A Fundamental Framework for the Robust Stabilization of Gas Turbine Combustion Instability

该项目解决了主动控制燃烧系统中不稳定压力振荡的挑战。虽然这项研究的重点是燃气轮机的应用，但它可能对其他燃烧装置如工业过程炉和锅炉产生影响。对于所有这些系统，社会都在大力推动提高燃烧效率，减少氮氧化物等有害排放，并适应消费者电力需求的变化。该项目的工作将推动燃烧系统稳定性、可靠性、耐久性和排放的改善。这些改进对于为电网提供备用发电能力的固定式燃气轮机尤其有价值。要实现这些高效率和低排放的目标，通常需要在接近燃烧不稳定的情况下运行。不稳定燃烧具有破坏性和物理破坏性，当热释放动力学、流体力学和声波传播相互增强时，就会触发不稳定燃烧。有丰富的现有文献提供了对燃烧不稳定性背后的物理学的见解，也表明主动控制可以减轻这种不稳定性。然而，许多控制策略过于简化了建模动力学，并且没有利用先进的控制策略，因此阻止了主动燃烧稳定性控制具有与先前实验室研究成功相称的实际/工业影响。该项目将产生用于数学描述燃烧现象的新公式，以及通过实验验证控制燃烧过程中发生的不稳定性的创新方法。该研究计划包括本科生和研究生通过独特的多学科培训机会的参与。这项工作将发展成一个本科水平的实验室实验，为大量工程本科学生提供将非线性控制理论应用于关键电力技术的机会。这项工作将为稳态和瞬态运行期间的鲁棒、多变量、组合被动/主动燃烧稳定提供一个新的基本框架，并使用灵活的实验室燃烧装置进行验证。该框架将利用非线性模型简化来开发面向控制的燃烧不稳定性简化模型，包括更精确的物理，并通过实验验证。此外，它将分析燃烧建模和参数化不确定性对基于模型的燃烧控制的性能、稳定性和鲁棒性的影响程度。该框架利用鲁棒的多变量控制理论来设计能够利用多个传感器和执行器的燃烧控制算法。非线性模型预测控制不仅可以稳定特定工况下的燃烧动力学，还可以稳定不同工况间的瞬态切换。最后，该研究将利用设计/控制组合优化来开发有利于主动稳定的被动燃烧室设计。这项工作的更广泛的影响将达到技术社区，行业和研究生和本科生水平的学生。在这项工作中取得的技术进步将证明鲁棒多变量控制理论在不稳定性控制和设计更好的燃烧室系统方面的应用。改进燃烧室设计的潜力可以转化为工业用途和燃气轮机燃烧室设计。

This project addresses the challenge of actively controlling unstable pressure oscillations occurring in combustion systems. While the research focuses on gas turbines applications it can have impact on other combustion devices like industrial process furnaces and boilers. For all of these systems, there is a strong societal push to improve combustion efficiency, reduce harmful emissions such as nitrogen oxides, and accommodate variations in consumer power demand. The work of this project will drive improvements in combustion system stability, reliability, durability, and emissions. These improvements will be especially valuable for the stationary gas turbines providing reserve generation capacity to the power grid. Achieving these goals of high efficiency and low emissions often involves operating closer to combustion instability. Unstable combustion can be disruptive and physically damaging, and is triggered when the dynamics of heat release, fluid mechanics, and acoustic wave propagation are mutually reinforcing. There is a rich existing literature that provides insights into the physics behind combustion instability, and also shows that active control can mitigate this instability. However, many of the control strategies over-simplify the modeled dynamics and do not take advantage of advanced control strategies and consequently prevent active combustion stability control from having a practical/industrial impact commensurate with previous laboratory research successes. This project will lead to novel formulations for mathematically describing the combustion phenomena and innovative ways to control instabilities occurring during the combustion process with experimental validation. The research plan includes the participation of both undergraduate and graduate students through a unique multi-disciplinary training opportunity. An undergraduate-level laboratory experiment will be developed from this work, providing a large number of undergraduate engineering students with the opportunity to apply nonlinear control theory to critical power technologies.This work will furnish a novel fundamental framework for robust, multivariable, combined passive/active combustion stabilization during both steady-state and transient operation, and validate it using a flexible laboratory combustion rig. The framework will utilize nonlinear model reduction to develop control-oriented, reduced models of combustion instability that include more accurate physics, validated using experiments. Further, it will analyze the degree to which combustion modeling and parameterization uncertainties can penalize the performance, stability, and robustness of model-based combustion control. The framework utilizes robust, multivariable control theory to design combustion control algorithms capable of exploiting multiple sensors and actuators. Nonlinear model predictive control will stabilize combustion dynamics not just around a particular operating condition, but also during transient switching between operating conditions. Finally, the research will exploit combined design/control optimization to develop passive combustor designs inherently conducive to active stabilization. The broader impacts of this work will reach the technical community, industry, and students at both the graduate and undergraduate level. The technical progress made in this work will demonstrate the use of robust, multivariable control theory for both instability control and design of better combustor systems. The potential for improving combustor design can be translated to industrial use and gas turbine combustor design.