Flame Tomography

Direct 3D Tomography of Flames using Maximization of Entropy Technique (2011 - 2012)

Adviser: Dr. P M V Subbarao, Department of Mechanical Engineering, IIT Delhi

Tomographic methods are ubiquitous in medical diagnosis and non-destructive evaluation techniques in a variety of engineering applications. Flame tomography is particularly useful because it is often difficult to acquire measurements inside a flame using physical probes. Using projection data and implementing reconstruction algorithms can help in revealing and understanding important features of flame intensity field and combustion process. We developed a code for direct 3D reconstruction using the maximization of entropy (MENT) technique, applied to emission tomography. This method improves reconstruction capabilities using sparse data collected from any point in 3-D space. We carried out reconstructions using a model phantom object, and actual flame images obtained from a gas burner and a candle. The results show significant decrease in reconstruction error as compared to previously established methods using 2-D slice-by-slice reconstruction. The algorithm was tested for robustness by introducing noise in the data and was found to work satisfactorily up to a noise level of 10%. The reconstructed intensity field of actual flames show striking similarity in flame geometry and normalized intensity distribution. Several-fold enhancement in computation efficiency was also observed by using an efficient optimization algorithm.

Pictures below show the original flame on the left, the reconstructed flame in the center, and and some planar intensity surfaces on the right.

Publication: Goyal, A., Chaudhry, S., Subbarao, P.M.V., 2014. Direct three dimensional tomography of flames using maximization of entropy technique, Combustion and Flame, 161(1). pp. 173-183 DOI: http://dx.doi.org/10.1016/j.combustflame.2013.07.024 (PDF)