GCN-Based LSTM Autoencoder with Self-Attention for Bearing Fault Diagnosis

Sensors, Journal Year: 2024, Volume and Issue: 24(15), P. 4855 - 4855

Published: July 26, 2024

The manufacturing industry has been operating within a constantly evolving technological environment, underscoring the importance of maintaining efficiency and reliability processes. Motor-related failures, especially bearing defects, are common serious issues in Bearings provide accurate smooth movements play essential roles mechanical equipment with shafts. Given their importance, failure diagnosis extensively studied. However, imbalance data complexity time series make challenging. Conventional AI models (convolutional neural networks (CNNs), long short-term memory (LSTM), support vector machine (SVM), extreme gradient boosting (XGBoost)) face limitations diagnosing such failures. To address this problem, paper proposes model using graph convolution network (GCN)-based LSTM autoencoder self-attention. was trained on extracted from Case Western Reserve University (CWRU) dataset fault simulator testbed. proposed achieved 97.3% accuracy CWRU 99.9% dataset.

Language: Английский

---

Anomaly Detection in Gas Turbines Using Outlet Energy Analysis with Cluster-Based Matrix Profile

Energies, Journal Year: 2024, Volume and Issue: 17(3), P. 653 - 653

Published: Jan. 30, 2024

Gas turbines play a key role in generating power. It is really important that they work efficiently, safely, and reliably. However, their performance can be adversely affected by factors such as component wear, vibrations, temperature fluctuations, often leading to abnormal patterns indicative of potential failures. As result, anomaly detection has become an area active research. Matrix Profile (MP) methods have emerged promising solution for identifying significant deviations time series data from normal operational patterns. While most existing MP focus on vibration analysis gas turbines, this paper introduces novel approach using the outlet power signal. This modified approach, termed Cluster-based (CMP) analysis, facilitates identification subsequent within turbine engine system. Significantly, CMP not only accelerates processing speed, but also provides user-friendly support information operators. The experimental results real-world demonstrate effectiveness our early anomalies system

Language: Английский

---

MEMS vibration sensor-based edge AI for machinery fault prediction: feasibility study using a petrochemical plant process simulation facility

Journal of Inclusion Phenomena and Macrocyclic Chemistry, Journal Year: 2025, Volume and Issue: unknown

Published: Jan. 22, 2025

Language: Английский

---

Generative adversarial message passing-based anomaly detection

Journal of King Saud University - Computer and Information Sciences, Journal Year: 2025, Volume and Issue: 37(1-2)

Published: March 18, 2025

Language: Английский

---

Anomaly Detection in Bearing Temperature Data of Industrial Centrifuge Device Using One-Class SVM for Predictive Maintenance in the Mining Sector

Lecture notes in networks and systems, Journal Year: 2025, Volume and Issue: unknown, P. 594 - 607

Published: Jan. 1, 2025

Language: Английский

---

Unsupervised Anomaly Detection using Vibration Signals for Milling Processes

Procedia CIRP, Journal Year: 2025, Volume and Issue: 133, P. 710 - 715

Published: Jan. 1, 2025

Language: Английский

---

A review of artificial intelligence techniques in blast field shockwave pressure testing

Journal of Measurements in Engineering, Journal Year: 2025, Volume and Issue: unknown

Published: May 11, 2025

Ammunition explosion shockwave pressure is an important war technology indicator to evaluate the explosive damage power of ammunition, and it great significance accurately obtain law distribution guide design ammunition. With rapid development artificial intelligence technology, researchers have applied technologies such as neural networks, machine learning, deep big data large models shock wave testing, field reconstruction assessment, greatly improving working bandwidth measurement system, accuracy calculation results efficiency assessment under current test conditions. This study summarizes application above methods in relevant achievements, discusses shortcomings research, puts forward problems that should be analyzed follow-up points out direction research.

Language: Английский

---

A Machine Learning Operations Platform for Streamlined Model Serving in Industry 5.0

Published: May 6, 2024

Language: Английский

---

An Experiment on Anomaly Detection for Fault Vibration Signals Using Autoencoder-Based N-Segmentation Algorithm

PHM Society European Conference, Journal Year: 2024, Volume and Issue: 8(1), P. 8 - 8

Published: June 27, 2024

Most manufacturing facilities driven by motors generate vibration and noise representing critical symptoms against facility malfunctioning conditions in the industry. Due to difficulty of obtaining abnormal data from sites, many prior researchers who have studied predicting faults adopted unsupervised learning-based anomaly detection approaches. Although these approaches a strength requiring only on normal behaviors, it is not clear that anomalies detected an model are due real component faults. Also, performance likely change according diverse given facility. In this paper, we took experiment with fault simulator measure one-dimensional convolutional autoencoder different conditions. experiment, used four conditions: imbalance, misalignment, looseness, bearing faults, which most frequently occurring failures rotating machineries. Data were gathered IEPE(Integrated Electronics Piezo-Electric) type sensor. We proposed N-Segmentation algorithm performs segmented frequency region corresponding for better performance. conclusion, showed about 15 times rate than applying it.

Language: Английский

---

An Echo State Network-Based Light Framework for Online Anomaly Detection: An Approach to Using AI at the Edge

Machines, Journal Year: 2024, Volume and Issue: 12(10), P. 743 - 743

Published: Oct. 21, 2024

Production efficiency is used to determine the best conditions for manufacturing goods at lowest possible unit cost. When achieved, production leads increased revenues manufacturer, enhanced employee safety, and a satisfied customer base. not only measures amount of resources that are needed but also considers productivity levels state lines. In this context, online anomaly detection (AD) an important tool maintaining reliability ecosystem. With advancements in artificial intelligence growing significance identifying mitigating anomalies across different fields, approaches based on neural networks facilitate recognition intricate types by taking into account both temporal contextual attributes. paper, lightweight framework Echo State Network (ESN) model running edge introduced AD. Compared other AD methods, such as Long Short-Term Memory (LSTM), it achieves superior precision, accuracy, recall metrics while reducing training time, CO2 emissions, need high computational resources. The preliminary evaluation proposed solution was conducted using low-resource computing device real machine through Industrial Internet Things (IIoT) smart meter module. test provided Italian company SIFIM Srl, which manufactures filter mats industrial kitchens. Experimental results demonstrate feasibility developing method with ESN-based reaching 85% compared 80.88% LSTM-based model. Furthermore, requires minimal hardware resources, time 9.5 s 2.100

Language: Английский

---