Tunable diode laser spectroscopy (TDLS) is a widely used technique for the measurement of gas species and offers in-situ operation, accuracy and faster response time compared to other optical and non-optical gas sensing techniques.The work in this thesis focusses on the measurement of CO2 in the harsh environment of a gas turbine engine (GTE). The work is part of a much larger initiative called Fibre Laser Imaging of gas Turbine Exhaust Species (FLITES) aimed at obtaining concentration distributions of gas species such as CO2 and NO, unburnt hydrocarbons, and soot in a gas turbine exhaust plume using optical tomography. In the FLITES system, a thulium doped fibre amplifier (TDFA) is used to boost the optical power output from a 2 mW, 1997 nm, multi-quantum well distributed feedback (DFB-MQW) laser to feed 126 measurement channels arranged in dodecagon geometry for optical tomography. Hence, agile TDLS techniques need to be developed which can be scaled up to the multi-channel measurement system.Attributed by the interference from noise in the measurement environment of a GTE, phase sensitive detection using a lock-in amplifier (LIA) has to be employed where an additional current modulation is applied to the DFB laser, creating an instantaneous intensity modulated output and a delayed wavelength modulation (WM) output. This technique falls under a metrology branch known as wavelength modulation spectroscopy (WMS).The unknown measurement conditions expected in a GTE engine necessitates the use of calibration-free WMS techniques for the simultaneous measurement of gas concentration and temperature. Calibration-free techniques in WMS have been developed at the Centre for Microsystems and Photonics (CMP) of Strathclyde University. These are known as the phasor decomposition method (PDM) and the residual amplitude modulation (RAM) technique. They employ the signals obtained using the first harmonic demodulation of the WMS signals, followed by post processing to recover the gas absorption line shape. It was known in the CMP group that the accuracy of these techniques was limited by the variation in the laser modulation parameters such as the phase of the wavelength modulation relative to the intensity modulation (WM-IM phase lag) and the wavelength modulation amplitude across the laser current scan.The solutions to two problems are addressed in this thesis, viz. the implementation of correction procedures to account for the variation in the laser modulation parameters across the current scan and the need for a calibration-free technique for the measurement of CO2 in a GTE exhaust plume scalable to a multi-channel measurement system.Accurate measurements of the wavelength modulation parameters were made across the current scan and correction algorithms were implemented to compensate for its effects on the recovered gas absorption line shape.The gas spectral parameters were measured in the lab for the R48 absorption line of CO2 near 1997.2 nm at the higher temperatures (up to 500°C) expected in a GTE exhaust plume, using a heated gas cell. A Fourier expansion model was developed for the WMS signals which employ the measured laser modulation and gas spectral parameters. 1f normalised 2f WMS technique was chosen as the calibration-free measurement approach due to the advantages of cancellation of the transmission fluctuations as well as signal normalisation. The 2f/1f measurement technique was validated in the lab at higher temperatures for the simultaneous recovery of the CO2 concentration and temperature with an accuracy of 3.39 % and 3.72 %, respectively. Subsequently, field campaigns were conducted at the Rolls-Royce test facility at East Kilbride, yielding concentration and temperature values having good correlation to the engine operating conditions such as the throttle and core temperature.Multi-channel tomographic measurements were conducted on the test phantoms at INTA, Madrid, using TFLAS-WMS (tunable fibre laser absorption spectroscopy). Accurate concentration images could be recovered using tomographic reconstruction algorithms.
|Date of Award||19 Jan 2017|
- University Of Strathclyde
|Sponsors||University of Strathclyde & EPSRC (Engineering and Physical Sciences Research Council)|
|Supervisor||Michael Lengden (Supervisor) & Walter Johnstone (Supervisor)|