OCR Text |
Show 4.4 Non-Intrusive Flame Temperature Measurements Knowledge of the flame temperature is important for studying the combustion efficiency, pollutant formation, heat transfer and flame structure. In addition, temperature information is required for validation of flames mathematical modeling. Numerous techniques both probe and non-intrusive optical methods including laser techniques, have been developed over the years for measuring flame temperatures. Laser techniques such as C A R S (Coherent Anti-Stokes Raman Spectroscopy) have been demonstrated successfully on both laboratory and industrial scale air-fuel flames. C A R S offers the adv antages of high temporal and spatial. However, extension to oxy-fuel flames poses special problems due to the higher temperatures and steeper gradients, e.g., the adiabatic flame temperature for a stoichiometric CH4-O; mixture is 820 K higher than when using air in place of oxygen. A number of problems were encountered in applying C A R S to industrial scale oxy-fuel flames [3,22]. First, the gas must be seeded with N : at concentrations of at least 2 5 % for use as a probe molecule. Nitrogen is the recommended probe gas since the molecular parameters used in calculating the temperature from the C A R S spectra are well known at high temperatures and N : is relatively stable, with only a small fraction being converted to NOx . The second problem encountered is beam stirring effects from the refractive index fluctuations. This effect leads to temperature measurements biased toward lower temperatures since the C A RS signal is proportional to the square of the local flow density. It was concluded that C A R S measurements in oxy-fuel flames were not practical in combustion chambers with an internal width greater than 30 cm. The difficulties in conducting temperature measurements on oxy-fuel flames are clearly outlined above and a search for an alternative technique applicable to industrial scale flames is needed. One alternative optical technique that has been implemented on the C R C pilot furnace is modified line reversal. Though this technique is line-of-sight. it has been demonstrated to be robust, easy to implement, and can provide essentially real-time temperature information. In traditional line reversal, an alkali species such as N a or K is seeded into the flame, resulting in emission due to electronic excitation. A background light source giving a continuum is directed through the flame and is adjusted until the spectrum line from alkali species appears neither in absorption nor emission. It can be shown from Kirchoffs law that the effective excitation temperature of the spectrum line is the same as the background light source temperature. which can be measured with a pyrometer. This technique works adequately on laminar flame conditions, but not in the case of turbulent flow where temperature fluctuations are present, since it is not possible to adjust the lamp on the same time scale as the temperature fluctuations. To overcome the difficulties with traditional line reversal on turbulent flames Air Liquide has implemented a modified line reversal technique [23]. This method allows nearly simultaneous detection of both the alkali emission and the transmission through the flame using a synchronized chopper system, as shown in Figure 18. The configuration gives a temporal resolution of 0.5 to 2 m s making the technique attractive for turbulent flow measurements. In this case, the reference lamp intensity is held constant and the flame temperature is obtained by recording the reference lamp intensity, the lamp transmission across the flame, and the flame emission. With this information, the radiative transfer equations can be solved for the ratio of the Planck blackbody functions at the flame and lamp temperatures. Thus, the flame temperature is found relative to that of the known fixed lamp temperature. Figure 19 shows an example of a realtime spectra captured from an A L G L A S S ™ burner seeded with N a firing at 1.2 MMBtu/hr. along with the temperature obtained by solving the radiative transfer equations. Near line center, the average temperature is obtained and from the wings of the spectra a temperature approaching the maximum along the path is obtained. The M L R measurements give only line-of-sight information since the detector views the transmitted lamp signal and flame emission signal across the entire optical path length. The spectral line observed is a path integral of the temperature and atomic number density. However, local temperature information can be obtained by fitting the spectral line shapes to a model using an accurate model of the atomic absorption coefficient and computation of the radiative transfer integral. In conducting the spectral fit for the observed radiation along the path information on the radial temperature profile is needed for an initial guess. Here C F D modeling is useful to provide the fitting program with an initial temperature profile. Another method for obtaining local temperature information is using inversion techniques (tomography) from a number of equally spaced line-of-sight measurements. Experiments using both line shape fitting and tomography techniques are planned. 8 |