Study On The Kinetic Model Of Dyeing With Supercritical CO₂

With the problem of environmental pollution becoming increasingly severe in China,textile dyeing and finishing industry,as a typical industry of high energy consumption,pollution and emission,its development will have to comply with the trend of energy saving,low carbon economy and environmental protection.Supercritical carbon dioxide?SC-CO₂?dyeing technology has received extensive attention due to its many advantages such as no pollution and zero discharge.In recent years,numerous experimental studies conducted by experts and scholars on supercritical fluid dyeing technology have provided the basis for the industrialization of supercritical dyeing technology.However,the supercritical dyeing kinetics model,as the basis for the realization of the design and production of supercritical dyeing industry,has not yet been reported.Based on the method of data-driven and the analysis of the main factors affecting the K/S value of fabric and dye uptake,this paper summarizes and reports a research on the kinetic model of supercritical carbon dioxide dyeing after the data induction and classification of 12 groups of published supercritical carbon dioxide dyeing.Based on MATLAB software,a supercritical CO₂ dyeing kinetics model based on generalized regression neural network?GRNN?was established.After seeking the best spread value by the 7-fold cross-validation method,the maximum relative error of the model is between 2.55%and 11.67%,among the groups of data one is 11.67%and the others are less than 10%.The results above indicate the feasibility of the modeling method and that the GRNN model can accurately reflect the variation of fabric K/S value and dye uptake during supercritical dyeing,which can correctly predict the K/S value of fabric and dye uptake.Based on MATLAB software,a supercritical CO₂ dyeing kinetics model based on back propagation neural network?BPNN?was established.After reasonable selection of network parameters and training,the network structure of n-7-1 was determined.In the training process,the BP model has a high degree of fit to each group of data,and the average regression coefficient is 0.99932.It can be seen from the results that the maximum relative error of the BP model is between 1.44%and 8.94%,and the maximum relative error is less than 10%,which indicates that the BP model can well reflect the variation of the K/S value of fabric and dye uptake in the supercritical dyeing process,with a predictive ability.In this paper,the supercritical CO₂ dyeing kinetics model based on GRNN and the supercritical CO₂ dyeing kinetics model based on BP neural network were established respectively.Both models can well predict K/S value and dye uptake under different conditions.It can be seen that the prediction accuracy of BP model is higher,but the prediction result reliability of GRNN model is higher.The two models provide a theoretical basis for the design and production of supercritical dyeing industry,and are of great significance for advancing the industrialization process of supercritical dyeing technology.