A Dataset of Univariate Crimp Force Curves for Data-Driven Time Series Analysis

Area of Science:

• Manufacturing Engineering
• Data Science
• Quality Control

Background:

• Data availability is a major challenge for developing data-driven tools in manufacturing.
• Crimp connections are crucial in industrial settings and require robust quality control.
• Existing methods for analyzing crimp force data are limited.

Purpose of the Study:

• To introduce a novel, comprehensive dataset of crimp force curves for manufacturing applications.
• To facilitate the development of domain-specific machine learning models for crimping processes.
• To provide a benchmark dataset for univariate time series analysis in industrial quality control.

Main Methods:

• Collected 2,439 crimp force curves from a semi-automatic crimping machine.
• Annotated curves using a state-of-the-art anomaly detection system.
• Provided detailed manual classification of curves into multiple quality categories.

Main Results:

• A unique dataset of 2,439 annotated crimp force curves is now available.
• The dataset includes both automated anomaly detection and detailed manual quality classifications.
• This facilitates research in machine learning and time series analysis for manufacturing.

Conclusions:

• The presented dataset addresses the critical need for data availability in manufacturing.
• It serves as a valuable resource for advancing data-driven quality control in crimping processes.
• Enables further research and development of intelligent manufacturing systems.