<p>Tunable diode laser absorption spectroscopy (TDLAS) has been an effective technology to measure the velocity of the combustion flow in hypersonic engine. The measurement theory is based on the relative Doppler shifts of the absorption transition for the upstream and downstream beams in the flow. The resolution of the TDLAS system for velocity measurement, which is defined as the minimum velocity change value that the TDLAS system can identify, is analyzed in three aspects: the structure of the beam path, the absorption transition from the laser, and the quantization of the digital signal sampling. The measured accuracy of velocity is also investigated based on the study of the resolution. An experiment was executed on a scramjet engine with a TDLAS system to validate the resolution and the accuracy of the measured velocity.</p>
Tomographic absorption spectroscopy (TAS) is a promising combustion diagnosis technique. It can simultaneously provide 2D temperature and concentration of the flow field. Some engine combustors, such as scramjet combustors and internal-combustion engines, only allow limited optical access. Only a few beams (ca. 20-40) are available to measure the combustor and beam layout becomes a key factor affecting TAS accuracy. In engineering application, probes are embedded in the wall of the combustor for laser propagation. The probe size should be taken into account when designing beam layout. In this paper, we considered the probe locations into beam optimization. The math formulation of the problem and the solving algorithm (genetic simulated annealing algorithm) is developed.
Two-dimensional imaging of gas temperature and concentration is realized by hyperspectral tomography, which has the characteristics of using multi-wavelengths absorption spectral information, so that the imaging could be accomplished in a small number of projections and viewing angles. A temperature and concentration model is established to simulate the combustion conditions and a total number of 10 near-infrared absorption spectral information of H<sub>2</sub>O is used. An improved simulated annealing algorithm by adjusting search step is performed the main search algorithm for the tomography. By adding random errors into the absorption area information, the stability of the algorithm is tested, and the results are compared with the reconstructions provided by algebraic reconstruction technique which takes advantage of 2 spectral information contents in imaging. The results show that the two methods perform equivalent in low-level noise environment, but at high-level, hyperspectral tomography turns out to be more stable.