Anomaly Detection in Batch Manufacturing Processes Using Localized Reconstruction Errors From 1-D Convolutional AutoEncoders

Multivariate batch time-series data sets within Semiconductor manufacturing processes present a difficult environment for effective Anomaly Detection (AD). The challenge is amplified by the limited availability of ground truth labelled data. In scenarios where AD possible, black box modelling approaches constrain model interpretability. These challenges obstruct widespread adoption Deep Learning solutions. objective study to demonstrate an approach which employs 1-Dimensional Convolutional AutoEncoders (1d-CAE) and Localised Reconstruction Error (LRE) improve performance Using LRE identify sensors that result in anomaly, explainability solution enhanced. Tennessee Eastman Process (TEP) LAM 9600 Metal Etcher datasets have been utilised validate proposed framework. results show outperforms global reconstruction errors similar architectures achieving AUC 1.00. unsupervised learning with AE improves expected be beneficial deployment semiconductor interpretable trustworthy are critical process engineering teams.