Turbines for flexible power plant operation

Thesis

More than 70 percent of power in the world is generated by gas and steam turbines. Whilst renewables are desirable and continue to provide a growing contribution to the energy portfolio, turbine technology is projected to play an important role in power generation to 2060. The increased capacity from renewables is imposing new challenges and operational requirements on conventional power systems. Traditional designs, optimised for peak performance at constant load, must be adapted for load-levelling flexible operation, accepting more frequent and demanding start-stop cycles.

These challenging operating conditions are driving the need for advanced online diagnostic and monitoring tools. The harsh internal conditions of power turbines mean limited access and data is available to measure the thermal behaviour directly. These restrictions force the need for fast simulation methods to remotely assess the turbine condition. Detail knowledge of the thermal profile, and associated clearances, is essential for optimising transient control without compromising reliability.

Numerical methods for the fast simulation of thermal behaviour in 1D and 3D have been evaluated. New solution methods are presented to support fast 1D modelling of transient heat flow and allow the accuracy of traditional methods to be quantified. A novel hybrid methodology is developed, enabling data from multiple fidelity sources to be combined, thereby bridging the limitations in the independent analyses. New concepts in hybrid data transfer and thermal network modelling are demonstrated in the case of analysing temperatures in a Mitsubishi Heavy Industries steam turbine. A new multi-fidelity thermal analysis software is developed utilising plant measurements, thermal networks, neural networks and simulation data.

Validation cases and future developments are explored, highlighting the potential of the hybrid modelling concept. Demonstrated in the case of thermal analysis for flexible operation of power turbine, the hybrid methods offer new and exciting opportunities for rapid design and online diagnostic monitoring.

Authors: Baker, M

• DOI: 10.5287/ora-r8n9mg50r
• Type of award: DPhil
• Level of award: Doctoral
• Awarding institution: University of Oxford
• Language: English
• Keywords: impulse response; thermal network; thin film gauge; neural network; heat transfer
• Subjects: Thin films, Multilayered; Thermal analysis; Python (Computer program language); Heat equation; Neural networks (Computer science)