Data for: Online motion accuracy compensation of industrial servomechanisms using machine learning approaches

Published: 19 July 2024| Version 1 | DOI: 10.17632/g8mrry54j8.1
Contributors:
Pietro Bilancia,

Description

Here are shared the files related to the study presented in the paper: Title: "Online motion accuracy compensation of industrial servomechanisms using machine learning approaches" Authors: Pietro Bilancia, Alberto Locatelli, Alessio Tutarini, Mirko Mucciarini, Manuel Iori and Marcello Pellicciari Two folders are shared: 1) Experimental Data --> contains the results obtained from an extensive experimental campaign carried out on a test rig for industrial servomechanisms. - Speed --> {100,200,...,18000] rpm (18 levels) - Output torque --> {0,100,...,1800} Nm (19 levels) - Oil temperature --> {25,30,35}°C (3 levels) 2) Machine learning models --> ONNX files saved from Python after training and readily available to be imported and utilized within Programmable Logic Controllers to achieve motion prediction and compensation.

Files

Steps to reproduce

The experimental data can be plotted with the Matlab script provided in the folder. The ONNX files are organized based on the related paper and can be imported and managed within the Beckhoff TwinCAT environment (guidelines are available here https://www.beckhoff.com/en-en/products/automation/twincat-3-machine-learning/). For any specific questions, feel free to contact the Authors.

Institutions

Universita degli Studi di Modena e Reggio Emilia - Sede di Reggio Emilia

Categories

Mechanical Engineering, Robotics, Machine Learning, Compensation Method

Licence