The Research Of Power Plant Combustion Characteristics Extraction Method

The combustion characteristic extraction is the research premise of boiler combustion stability estimation with the low load, the combustion fault diagnosis and the combustion control and optimization. The furnace flame detection is one of the characteristic signals that reflect the boiler combustion state directly. The software Matlab is used as a platform for signal analysis and the Butterworth filters of low-pass, band-pass and high-pass with different frequency bands are desinged.With the filtering of the furnace flame detection signals, more information is obtained apart from reflecting the flame on and off. Through the analyses of the 24 coal flame detectors and 16 oil flame detectors output signals of 600 MW coal-fired generating unit with the filtering method, it is found that the intermediate frequency signal of single flame detector combustion intensity is correlative with that of the layer fuel amount, the intermediate frequency signal of the weighted sum single flame detector combustion intensity is very correlative with that of total fuel amount.The flame detection signal of filtering could reflect the boiler combustion state quickly,which could be used for feed-forward control and combustion fault diagnosis.