International Journal of Prognostics and Health Management

Joint Prescriptive Maintenance and Production Planning and Control Process Simulation for Extrusion System

Kevin Wesendrup — 2024-06-26

Production planning and control (PPC) is the mainstay of every manufacturer and ensures flawless production processes. However, PPC is jeopardized by breakdowns that can only be tackled with appropriate maintenance. In the past, static strategies, such as reactive and scheduled maintenance, have been used. Yet, with growing system complexity, Industry 4.0, and abundant sensor data, dynamic strategies through PHM have emerged. The most advanced maintenance strategy is prescriptive maintenance (PxM), which allows manufacturers not only to predict failures but also to establish condition-based production plans and controls. To this end, our study explores the integration of PxM with PPC. First, we propose a fault prediction model based on health indicators and future loads of a single-machine system. The proposed fault prediction is integrated into a joint PxM and PPC simulation model that compares the makespan of three joint PxM and PPC strategies inter se and versus reactive and scheduled maintenance. A simulation study using industrial data from an extrusion process evaluates the different strategies across different time horizons (one month to a year). The findings indicate that joint PxM and PPC outperform other strategies, providing significant time savings over traditional methods. Further, a sensitivity analysis is conducted to assess the robustness of the PxM strategies under varying levels of measurement noise, revealing potential challenges under high noise conditions. The study contributes to the field of PHM by providing insights into the effectiveness of joint PxM and PPC strategies and offering a comprehensive analysis of their performance under different conditions.

A Comprehensive Review of Machine Learning Techniques for Condition-Based Maintenance

Tyler Ward — 2024-06-26

While most industrial maintenance strategies are centered on optimizing machine runtime and cost reduction, the condition-based maintenance (CBM) strategy distinguishes itself from others in its use of real-time operational data from machines to help engineers make informed decisions. The introduction of machine learning (ML) into a CBM strategy can increase its effectiveness, enabling more accurate predictions and making the decision-making process more efficient. In this review paper, we seek to provide a comprehensive overview of the role ML plays in modern CBM systems, beginning by outlining the core concepts and historical development of CBM and briefly introducing various ML techniques being employed in industry today. We then review numerous real-world cases where ML-based CBM systems have been implemented and discuss some of the technological, human, and ethical challenges faced by organizations seeking to integrate sophisticated ML models into existing CBM systems. We end by highlighting some of the current limitations of ML-based CBM systems, paving the way for a discussion on emerging trends and future research directions in this area.

A Novel Prognostics and Health Management Framework to Extract System Health Requirments in the Oil and Gas Industry

Khalid Alfahdi — 2024-06-26

The paramountcy of Prognostics and Health Management (PHM) within the oil and gas sector is instrumental in ensuring safety, reliability, and economic efficiency by optimizing system availability. However, a prevalent industrial challenge is the lack of a comprehensive identification of health management requirements from actual operational situations. This study introduces an innovative Prognostics and Health Management Framework (PHMF), encompassing a methodical procedure to discern health management necessities systematically. The PHMF consolidates structured causal factors, foundational elements of functional failure, and the antecedents of unplanned downtime, which collectively inform the PHM strategy.
This framework offers an integrated view of multiple dimensions of system health, facilitating accurate portrayal and proactive monitoring. It particularly underscores a reverse engineering approach to scrutinize the root causes of system failures and unexpected operational halts. To validate the practicality and efficacy of the PHMF, it has been applied to a real-world scenario: a lubrication oil system within a gas turbine equipment, thereby elucidating the specific PHM strategy prerequisites.

Telemetry Monitoring System with Features Explaining Anomalies Based on Mahalanobis Distance

Shun Katsube — 2024-06-26

Because satellites cannot be repaired once launched, operators must detect anomalies early and prevent failures before they occur. Thus, satellite telemetry monitoring systems need to alert operators of anomalies and provide them with useful information to deal with these anomalies. However, traditional knowledge-based monitoring systems have the challenges of difficulty in building comprehensive models and a high dependency on experts. In recent years, data-driven approaches have been actively studied with the development of machine learning algorithms. These approaches solve the challenges of knowledge-based methods; however, they are often less capable of explaining anomalies than knowledge-based methods. In this study, we propose the new telemetry monitoring system with feature engineering to explain anomalies. The proposed method realizes identifiability of anomaly types and unusual telemetry by designing features based on moving averages, telemetry periods, waveform differences, and the Mahalanobis distance. We applied the proposed features to artificial and practical abnormal datasets and evaluated their usefulness. The results showed that the proposed method is capable of identifying trend, periodic, and waveform anomalies, specifying the telemetry in which the anomaly occurred and providing the information to operators.