Introducing anomdb: An unsupervised anomaly detection method for cnc machine control data

Abstract

We propose the application of unsupervised machine learning to automatically detect anomalous behavior on Computer Numerically Controlled (CNC) machines. We achieve this through an approach that utilizes Principal Component Analysis (PCA), time series feature extraction with the anomalous package in R, and Density Based Scanning of Applications with Noise (DBSCAN). We call this method AnomDB. Time series data collected from CNC machines may benefit from this technique due to its ability to consolidate noisy, multivariate data from CNC machine controls and detect anomalies without reliance on periodicity of signal. We perform experiments on CNC machine control data to demonstrate the effectiveness of this method in discovering anomalies over other commonly used methods of anomaly detection such as Interquartile Range (IQR) and kmeans clustering. We show the effectiveness of this method through a case study of an actual machine anomaly, and then on a series of real machining data with synthetic anomalies injected.