On-line identification of biomass fuels based on flame radical and application of support vector machine techniques

In biomass fired power plants, a range of biomass fuels are used to generate electric power. It is vitally important to identify the type of biomass on-line in order to improve combustion efficiency, reduce emissions and ensure the boiler safety. Present research focuses on the on-line identification of biomass fuels using flame radical imaging and SVM (Support Vector Machine) techniques. The characteristic values of flame radicals, including OH*, CN*, CH* and C2*, are extracted and used to reconstruct the SVM for on-line fuel identification. Experimental results obtained on a laboratory-scale biomass-gas-fired combustion test rig demonstrate the effectiveness of the proposed method.