Estimation of excess air coefficient on coal combustion processes via gauss model and artificial neural network

It is no doubt that the most important contributing cause of global efficiency coal fired thermal systems combustion efficiency. In this study, relationship between flame image obtained by a CCD camera and excess air coefficient ({\lambda}) has been modelled. The model with three-stage approach: 1) Data collection synchronization: Obtaining images means mounted on 10 cm diameter observation port, {\lambda} data coordinately measured recorded flue gas analyzer. 2) Feature extraction: Gridding image, it divided into small pieces. uniformity each piece to optimal calculated modelling single multivariable Gaussian, calculating color probabilities Gauss mixture approach. 3) Matching testing: A multilayer artificial neural network (ANN) used for matching feature-{\lambda}.