Title Dataset: Tool Wear Experiment Dataset for the Evaluation of Sensor Informativeness in Longitudinal Turning (ExtraDrey) Dataset DOI: https://doi.org/10.15480/882.17237 Publication DOI: 10.1515/zwf-2026-1068 Author: Sebastian Schibsdat, sebastian.schibsdat@gmx.de, ORCID: 0009-0001-0605-9754 , affiliation: Hamburg University of Technology - Institute of Production Management and Technology Dataset creation date: 01.06.2026 Version: 1.0 License: Public Domain Mark 1.0 Universal Context: Investigation of the wear behaviour of coated indexable inserts in CNC longitudinal turning. Multi-sensor process signals are recorded together with interval microscopy of the cutting tool to study the informativeness of different sensors for predicting flank wear (VB_Max). Acknowledgement: This work was funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) – 521385417 (SPP 2402 - Greybox models for the qualification of coated tools for high-performance cutting, subproject D2: Extrapolative digital greybox models for describing and predicting the macroscopic system behavior of TiAlN-coated cutting tools) Structure of the dataset: The data is organised per tool edge. Each tool edge (CT0XX_Y) is packed into its own uncompressed .tar archive. The readme, the mask label definition and the experiment documentation spreadsheet are provided unpacked next to the archives. - CT0XX_Y.tar (one per tool edge, 67 edges total) Each archive extracts to a single folder CT0XX_Y/ containing: CT0XX_Y/ Experiment_Data/ sensor .mat files of all experiments run on this edge Images/ Image/ microscope images Mask/ segmentation masks (sparse - not every image has a mask) Heightmap/ extracted heightmaps (gzip-compressed CSV) - LabelDefinition.txt Class labels used in the segmentation masks (e.g. Background, Flank_Wear, Crater_Wear, Breakage, Tool_Area, Adhesion). - Experiment_Documentation.xlsx Combined metadata of all experiments (identical to the MetaData stored inside the .mat files). Experimenter names have been removed; the VB_Max (maximum flank wear) column is populated. Information about the data: - Images: [CT0XX_Y_Zs_AAB.jpg] Microscope images taken in intervals including 0 s. CT = Cutting Tool, XX = Cutting Tool Number, Y = Cutting Edge Number, Z = Time in use [s], AA: FF = Free Flank, CE = Cutting Edge, RF = Rake Face B : [] = 150x Magnification (Standard), L = 100x Magnification - Masks: [CT0XX_Y_Zs_AA_mask.png] Pixel-wise segmentation masks of the corresponding microscope image. Not every image has a mask. Class labels are defined in LabelDefinition.txt. - Heightmaps: [CT0XX_Y_Zs_AA_3D.csv.gz] Extracted heightmap of the microscope image, stored as a gzip-compressed CSV. Naming follows the image convention. - Experiment_Data: [ES0XX_YYY_Data.mat] Sensor data of the experiments (30 s cutting) combined, synchronised and enhanced. Metadata included. Experimenter names removed; VB_Max filled in the MetaData. File format: MAT-File Version 7.3 (HDF5-based format) Data structure: - MetaData (struct) -- Data (cell) -- Description (cell) - SensorNI (struct) -- Data (cell) -- Description (cell) - SensorOptosurf (struct) -- Data (cell) -- Description (cell) - SensorQUASS (struct) -- Data (cell) -- Description (cell)