Work

Multivariate time series dataset for predicting surface roughness and dimensional error in longitudinal turning

Year: 2026

Type: data-paper

Abstract: Quality control in machining is essential for consistent workpiece quality. Predictive quality methods that predict the workpiece quality based on monitored signals are used instead of real measuring devices in order to reduce non-productive times and save costs by detecting scraps already during the process. The methods often apply machine learning models since they can model complex relationships and can handle a wide range of signal sources. While machine learning models require a lot of training data, only few datasets are publicly available that include process signals from various sensors. The presented dataset provides multivariate time series data of 453 longitudinal turning processes with varying feed, cutting speed, depth of cut and workpiece material. Additionally, the resulting surface roughness and diameter deviation are provided for each cut. All cuts were carried out using new cutting tools. To minimize measuring errors, the surface roughness and diameter deviation were measured on-machine using machine integrated touch probes. This dataset supports research into predictive quality using machine learning models.

Source: Data in Brief

DOI: 10.1016/j.dib.2026.113325

Authors: Berend Denkena, Benjamin Bergmann, Henning Buhl, Dennis Kowalke, Miriam Handrup

Institution: Leibniz University Hannover

Language: English

Topic: Advanced machining processes and optimization

Subfield: Mechanical Engineering

Field: Engineering

Domain: Physical Sciences

Open Access status: gold

Estimated APC paid: $1,220

Funder: Deutsche Forschungsgemeinschaft