High precision temperature measurements for reacting flows

点击翻译按钮获取中文摘要

The effective and fast design of low emission gas turbines depends critically on the ability of engineers to make accurate and precise predictions of gas temperatures within the combustion chamber. This project aims to produce instantaneous temperature measurements of the highest accuracy and precision ever in model and industrial scale combustors. These precision measurements aim not only provide the basis for validation of models by industrial and academic users, but also to create a path for development of a lower cost, high precision thermometry technique for deployment in realistic combustors. The two key factors governing the design of continuous flow combustors are maintaining low emissions - particularly nitric oxides - and keeping the system away from thermoacoustic instabilities. The spatial and statistical distribution of burned gas temperatures is the single most important factor governing the formation of nitric oxide (NO): a local change of 50 K can lead to a change of 70% in thermal NO formation rates at typical combustion temperatures. Validation of emission prediction models is hemmed by the lack of availability of statistical and spatial information on temperatures. Thermoacoustic instabilities are created by a feedback effect in which acoustic waves generated by the unsteady acceleration of the flow during combustion in a confined environment lead to further unsteadiness in heat release. Two factors associated with the flame are important: the response of the flame to acoustic perturbation, and the generation of temperature non-uniformities (called entropy spots): the former leads directly to density fluctuations and acoustic waves, and the latter couple the boundary conditions to reflect as pressure waves. The identification of the origin of combustion instabilities is complex, as several factors can contribute fluctuations, yet usually only pressure information is available, sometimes aided by relative total heat release fluctuations via chemiluminescence. Nevertheless, statistical measurements of temperatures in either model or industrial scale gas turbine flames are relatively uncommon, because of difficulties with physical probes or optical methods relying on calibration of signal amplitudes. The proposed measurements do not rely on amplitudes, but on the measurement of signal frequency, which can be made significantly more precisely (down to errors of 0.2%) than comparable techniques. Furthermore, the present measurements will enable the direct simultaneous measurements of NO and temperature with a single laser, thus creating a unique statistical database for model validation. Finally, the technique will enable for the measurement of temperature fluctuations through a nozzle at very high precision, which has not been done previously. The high precision measurements will have a direct impact on assessing the quality of model predictions for NO and instabilities, and when translated into design codes, into the design of cleaner and more stable power and propulsion systems.