23 European Symposium on Computer Aided Process Engineering

There is a growing interest in the use of gasification to contribute to the energy share, as a more efficient alternative to fossil fuels combustion, especially with coal and biomass in large power plants. In order to unveil the effect of time in the degradation of the gasifier operation and the influence that operating conditions have over reactor performance, the evolution of controlled and observed variables should be studied. Two statistical techniques are used in the present work: (i) multiple regression analysis, which provides information about the relationship between several predictor variables and a dependent variable, and (ii) time series analysis, which may allow identifying the nature of the phenomenon associated with reactor degradation as a function of time. The final goal consists of evaluating and predicting the reactor natural instability by monitoring CO2 emissions. Results demonstrate the capability of forecasting techniques to predict reactor degradation with time. Thus, the tendencies of the variables and parameters in the gasifier may be better determined to avoid faulty operation.