Research On The Influence Law And Prediction Model Of Hydrocarbon Oil On The Flotation Of Coal Slime With Different Continuation Degrees Degrees

Coking coal,as a high-quality mineral resource,is widely used in various industries.However,domestic coking coal resources have characteristics such as low reserves of high-quality coal and poor washability.Flotation is the main source of obtaining high-quality coking coal.Due to its good natural floatability,hydrocarbon oils(such as kerosene,diesel,etc.)are often used as collectors to improve the flotation quality of coking coal.However,in industrial applications,there are strict requirements for the ash content of coking coal.One of the main problems in coking coal flotation is that the ash content of cleaned coal is easily affected by collectors and the content of raw coal in flotation,making it difficult to adjust the ash content of flotation cleaned coal in a timely manner based on the fluctuations in the properties of the selected raw coal.Building a flotation prediction model has been proven to effectively improve the flotation efficiency of coking coal.By calculating the dosage of reagents,the optimal ash content index for economic benefits can be obtained,thereby improving the economic benefits of the beneficiation plant.However,traditional models typically only target a specific collector or raw ore,and their applicability is relatively narrow.Therefore,research methods such as small floatation and settling tests,contact angle testing and analysis,single mineral flotation tests,and neural network models were used to reveal the flotation laws of various hydrocarbon oils and coking coal with different coalescence degrees.A neural network model for the flotation process of various hydrocarbon oils and raw coal was established.Divide the coal sample into components with different content of connected bodies through small float and sink tests,and calculate the content of connected bodies for each component.The results show that the higher the density of the coal sample,the higher the content of connected bodies.After analyzing different types of coal samples with connected bodies,the influence of connected body content on the floatability of coking coal was analyzed through reagent free flotation tests,contact angle measurements,and other methods.It was found that the content of connected body has a significant impact on the floatability of coking coal.The higher the content of connected body,the poorer the floatability of coal samples,and the higher the ash content of flotation clean coal.The contact angle results show that the contact angle of coal slurry with high connectivity content is small,indicating a decrease in its floatability.Flotation experiments were conducted on different connected coal samples and alkanes,and it was found that as the carbon chain length of hydrocarbon oil increased,the ash content of flotation coal decreased,and the recovery of combustible materials increased first and then decreased,reaching its maximum at 14 alkanes.By analyzing the flotation test results,it was finally determined that a solvent with 14 alkane as the composite reagent was used to make a composite collector with other reagents and four different ash contents of raw coal for flotation tests.The results of the composite reagent flotation test showed that the ash content covered a wide range and could be used as subsequent network fitting data.Coking coal flotation is a complex physical and chemical process,which has many influencing factors.This paper combs the literature from several aspects,such as the influence of conjoined bodies on the floatability of coking coal,the research status of coking coal flotation reagents,and the research on mathematical models of beneficiation,and summarizes the technology and theoretical methods provided by the current research,which provides strong support for the exploration of flotation influence laws and modeling in this paper.The flotation principle and the process parameters affecting its performance are analyzed,and the correlation between the flotation parameters and the ash content of flotation clean coal is sorted out.The flotation data of coking coal was imported into the BP neural network to construct a coking coal flotation model.The modeling ideas and model building of the BP neural network were elaborated.The main process of model building was divided into data acquisition,parameter setting,and test result analysis.The model was ultimately determined as a network structure with 4 input nodes,6 hidden layer nodes,and 1 output node.Import flotation data into a neural network model with set parameters for operation.After the prediction model meets the expected error,select J1 coal sample flotation results for prediction testing.The actual coal sample inspection results show that the model accuracy is 88.1%,with a maximum error of 0.53.Among them,there are 25 groups below 0.1 error,exceeding half of the total experimental groups,and the root mean square error is 0.03,The accuracy of the model generally requires optimization.Import the flotation data of coking coal into BP neural network,build the flotation model of coking coal,and optimize the model with particle swarm optimization algorithm to make the prediction accuracy of the model reach the set accuracy,and test and verify the model.It is found that the prediction effect of the model is good,and the flotation hydrocarbon oil collector of coking coal can be selected through the model,so as to achieve the purpose of predicting and controlling the ash content by controlling the collector.Using particle swarm optimization to optimize the BP neural network,a PSO-BP neural network model was constructed.By referring to existing literature and empirical formulas,parameters such as particle dimension,population size,learning factor,initial speed,iteration number,and weight coefficient of the PSO algorithm were set.After setting the parameters,the model began to be trained.According to the convergence curve of the particle swarm algorithm,under the set iteration number and population number,The convergence of fitness is relatively smooth,and the curve changes basically smoothly after reaching 50 evolution times,indicating that the set parameters are reasonable.After the training value reaches the set accuracy,the J1 slime is predicted,and the predicted value is compared with the J1 flotation data.It is found that the average error of the optimized PSO-BP model is small,the model accuracy reaches 94.2%,and the root mean square error of the model is 0.01,The model can be well used to predict and guide actual flotation experiments.This article by using neural network modeling to link the content of coking coal aggregates with different collectors,a multi collector model for predicting the ash content of coking coal flotation clean coal is constructed.By selecting appropriate hydrocarbon oil as the collector based on the content of coking coal aggregates,the aim of controlling ash content is achieved,providing a new method for regulating the ash content of coking coal flotation clean coal.