Development of Real-Time Smart Data Analytic Tools for Monitoring and Optimum Operation of MGT Systems

Abstract

In the evolving energy landscape, a shift away from traditional centralized power models is underway. Distributed energy generation (DEG) takes the spotlight, enabling consumers to utilize a tailored mix of energy sources. Micro gas turbines (MGTs) emerge as key players, providing dispatchable power to seamlessly address renewable source intermittency. Aligned with global energy policies emphasizing renewables and efficiency, MGTs contribute significantly to sustainability goals. This study aims to actively advance power generation technology towards higher efficiency and environmental responsibility, supporting the vision for a cleaner and more resilient energy future. The focus centers on enhancing the fuel versatility of MGTs and optimizing their integration within distributed energy systems. Aligned with the visionary goals of the NextMGT project, this endeavor focuses on advancing MGT technology for high efficiency, low emissions, and enhanced fuel flexibility. The journey begins with an exploration into optimizing an MGT for efficient hydrogen operation — a clean fuel and potential storage solution for surplus renewable power. Despite substantial progress in the laboratory and theoretical realms, the research spotlights a critical gap: the absence of reported operational instances of MGTs running with hydrogen. This underscores the imperative to bridge the divide between theoretical prowess and real-world applications, a recurring theme in the thesis. Concurrently, the research navigates the intricate integration of MGTs into DEG, particularly those fueled by hydrogen. Addressing the challenges of integration and optimization with renewable systems, artificial intelligence (AI) based on real-world data is employed to enhance microgrid performance. Undertaking the mission to create a functional hydrogen-fueled MGT, the research confronts challenges such as combustion stability and emissions control. Through targeted modifications in combustor design and operational adjustments, the thesis emphasizes real-world testing, highlighting the crucial need for practical implementations. Notably, the outcome is an MGT demonstrating fuel flexibility with various methane and hydrogen combinations, capable of running on pure hydrogen, all while maintaining NOx emissions below the permitted values. A significant step in the research narrative involves adopting a dual-modeling approach—utilizing both physics-based and datadriven models. The physics-based models, also known as whitebox models, rooted in physics for theoretical understanding, exhibit adaptability to diverse operational scenarios, aligning with steady-state and transient responses. This model plays a crucial role during the developmental phase of the MGT for hydrogen and hydrogen-blended methane, assessing its operation in different scenarios. In addressing the integration of MGT within a microgrid, a datadriven or black-box modeling approach is employed. These blackbox models, driven by empirical data, incorporating artificial neural networks (ANNs) and recurrent neural networks (CNNs), emerge as a robust framework for MGT modeling. The versatility of the method extends beyond MGTs, laying the groundwork for advancements in various renewable energy contexts. In a dedicated chapter, the study delves into microgrid integration and optimization. Here, a smart management system coordinates interactions among wind turbines, an MGT, and an electrolyzer. The optimizer navigates the complex terrain of economic gains and environmental sustainability. The findings emphasize the practical application of a smart management system in optimizing microgrid operations for economic efficiency, demonstrating the relevance of the research insights. In response to Norway’s imperative to curtail emissions from offshore oil and gas operations, the research broadens its focus to optimize gas turbine operations within integrated systems. The research demonstrates adaptability by transitioning from onshore microgrids with MGTs to offshore scenarios with larger gas turbines, highlighting the transformative and generalizable capacity of methodologies and insights. The optimization of offshore microgrids results in considerable cost and emission reductions. The hybrid optimization approach, efficiently utilizing genetic algorithms alongside rapid database searches, enhances efficiency without an excessive demand for computing resources. Throughout this project, the strategic adoption of an infrastructural approach has been pivotal in the development of all models and programs. This deliberate choice ensured the effectiveness of transformative insights and a seamless adaptability and expandability of the work. Integrating hydrogen-fueled MGTs with advanced AI management tools moves beyond theory; it represents a practical step toward achieving sustainable development goals. From onshore microgrids to offshore scenarios, the research illustrates a commitment to real-world applications. Its impact extends beyond theoretical contributions, actively shaping a more sustainable, resilient, and eco-friendly energy future. Additionally, by identifying areas for future research, this thesis lays the foundation for ongoing advancements in sustainable energy solutions.