Diagnostics of combustion process based on flame images analysis and genetic programming

One of the means of assessing the state of combustion process is analyzing data obtained from flame images. Flame images can be used as a source of information of combustion process input parameters such as air flow, fuel expense or fumes temperature. These parameters are crucial for stable and effective combustion. Considering complexity of combustion process and large number of fluctuating variables, genetic programming is an approach that seems most appropriate. Several video streams of co-combusting biomass and coal were recorded at the rate of 150 frames per second with 800x800 resolution. Different image and combustion parameters were processed to determine dependencies between them.