Research On Oxygen-to-Coal Ratio Optimization Of Coal Water Slurry Gasifier Based On Hybrid Model

With the development of industrial intelligence,the method of manually controlling the ratio of oxygen to coal in coal water slurry gasification can no longer meet the needs of industrial production.As one of the important process parameters in coal gasification process,the ratio of oxygen to coal is susceptible to fluctuations due to many internal and external factors such as coal quality and process conditions.The single mechanism and empirical model cannot accurately predict the gasification process.Therefore,this article takes coal water slurry gasifier as the object,and optimizes the oxygen to coal ratio by establishing the mixing model of the gasifier.Based on the partition of gasifier and neural network,the partition model and GRNN model of gasifier were established respectively in combination with the operation data of the factory.It is verified that the prediction errors of the two models are 1.42%and 1.7%,and the calculated results of the models are in close agreement with the industrial data,which verifies the effectiveness of the models,but it is poor in the case of coal quality fluctuations.A hybrid model based on CWS gasifier was established by combining the established regional model and GRNN model with parallel hybrid modeling method.The mechanism model is the main body of the hybrid model,and the GRNN model is the function approximator,which plays the role of residual compensation.It is proved that the prediction errors of the mixed model regarding the gasification results are about 0.9%and 0.5%,respectively,under coal quality is fixed and the coal quality changes.and the prediction accuracy is higher than that of the single mechanism model and the data-driven model.The effect is good under the condition of coal quality fluctuations.The relationship between the ratio of oxygen to coal and gasification performance was analyzed based on the mixing model.Based on the hybrid model,an automatic method for predicting the set value of oxygen to coal ratio is proposed by using R-type clustering method and nonlinear fitting method.The key parameters related to oxygen/coal ratio were obtained by R-type clustering analysis,including temperature and CH4 content.Combined with the nonlinear fitting method,the CH4 content,gasification temperature and oxygen-coal ratio are quantified,and the prediction error is about 0.15%,which can be used to predict the oxygen-coal ratio in actual production.