Machine-Learning-Based Condition Assessment of Gas Turbines—A Review

Condition monitoring, diagnostics, and prognostics are key factors in today’s competitive industrial sector. Equipment digitalisation has increased the amount of available data throughout the industrial process, and the development of new and more advanced techniques has significantly improved the performance of industrial machines. This publication focuses on surveying the last decade of evolution of condition monitoring, diagnostic, and prognostic techniques using machine-learning (ML)-based models for the improvement of the operational performance of gas turbines. A comprehensive review of the literature led to a performance assessment of ML models and their applications to gas turbines, as well as a discussion of the major challenges and opportunities for the research on these kind of engines. This paper further concludes that the combination of the available information captured through the collectors and the ML techniques shows promising results in increasing the accuracy, robustness, precision, and generalisation of industrial gas turbine equipment.

Operational Scenarios of a Gas Turbine Using Exhaust Gas Recirculation for Carbon Capture

Finding viable economic solutions to significantly reduce or eliminate greenhouse gas emissions from energy and transportation products in the near future is paramount for the long-term survival of fossil fuel burning systems. One of which, the industrial gas turbine, has proven for decades to be a versatile energy system providing high efficiencies in combined heat and power applications melding well within existing infrastructure. Applying appropriate technology, the industrial gas turbine could be augmented to both sequester carbon and improve efficiency leveraging the full heating value of the fuel. The paper considers a more detailed operational assessment of a gas turbine using exhaust gas recirculation (EGR) to enable cost effective post combustion carbon sequestration and utilization. In this study, the effect of using EGR will be assessed at part load and throughout the operational envelope quantifying component and overall performance, detailed combustion characteristics, and maximizing the utilization of exhaust heat and sequestered carbon in various applications. This study will also attempt to quantify true carbon footprint of gas turbine installations and endeavor to understand the relative change of replacing the gas turbine with an all-electric alternative. Fundamentally, we are looking to see if there is a future to sustain and adapt this significant natural gas (NG) energy infrastructure to a net-zero carbon emissive future by 2050.

NUMERICAL INVESTIGATION OF UNSTEADY COMBUSTOR TURBINE INTERACTION FOR FLEXIBLE POWER GENERATION

In the present study low-frequency disturbances introduced by a periodic load variation have been simulated and superimposed to the inhomogeneous, unsteady flow entering a 3-stage, high-pressure industrial gas turbine fed by a can-type combustion chamber comprising 6 silo-burners. The effects of the unsteadiness realized at the combustor exit have been investigated by means of Detached Eddy Simulations, whereby a density-based solution approach with detailed thermodynamics has been employed. The periodic disturbances at the turbine inlet have been obtained by means of an artificially generated, unsteady field, resulting from a two-dimensional snapshot of the flow field at the combustor exit. Also, a combustor failure has been mimicked by reducing (respectively increasing) the mean temperature in some of the turbine inlet regions corresponding to the outlet of two burners. The propagation and amplitude changes of temperature fluctuations have been analyzed in the frequency domain. Tracking of the temperature fluctuations' maxima at the lowest frequencies revealed characteristic migration patterns indicating that the corresponding fluctuations persist with a non-negligible amplitude up to the last rows. A distinct footprint could also be observed at the same locations when a combustor failure was simulated, showing that, in principle, the early detection of combustor failures is indeed possible.