Bed expansion characterization of fluidized beds with Geldart A dense medium based on machine learning
In order to predict the expansion characteristics of Geldart A particles,we investigated the correlation between the composition of the bubble and emulsion phases in the fluidization process and the operational factors,constructed data sets of operating gas velocity,static bed height and bed expansion height by bed collapse experiments,performed statistical distribution and correlation analysis,and then successfully simulated the nonlinear relationship between expansion height and influencing variables by using GBDT model with optimal hyperparameters;and finally,performed sensitivity analysis of the characteristic variables.The results show that,with the gradual increase of gas velocity,the expansion of emulsified phase and bubble phase shows a pattern of increasing and slightly decreasing first.For different initial bed heights and bed expansion heights,the composition of the bubble phase is not affected by their changes and the composition ratio is relatively stable.The importance score of operating air velocity is 0.68,which is the most sensitive variable for the expansion height.Compared to the operating gas velocity,the importance score of the static bed height is 0.32,and the influence of the static bed height on the bed expansion is smaller.In addition,from the partial correlation analysis,it can be seen that the dependence of bed expansion height on operating gas velocity has a certain sensitivity interval.
NETL
NSTL
万方数据
