requires minimum modifications commercially available machines.