From Corrective to Predictive Maintenance - A Review of Maintenance Approaches for the Power Industry

Weiping Ding (ae of Tnnls, Tfs, Tits, Jas, Tetci, Tiv, Tai, Inffus, Ins, Asoc, Neucom, EAAI,et al), Marek Moleda, Bozena Malysiak-Mrozek, Weiping Ding, Vaidy S. Sunderam, Dariusz MrozekSensors 2023 · DOI

TL;DR

This review paper provides a comprehensive overview of maintenance strategies in the power industry, tracing the evolution from traditional corrective and preventive methods to advanced predictive and prescriptive approaches. It highlights the growing importance of Industry 4.0 technologies in enabling more intelligent and efficient maintenance practices, crucial for ensuring the reliability and cost-effectiveness of power generation.

What Problem Does This Paper Address?

The power industry faces significant challenges in maintaining the reliability and efficiency of its complex infrastructure. Equipment failures, though rare, can lead to substantial economic losses and societal disruption. Traditional maintenance approaches, like corrective (fixing after failure) and preventive (scheduled maintenance), are often reactive or inefficient, leading to unnecessary costs and potential downtime. The paper addresses the need for more proactive and intelligent maintenance strategies that can leverage modern technologies to predict failures, optimize maintenance schedules, and ultimately enhance operational performance and safety.

Methods

The paper reviews a wide spectrum of maintenance techniques, categorizing them from traditional to advanced:

  • Corrective Maintenance: The most basic approach, involving repairs after a failure occurs.
  • Preventive Maintenance: Scheduled maintenance based on time intervals (e.g., MTBF, MTTF) or manufacturer recommendations.
  • Predictive Maintenance (PdM): This is a core focus, involving condition-based monitoring and analysis to predict failures before they happen. Techniques discussed include:
  • Vibration Monitoring: Analyzing vibrations to detect wear and anomalies.
  • Thermography: Using infrared imaging to detect temperature anomalies.
  • Oil Analysis: Examining oil for wear particles and contamination.
  • Acoustic Analysis: Monitoring acoustic signals for signs of deterioration.
  • Motor Current Analysis: Analyzing motor current signatures for electrical and mechanical faults.
  • Process Parameter Analysis: Monitoring key performance indicators.
  • Visual Inspection: Supported by technologies like AR and RFID.
  • Advanced Analytics: The paper delves into methods that support PdM, including:
  • Descriptive Analytics: Understanding past events.
  • Diagnostic Analytics: Identifying the "why" behind failures (e.g., Fault Tree Analysis, Root Cause Analysis).
  • Predictive Analytics: Forecasting future states and remaining useful life (RUL).
  • Prescriptive Analytics: Recommending optimal actions.
  • Industry 4.0 Enablers: The review also covers technologies that facilitate these advanced approaches, such as IIoT, Big Data, Cloud/Edge Computing, Augmented Reality, RFID, and 3D Printing.

Datasets

Not explicitly detailed in the paper as specific datasets used for training or evaluation. The paper discusses the need for large datasets of historical data for training machine learning models and the challenges associated with data acquisition, integration, and quality.

Key Insights and Ideas

  • Evolution of Maintenance: The paper clearly illustrates the shift from reactive (corrective) to proactive (preventive) and then to predictive and prescriptive maintenance, driven by technological advancements.
  • Industry 4.0 as a Catalyst: Industry 4.0 technologies are not just enablers but fundamental drivers for the widespread adoption of advanced maintenance strategies, particularly in the power sector.
  • Data is King: The effectiveness of predictive and prescriptive maintenance hinges on the availability, quality, and intelligent analysis of vast amounts of data from sensors and operational systems.
  • Holistic Approach Needed: The most effective maintenance strategies integrate data from various sources (operational technology, enterprise asset management) and leverage a combination of methods.
  • Power Industry Specifics: The unique demands of the power industry (high availability, safety, regulatory compliance) present specific challenges and opportunities for maintenance innovation.

Key Results

As a review paper, it does not present novel experimental results but rather synthesizes existing knowledge. The key "results" are the comprehensive categorization and comparison of various maintenance strategies and technologies, highlighting their applications, advantages, and limitations within the power industry context. It provides a roadmap for understanding the current state-of-the-art and potential future directions.

Limitations and Future Work

  • Challenges in Adoption: The paper acknowledges that despite the potential of advanced technologies, the energy industry cautiously benefits from them, often due to integration barriers with existing systems, the need for strong business cases, and the inherent risk aversion in critical infrastructure.
  • Uncertainty in Decision Support: A significant challenge remains in handling uncertainty within decision support systems, especially in high-stress operational environments.
  • Cybersecurity and Regulatory Compliance: The stringent safety and regulatory requirements, particularly for nuclear power plants, pose significant hurdles for implementing new technologies and may necessitate on-premise solutions over cloud-based ones.
  • Future Work: The authors suggest that further research is needed to develop solutions that better align with industry requirements, address the challenges of uncertainty, and leverage Industry 4.0 technologies to enhance protection against threats. There is also a call for more research and development tasks and proof-of-concept projects to bridge the gap between theoretical solutions and practical application.

Summary updated

BibTeX
@article{al2023from,
  author = {Weiping Ding (AE of TNNLS, TFS, TITS, JAS, TETCI, TIV, TAI, INFFUS, INS, ASOC, NEUCOM, EAAI,et al) and Marek Moleda and Bozena Malysiak-Mrozek and Weiping Ding and Vaidy S. Sunderam and Dariusz Mrozek},
  title = {{From Corrective to Predictive Maintenance - A Review of Maintenance Approaches for the Power Industry}},
  journal = {Sensors},
  year = {2023},
  doi = {10.3390/s23135970},
}
Figure