Skip to content

AAAipa/dataset_unfastening

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

dataset_unscrewing

This repo contains the data set, which has been used in the technical paper Machine learning based screw drive state detection for unfastening screw connections published in Journal of Manufacturing Systems, Volume 65, October 2022.

Dataset Structure

Each of the different datasets contains two subfolders, one containing training data and the other containing test data. The train and test data is split by lables into loesbar (releasable) and nicht_loesbar (non_releasable) in each folder. Each json file contains the data of one unscrewing process and contains different key values that provide information about the unscrewing process.

dataset overview
The dataset overview

Key parameter

The tourque can be optained with following python code:

path = 'your-path-to-the-dataset'
with open(path) as f:
   df = json.loads(f.read())
torque = np.array(df['tightening steps'][0]['graph']['torque values'][0:self.sequence_length])

The corresponding angle values have equidistant steps of 5.320°.

The file also contains other important keyvalues:

  • df['prg name']: screw type
  • df['tightening steps']['torque']: max tourque
  • df['tightening steps']['speed']: max speed
Dataset Description Trial Number Train/Test
1 Benchmark 1 & 2 80%/20%
2 Screw Size Independency 3 80%/20% (M4,M5/M6,M8)
3 Screw Head Independency 4 80%/20% (Torx, E. Hexagon/ Philips, I. Hexagon)
4 Sparse Dataset1 5 20%/80%
5 Sparse Dataset2 6 160/80 datasets
6 Sparse Dataset3/Half Data 7 80%/20%

Used Hardware

Bosch Rexroth Nexo NXP for manual unscrewing operations MINIMAT ®-EC-SERVO and deprag AST40 for the robot-based unscrewing operations

Reference

@article{ALASSADI202219,
title = {Machine learning based screw drive state detection for unfastening screw connections},
journal = {Journal of Manufacturing Systems},
volume = {65},
pages = {19-32},
year = {2022},
issn = {0278-6125},
doi = {https://doi.org/10.1016/j.jmsy.2022.07.013},
url = {https://www.sciencedirect.com/science/article/pii/S0278612522001248},
author = {Anwar {Al Assadi} and David Holtz and Frank Nägele and Christof Nitsche and Werner Kraus and Marco F. Huber},
keywords = {Screws, Machine learning, Neural networks, Unfastening, Disassembly, Battery systems}
}

License

This work is licensed under a Creative Commons Attribution 4.0 International license.

SPDX-License-Identifier: CC-BY-4.0

Acknowledgement

Sponsored by the Ministry of the Environment Baden-Württemberg, in the context of the Strategic Dialogue Automotive Industry, and supervised by the Project Management Agency Karlsruhe (PTKA). Funding number: L7520101

Link zur Projektseite: Industrielle Demontage von Batteriemodulen und E-Motoren DeMoBat

About

This repo contains the data set of the publication "Machine learning based screw drive state detection for unfastening screw connections", which have been published on in october 2022.

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published