Study On Prediction Of Coal Blending Slurry Concentration Based On Integrated Learning Model

In this paper,Pearson correlation coefficient is used to analyze the correlation between coal quality parameters and slurry concentration.Parameters with strong correlation and parameters commonly used in literature are selected as variables to predict the slurry concentration of coal.An integrated learning model is established by using python.Choose the suitable model by comparing the index arguments of the model and testing.Using the application design tool in Matlab and the machine learning module in python,and taking the coal quality database established by SQL Server2014 as the background,develop a set of prediction system for coal blending slurry concentration based on Ada Boost.The research results are as follows:(1)Through Pearson correlation coefficient analysis of coal particle size distribution,industrial analysis,element analysis,silicon aluminum and the correlation with slurry concentration,it is found that the 160-220 distribution interval,Mad,O,oxygen carbon ratio in the particle size distribution are strongly related to the slurry concentration,the10-20,20-30,30-50,110-160 distribution interval,C,N,S,silicon aluminum and the slurry concentration are moderately related,and the 0.04-10,50-80,80-110 >220distribution range,Aad,Ad,Vad,Vdaf,H are weakly correlated with the slurry concentration.(2)The combined learning algorithm is used to create a design model that does not reflect granularity factors and granularity factors.It is found that the R2 of random forest algorithm is 0.667 and that of Ada Boost algorithm is 0.729 in the model without considering the granularity factor.In the model considering the granularity factor,R2 of random forest algorithm is 0.667,and R2 of Ada Boost algorithm is 0.915.Through comparison,it is found that the accuracy of Ada Boost algorithm is always higher than that of random forest;The accuracy of the algorithm considering granularity factor is higher than that of the algorithm without granularity factor;Comparing the four algorithms,we can find that Ada Boost algorithm considering granularity is the optimal algorithm.(3)Use Matlab to call the formed Py function compiles the code of the prediction module of the slurry concentration of single coal and coal blending,and establishes the prediction system software.Through the test of the normal operation of the system,it can quickly and accurately predict the slurry concentration of coal blending.(4)The coal quality database is established,and the application design tool provided by Matlab is used to program and manage the coal quality database,realizing the "add","modify","delete" and "query" operations of coal quality data.Figure [38] Table [19] Reference [93]...