The development of a combustion temperature standard for the calibration of optical diagnostic techniques
This thesis describes the development and evaluation of a high-temperature combustion standard. This comprises a McKenna burner premixed flame, together with a full assessment of its temperature, stability and reproducibility. I have evaluated three techniques for high-accuracy flame thermometry: Modulated Emission in Gases (MEG), Rayleigh scattering thermometry and photo-acoustic thermometry. MEG: Analysis shows that MEG is not usable in this application because the sharp spectral features of the absorption coefficient of gases are represented within MEG theory as an average absorption coefficient over the optical detection bandwidth. A secondary difficulty arises from the lack of high power lasers operating at wavelengths that coincides with molecular absorption lines in the hot gas. Rayleigh Scattering: Applying corrections for the temperature-dependence of the scattering cross-section, it has been possible to determine the temperature of the combustion standard with an uncertainty of approximately 1%. The temperature dependence of the scattering cross-section arises from changes in the mean molecular polarisability and anisotropy and can amount to 2% between flame and room temperatures. Using a pulse Nd-YAG laser operating at 532 nm and high linearity silicon detectors, the Rayleigh scattering experimental system has been optimised. Temperatures measured over a three-month interval are shown to be reproducible to better than 0.4% demonstrating the suitability of the McKenna burner as a combustion standard. Photo-Acoustic: By measuring the transit time of a spark-induced sound wave past two parallel probe beams, the temperature has been determined with an uncertainty of approximate 1%. Flame temperatures measured by the photo-acoustic and Rayleigh scattering thermometry system show good agreement. For high airflow rates the agreement is better than 1% of temperature, but for low airflow rates, photo-acoustic temperatures are approximately 3.6% higher than the Rayleigh temperatures. Further work is needed to understand this discrepancy.