Industrial screw driving dataset collection: Time series data for process monitoring and anomaly detection

Industrial Screw Driving Datasets This comprehensive dataset collection captures time-series data from industrial screw driving operations in plastic components. Designed for research in manufacturing process monitoring, anomaly detection, and quality control, it includes over 34,000 individual screw driving operations across six distinct scenarios. Each scenario investigates specific aspects of the screw driving process, from natural wear patterns to controlled material variations. Overview Scenario name Observations Classes Purpose s01_variations-in-thread-degradation 5,000 1 Studies natural degradation of plastic threads over repeated use cycles, documenting wear patterns and failure progression s02_variations-in-surface-friction 12,500 8 Examines effects of different surface conditions (lubricants, surface treatments, contamination) on screw driving performance s03_variations-in-assembly-conditions-1 1,700 26 Investigates diverse component and assembly faults including washer modifications, thread deformations, and alignment issues s04_variations-in-assembly-conditions-2 5,000 25 Features methodically arranged assembly fault conditions in 5 distinct error groups with paired normal/abnormal operations s05_variations-in-upper-workpiece-fabrication 2,400 42 Analyzes how injection molding parameter variations in the upper component affect screw driving metrics s06_variations-in-lower-workpiece-fabrication 7,482 44 Explores effects of injection molding parameter variations in the lower component on fastening quality Experimental Setup Automatic screwing station (EV motor control unit assembly) Delta PT 40x12 screws optimized for thermoplastics Target torque: 1.4 Nm (range: 1.2-1.6 Nm) Sampling frequency: 833.33 Hz Data completeness: >95% Detailed time-series measurements including torque, angle, gradient, and time values Dataset Features These datasets are suitable for various research purposes. Each ZIP file contains: Complete raw data in JSON format for maximum flexibility Standardized labels.csv files for metadata and classification Comprehensive README.md documentation for each scenario Various error classes with varying degrees of severity Time series data for the complete screwing process Research Applications Development of machine learning models for anomaly detection Process monitoring and quality control system development Manufacturing analytics and parameter optimization Digital twin development for screw driving operations Material property influences on assembly processes Access and Usage For data handling, we recommend our PyScrew Python package (https://github.com/nikolaiwest/pyscrew). However, this is optional: The data is easily accessible using standard JSON and CSV processing. No changes were made to the raw data and all information on the experiments can be found in the labels.csv file. Citation Request When using this dataset in academic work, please cite this project or the mentioned papers from the README.md files. Contact and Support For questions, issues, or collaboration interests regarding these datasets, either: Open an issue in our GitHub repository PyScrew Contact us directly via email Acknowledgments These datasets were collected and prepared by: RIF Institute for Research and Transfer e.V. Technical University Dortmund, Institute for Production Systems The preparation and provision of the research was supported by: German Ministry of Education and Research (BMBF) European Union's "NextGenerationEU" program The research is part of this funding program More information regarding the research project is available here