Combustion of low-quality fuels or fuel blends will lead to flame instability

resulting in low combustion efficiency and high NOx emissions. Due to the inherent complexity of burner flames and the lack of an effective means for flame monitoring and characterization

it is difficult to evaluate the flame stability in a combustion process quantitatively. To solve this problem

a method based on digital image processing for co-firing biomass fuels is proposed in this paper to monitor various characteristic parameters of a burner flame and evaluate its stability. In this method

a general flame stability index with continuous values in the range of [0

1] is defined

and by using a digital CCD camera

the flame image information is collected. After the collected image is analyzed

the characteristic parameters like the flame length/height

brightness

temperature

flicker frequency and others are extracted. Then

statistical analysis and data fusion are carried out for theses characteristic parameters

and the flame stability index is obtained. Thus

the quantitative detection and evaluation of flame stability is realized. Moreover

this method was verified on a laboratory-scale combustion test rig. The combustion behaviours of different biomass blends(corncob-wheat straw blend

willow-peanut shell blend and peanut shell-wheat straw blend) were compared. The results show that

the defined flame stability index could effectively characterize the flame combustion state.