Advanced Gas Turbine cycles for high efficiency and sustainable future conventional generation

点击翻译按钮获取中文摘要

Gas Turbines (GTs) will figure prominently in complimenting the intermittent power generated by renewables, while varied fuel sources by 2050 are likely to include biofuels (the former a mixture of methane, carbon mono- and di-oxide and nitrogen - essentially low calorific value fuel) and perhaps shale gas and hydrogen. In meeting CO2 emissions targets, there will be a premium on designs that (i) have the highest fuel conversion efficiency and (ii) integrate with carbon capture and storage. Such designs include either humid air turbines (HAT) or schemes with extensive exhaust, or flue, gas recirculation together with the use of oxygen-enriched air. There is extensive techno-economic evaluation of these designs with no preferred 'winner' and it is likely that each will find extensive application. Thus, there will be a need to design combustion chambers to burn low calorific gases, with "oxidant" streams including up to 30% (w/w) of steam, pure oxygen or oxygen heavily diluted with Carbon dioxide. Such changes present formidable difficulties to flame stability and extinction. The design of low NOx combustion chambers has shown the value of computational fluid dynamics (CFD) in developing commercially viable designs and this trend will strengthen. Finally, the value of suitable sensors during development has proved its worth. This research identifies the gaps in existing physical understanding, CFD and optical sensors, to be addressed by "fundamental research", that need to be filled so that step change GT technologies can be developed by industry. This proposal will develop tools and understanding as follows:(i) On-line, near real time optical sensor to measure the 'Wobbe' index of fuel entering the gas turbine, since fast knowledge of the calorific value of highly variable bio- fuels is important for control of future GTs.(ii) Flame stability and extinction is associated with the existence of a critical 'rate of stretch' and the largest laminar flame speed that the flame can experience due to the aerodynamic flow field of the combustors. Designers, using CFD for flow prediction in combustion chambers, need to know these critical values for the range of fuels and oxidants, which will be in use up to 2050. Thus, this proposal will obtain measurements of these values in premixed and non-premixed flames as a function of preheat and pressure and analyse the process of flame extinction in laboratory and pilot scale model combustors using, amongst other instruments, detection of CO and formaldehyde by planar laser induced fluorescence.(iii) Low NOx emissions require the fuel to be well premixed and it is useful for development engineers to have access to an instrument, which can measure local fuel/air ratio on test stands. Building on previous successful development of an instrument based on natural chemiluminescent emissions from a flame, there will be an evaluation of its calibration as a function of pressure and humidity, the latter in the context of a HAT gas turbine design.(iv) Thermoacoustic instability is a destructive high intensity 'limit cycle', which is either avoided operationally or designs are improved largely by cut and try methods. Until recently, the transition to this limit cycle and the limit cycle itself were characterised by frequency and phase spectral analysis. Our recent work has shown that non-linear time series analysis reveals that transition to high amplitude oscillations retains a structure as determined by chaos theory. We will use this form of analysis to identify the fluid mechanical structures responsible for this behaviour, with the aim of devising methods to at least warn gas turbine operators of impending thermoacoustic instability.(v) The best available LES CFD methods will be evaluated using the measurements in the counterflow and model combustor geometries. There will also be direct assessment, through the measurements, of the 'sub-grid' contribution of LES methodology to calculations