Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Industrial Machines Health Prognosis using a Transformer-based Framework

David J Poland, Lemuel Puglisi, Daniele Ravi

TL;DR

Transformer Quantile Re-gression Neural Networks (TQRNNs) are introduced, a novel data-driven solution for real-time machine failure prediction in manufacturing contexts and can increase high-quality production, improving product yield from 78.38% to 89.62%.

Abstract

This article introduces Transformer Quantile Regression Neural Networks (TQRNNs), a novel data-driven solution for real-time machine failure prediction in manufacturing contexts. Our objective is to develop an advanced predictive maintenance model capable of accurately identifying machine system breakdowns. To do so, TQRNNs employ a two-step approach: (i) a modified quantile regression neural network to segment anomaly outliers while maintaining low time complexity, and (ii) a concatenated transformer network aimed at facilitating accurate classification even within a large timeframe of up to one hour. We have implemented our proposed pipeline in a real-world beverage manufacturing industry setting. Our findings demonstrate the model's effectiveness, achieving an accuracy rate of 70.84% with a 1-hour lead time for predicting machine breakdowns. Additionally, our analysis shows that using TQRNNs can increase high-quality production, improving product yield from 78.38% to 89.62%. We believe that predictive maintenance assumes a pivotal role in modern manufacturing, minimizing unplanned downtime, reducing repair costs, optimizing production efficiency, and ensuring operational stability. Its potential to generate substantial cost savings while enhancing sustainability and competitiveness underscores its importance in contemporary manufacturing practices.

Industrial Machines Health Prognosis using a Transformer-based Framework

TL;DR

Transformer Quantile Re-gression Neural Networks (TQRNNs) are introduced, a novel data-driven solution for real-time machine failure prediction in manufacturing contexts and can increase high-quality production, improving product yield from 78.38% to 89.62%.

Abstract

This article introduces Transformer Quantile Regression Neural Networks (TQRNNs), a novel data-driven solution for real-time machine failure prediction in manufacturing contexts. Our objective is to develop an advanced predictive maintenance model capable of accurately identifying machine system breakdowns. To do so, TQRNNs employ a two-step approach: (i) a modified quantile regression neural network to segment anomaly outliers while maintaining low time complexity, and (ii) a concatenated transformer network aimed at facilitating accurate classification even within a large timeframe of up to one hour. We have implemented our proposed pipeline in a real-world beverage manufacturing industry setting. Our findings demonstrate the model's effectiveness, achieving an accuracy rate of 70.84% with a 1-hour lead time for predicting machine breakdowns. Additionally, our analysis shows that using TQRNNs can increase high-quality production, improving product yield from 78.38% to 89.62%. We believe that predictive maintenance assumes a pivotal role in modern manufacturing, minimizing unplanned downtime, reducing repair costs, optimizing production efficiency, and ensuring operational stability. Its potential to generate substantial cost savings while enhancing sustainability and competitiveness underscores its importance in contemporary manufacturing practices.

Paper Structure

This paper contains 14 sections, 3 equations, 2 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. System Architecture: Proposed TQRNN pipeline
  4. Dataset
  5. Input data
  6. Quantile Regression Neural Networks
  7. Transformer network
  8. Experiment Results
  9. Evaluation Results and Metrics
  10. Ablation study
  11. Comparison against related state-of-the-art
  12. Computational cost analysis
  13. Machine downtime
  14. Conclusion

Figures (2)

  • Figure 1: Proposed Pipeline TQRNNs: Input data from 43 sensors is processed by a series of QRNNs, first with 10×43 networks, followed by 2×43 QRNNs. The final block employs a transformer network to synthesize outputs and predict machine anomalies.
  • Figure 2: Examples of damaged mechanical components, such as punch sleeves and swing levers, were predicted by our proposed approach before the damage occurred on two separate manufacturing machines.