Study On The Ash Fusion Property Of Biomass And Coal Blending

In recent years, due to the problem of energy crisis and environment pollution become more and more serious, people began to research the biomass energy to use cleanly and efficiently. Because biomass has high alkaline substance content, low ash fusion temperature and small difference between softening temperature and flowing temperature, the boiler have the problems of fouling, slagging and corrosion when burning biomass and blending with coal. So the studying on the ash fusion property of biomas and blending with coal can speed up resolving the problems of fouling, slagging and corrosion of biomass boiler, and prompt our country to have the ability of produce, reaserch and design of biomass boiler.First of all, a thermal analyzer modeled TGA/SDTA851e was used to study the ash fusion property of eight kinds of biomass, such as wheat straw, rice straw, corn stalk, cotton stalk, peanut shell, poplar chip, vinasse and paper sullage. Because of the differences in ash chemical composition and mineral composition of the biomass, there were obvious differences in ash fusion property through the thermal analysis methods. Thermal analysis indicates that the principal causes of weight loss of the biomass ashes are the decomposition of carbonates and sulfates and the evaporation of alkali oxides and alkali metal salts. The endothermic peaks on the DSC curves are mainly caused by the decomposition of carbonates and the melting of alkali silicates and alkali aluminosilicates. According to the migration regularity of chemical element during ash fusing, the function of weight loss rate, which deduced by regression analysis method, can reflect the weightless regulation of most element contents of biomass ashes.Secondly, the effects of biomass proportions on the ash fusion temperature of biomass and coal blending were researched through YX-HRD ash fusion analyzer. Because coal has higher ash content than biomass, ash fusion property of biomass and coal blending reflects the ash composition and fusion prpperty of the coal. With the increment of the biomass proportions, the relationship between ash fusion temperature and the biomass proportions is nonlinear in general, but the relationship is approximate to linear property in part proportions ranges. Through experimental research on ash fusion property of single biomass and coal blending, it is found that the sequencing of ability of biomass to reduce coal ash fusion temperature is ( rice straw, wheat straw ) >( corn stalk, vinasse, cotton stalk )>( poplar chip, peanut shell ) . Through experimental research on ash fusion property of double biomass and coal blending, it is found that the effect of double biomass on coal ash fusion property basically embodies the combined action of double biomass. Through the influence test of additives on biomass ash fusion temperature, it is verified that ash fusion temperature increases with increasing content of Al₂O₃, while the effect of SiO₂ on ash fusion temperature is not obvious, but the ash will be sintered with a higher content of SiO₂. CaO and MgO play a role as skeleton at low content, and ash fusion temperature is significantly increased with a higher content.According to the ash fusion temperature of biomass and coal blending in different proportions, multiple linear regression model and quadratic polynomial regression model between ash chemical composition and ash fusion temperature were estimated. Regressive analysis shows that the significance of quadratic polynomial regression model, which has five independent variables merged from eleven independent variables, is obviously better than multiple linear regression model.At last, the radial basis function neural network (RBFNN) and the generalized regressive neural network (GRNN) were applied to build models for forecasting the ash fusion temperature of biomass and coal blending. 190 group samples were used to train and 8 group samples were used to check up the validity of the network. The simulated result shows that GRNN has better predictive ability and generalization ability than RBFNN, and the prediction results of GRNN satisfy the error requirement. So GRNN is suitable to be used to build models for forecasting the ash fusion temperature of biomass and coal blending.