A predictive maintenance model for health assessment of an assembly robot based on machine learning in the context of smart plant - Archive ouverte HAL Access content directly
Journal Articles Journal of Intelligent Manufacturing Year : 2024

A predictive maintenance model for health assessment of an assembly robot based on machine learning in the context of smart plant

Ayoub Chakroun
Yasmina Hani
  • Function : Author
Abderrahman El mhamedi
Faouzi Masmoudi
  • Function : Author

Abstract

This paper introduces a predictive maintenance model based on Machine Learning (ML) in the context of a smart factory. It addresses a critical aspect within factories which is the health assessment of vital machinery. This case study specifically focuses on two brass accessories assembly robots and predicts the degradation of their power transmitters, which operate under severe mechanical and thermal conditions. The paper presents a predictive model based on ML and Artificial Intelligence (the Discrete Bayes Filter) to estimate and foresee the gradual deterioration of robots’ power transmitters. It aims at empowering operators to make informed decisions regarding maintenance interventions. The model is based on a Discrete Bayesian Filter (DBF) in comparison to a model based on Naïve Bayes Filter (NBF). The findings indicate that the DBF model demonstrates superior predictive performance compared to the NBF model. The predictive model’s investigation results were validated during testing on robots. This model enables the company to establish an informed and efficient schedule for maintenance interventions.
No file

Dates and versions

hal-04538673 , version 1 (09-04-2024)

Identifiers

Cite

Ayoub Chakroun, Yasmina Hani, Abderrahman El mhamedi, Faouzi Masmoudi. A predictive maintenance model for health assessment of an assembly robot based on machine learning in the context of smart plant. Journal of Intelligent Manufacturing, 2024, ⟨10.1007/s10845-023-02281-3⟩. ⟨hal-04538673⟩
13 View
0 Download

Altmetric

Share

Gmail Mastodon Facebook X LinkedIn More