Research On Optimization Test Of Thermal Coal Blending Heat Prediction And LIBS Detection

In China,with the continuous development of coking,steel and electric power industries,the modernization has put forward higher requirements for each industry.For the coal industry,the proposal and implementation of"green energy"put forward higher requirements for the quality of coal and the optimal utilization of coal resources.With the development of artificial intelligence technology,it is of great significance for the coal industry to realize the stable and accurate online detection of coal quality indexes.In view of the limitation of the traditional laboratory testing time,and the market based on microwave technology,neutron activation technology,X-ray fluorescence spectroscopy and other principles of the coal analyzer,due to their own shortcomings and harsh conditions and other problems can not be promoted and applied in a large scale.This thesis studies the application of laser induced breakdown spectroscopy in coal quality detection direction effect,through the optimization of LIBS experiment for power coal blending products of ash content and calorific value forecast analysis,prediction results show that the dynamic prediction of ash content relative error is0.61%,the calorific value relative error 4.12%,and the technology in coal quality detection with high accuracy.In this thesis,combining with the actual situation of the production site,through the analysis of the coal quality of the coal seam,the optimization scheme of coal blending is determined,and the coal blending experiment is carried out in the coal preparation plant.Through the data of coal blending,the coal quality index relationship between single coal and coal blending products is analyzed.Through Pearson correlation analysis,it is concluded that sulfur content,ash content and calorific value have a good linear addable relationship,while internal water has no significant linear relationship.According to the relationship between the indexes,the quality prediction model is established,which is used in the data analysis and application of thermal coal blending.According to the traditional test process,the calorific value of the coal sample needs to be converted into the calorific value of the cartridge,but the test process is tedious,time-consuming and laborious,so this article tries to optimize the test process.Establish calorific value prediction model by using total water,internal water,ash and sulfur as input parameters,apply genetic algorithm improved BP neural network algorithm and LS-SVM support vector machine algorithm,predict calorific value by establishing different models,and compare calorific value Forecast result error.The results show that the calorific value prediction model established by the LS-SVM support vector machine algorithm has the maximum absolute error of 176.51 J/g,the correlation coefficient R~2 value is 0.98,the calculation speed is fast,and the prediction result is good...