Research Dataset from a Smart Factory Model for Case-Based Predictive Maintenance

This dataset provides research data from a smart factory model for predictive maintenance and is intended to support reproducible research on temporal case-based reasoning on multivariate sensor time series. It is based on the smart-factory data introduced by Klein and Bergmann in Data Generation with a Physical Model to Support Machine Learning Research for Predictive Maintenance. Building on these data, Klein et al. later developed the corresponding similarity-learning approach for predictive maintenance. The underlying resources are made available through the companion repository PredM/SiameseNeuralNetwork. The underlying data originate from the IoT Lab Trier.

The original resources were consolidated and prepared into an XML-based dataset as described in the publication Towards Explainable and Reusable Temporal Case-Based Reasoning in Predictive Maintenance: A Research Environment. The resulting dataset follows the XML schema specified by the ProCAKE framework and comprises the case base, queries, and validation data in a format that can be used with the ProCAKE-PredM project. In this form, the dataset provides a reusable basis for research on temporal case-based reasoning, including future work on explanation and adaptation in case-based predictive maintenance.