Use this URL to cite or link to this record in EThOS:
Title: Numerical simulations of temperature mapping in industrial combustion environments
Author: Wood, Michael Philip
Awarding Body: University of Manchester
Current Institution: University of Manchester
Date of Award: 2013
Availability of Full Text:
Access from EThOS:
Access from Institution:
This thesis presents the results from a set of numerical experiments of two-dimensional gas temperature imaging using laser absorption spectroscopy inside a turbofan engine. This measurement environment is characterised by temperatures of 2000 K, pressures of 45 bar, and extremely limited access for the installation of measurement hardware, which renders invasive measurement (thermocouple arrays) or direct imaging (PLIF or pyrometry) methods unviable. An alternative approach is indirect imaging of the temperature, whereby the transmittance of a near-infrared laser light through the gas is measured and used to make inferences about the properties of the gas along the beam; specifically, its temperature, pressure, and molecular constitution. The frequency of the light is chosen to interrogate particular molecular transitions of a target species—water—in such a way that the fraction of light measured at the detector depends on the temperature of the gas through which it has passed. This is an established measurement technique known as tuneable diode laser absorption spectroscopy (TDLAS), but it is possible to extend this method to two dimensions if the transmittance measurements are made over set of coplanar beams that transect the measurement region. Using the principles of tomographic inversion, it becomes possible to image not only the two-dimensional temperature distribution within a gas, but also the pressure and molecular species concentration distributions. In this thesis, extensive numerical simulations are used to critically evaluate this approach when applied to the particular case of the turbine engine, and a new methodology is developed for use in this environment which opens up—for the first time, to the best of the author’s knowledge—the possibility of tomographic reconstruction of a gas pressure. This is challenging because the gas pressure has a strong influence on not only the width of absorption lines, but of their positions on the spectrum, with each line being affected in a different way. To overcome and eventually exploit this dependence, a robust approach which the author terms the spectral fitting approach is developed and tested against the two main existing methods found in the literature: integrated absorbance and peak absorption reconstructions. The spectral fitting approach was found to outperform both methods not only in the high-pressure regime, but throughout the tested pressure range (1-70 bar).The numerical tests were also applied to more realistic measurement environments, including annular measurement regions (modelling the opaque central driveshaft of a turbine engine) with non-uniform molecular species concentrations and gas pressures. In these investigations, the temperature was reconstructed with a relative root-mean-squared error of 2.47%. This demonstrates the theoretical feasibility of tomographic reconstructions of gas temperature in the turbine environment. Numerical optimisation of the methodology is also addressed. The geometric arrangement of beams through the measurement region is investigated with a view to maximise the quality of the reconstructed image, and a new design rule is analytically derived and then applied to generate a set of viable beam arrangements that perform competitively when compared to more conventional regular arrangements. The selection of laser frequencies is also optimised in the specific case of high-pressure spectroscopy, and two near-infrared transitions are suggested as a possible candidate pair for experimental verification.
Supervisor: Ozanyan, Krikor Sponsor: Rolls-Royce plc ; EPSRC
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available