By: Assoc. Prof. Dr. Mohd Azree Idris
Tube leaks in crucial boiler components like waterwalls present significant challenges for power generators worldwide. These failures, ranging from minor cracks to entire section ruptures, result in extended outages, substantial revenue losses (with power outages costing the US economy $150 billion annually), and safety risks (15% of workplace accident deaths relate to boiler failures according to OSHA). As a rapidly developing middle income country, Malaysia has seen electricity demand grow at over 5% annually, with thermal plants reliant on coal and gas supplying 45% of capacity.
Unplanned outages caused by boiler issues can severely impact economic growth given demand runs at near full capacity. As Malaysia works towards its 2025 renewable energy target of 31% of total capacity, reliable fossil fuel plants remain crucial in the interim. Traditional manual inspection methods for these high-value assets have shown limited advancements over decades.
However, emerging technologies such as physics-based boiler modelling and Artificial Intelligence (AI) stand poised as potential game-changers. Automated monitoring solutions can detect anomalies and predict failures 35-45% before manual inspectors, while reducing unplanned downtime by up to 45% and maintenance costs by 35%.
Virtual Boiler Models – Simulating the Unattainable
Physics-based virtual boiler models replicate these physical assets computationally. These models integrate actual geometry, specifications, and historical operational data using multifidelity modelling, encompassing Computational Fluid Dynamics (CFD), heat transfer, and material science concepts. CFD, in particular, accurately maps fluid flow velocities, temperatures, and thermal stress gradients across boiler components.
Engineers can visualize and simulate scenarios that are impossible to recreate physically. This includes mapping thermal profiles under various operating conditions; simulating transients like cold/hot starts based on as-built details; predicting responses to part loads aligning with operational strategies; and conducting forensic analysis to replicate causes behind past tube failures.
Such simulation capabilities offer invaluable perspectives on performance response, stress thresholds, and failure modes, surpassing physical testing limitations.
Enabling Virtual Inspection
These physics-based models facilitate ‘virtual inspection’ by simulating hypothetical failure scenarios and evaluating responses digitally. For instance, software sensors can introduce pinhole cracks into specific waterwall tubes and assess impacts on surrounding areas in the model. Performance degradation quantified through simulations provides targeted early warning systems.
Integration for Comprehensive Insights
Modern simulation philosophy emphasizes integrating multiple modelling dimensions to enhance decision support. Combining thermal-hydraulics modelling with combustion optimization and welding residual stresses simulation provides holistic diagnostics. Expert sub-models are assimilated into an ensemble boiler model for comprehensive virtual inspection.
AI and Big Data for Continuous Monitoring
Integration of Internet-of-Things (IoT) technologies has inundated power plants with vast volumes of boiler instrumentation data captured at high frequencies. However, uncovering critical correlations from this data flood is challenging for operators. AI and Machine Learning (ML) techniques excel in recognizing complex multivariate interactions from sensor data through feature extraction, clustering analysis, and deep learning.
AI models enable round-the-clock monitoring of key parameters affecting boiler health, predicting anomalies, and enabling pre-emptive maintenance to prevent catastrophic outages, maximizing availability and reliability.
Enhancing Decision Support: CFD and AI Integration
Utilizing Computational Fluid Dynamics (CFD) and Artificial Intelligence (AI) forms the cornerstone of modern techniques for detecting and preventing boiler tube leaks within virtual boiler models. CFD, integrated into physics-based virtual boiler models, meticulously maps fluid flow velocities, temperatures, and thermal stress gradients across the boiler components, offering unparalleled insights into thermal profiles under diverse operational conditions. Engineers leverage this technology to simulate scenarios, such as transient cold/hot starts or responses to varying part loads, which are unfeasible to replicate physically.
Through this simulation, even conducting forensic analysis to understand past tube failures becomes feasible, transcending the limitations of physical testing. Moreover, the incorporation of AI into these models facilitates continuous monitoring by analyzing vast volumes of boiler instrumentation data captured at high frequencies. AI’s prowess in recognizing intricate patterns and correlations within this data enables the prediction of anomalies, thereby enabling preemptive maintenance to avert potential catastrophic outages, ensuring enhanced availability and reliability of the boiler systems.
The Future Ahead: TNB’s Role in Technological Advancements and Energy Transition
Tenaga Nasional Berhad (TNB), a prominent electricity utility company in Malaysia, is strategically positioned to spearhead technological advancements in the evolving energy landscape. Leveraging its extensive expertise in power generation and distribution, TNB envisions collaborating with esteemed technology partners like Universiti Tenaga Nasional (UNITEN) and TNB Research (TNBR) to implement cutting-edge solutions such as physics-based virtual boiler models and AI for predictive maintenance across its power plants.
Assisting TNB in Energy Transition and Co-firing Implementation, UNITEN and TNBR are equipped with robust research capabilities to bolster TNB’s initiatives aligned with the National Energy Transition Roadmap (NETR). As TNB navigates the complex implementation of co-firing coal and biomass for decarbonization, UNITEN’s specialized research in energy and sustainability combined with TNBR’s technological prowess positions them to analyze and predict the impacts of this transition on boiler performance. Specifically, they can address uncertainty effects on boiler tube leaks resulting from co-firing through advanced modeling and predictive analytics. This collaborative
effort ensures a seamless transition towards cleaner energy sources while optimizing the performance and reliability of TNB’s power generation infrastructure.
Conclusion
Malaysia’s power generation industry is embracing advanced technologies like physics-based virtual boiler models and AI, signifying a ground-breaking shift in operational practices. Industry players spearhead innovation efforts, aiming to revolutionize maintenance methods and efficiency. This forward-looking approach strengthens Malaysia’s energy sector capabilities while aligning with the National Energy Transition Roadmap (NETR). By integrating computational modelling and AI for predictive maintenance, operators aim for heightened reliability and sustainability. Such initiatives promise reduced downtime, bolstered infrastructure safety, and optimized plant performance. Investing in these emerging technologies pays dividends through improved forecasting, planning, and integration of renewable energy.
Ultimately, the adoption of such smart solutions paves the way for a more sustainable energy future in Malaysia by enabling better grid management and security as the country transitions toward clean energy.
…
The author is an Associate Professor at Department of Mechanical Engineering, College of Engineering, Universiti Tenaga Nasional (UNITEN), and may be reached at azree@uniten.edu.m
Leave a comment