Research On The Process Parameters Online Predication Of Xingtai Mine Coal Preparation Plant

Gravity separation technology is widely used in coal preparation plant.With thedevelopment of the coal industry, the automation level of gravity separation processis payed more and more attention. However, there are a series of problems in theimplementation of the automation process of gravity separation process. First of all,the means and the accuracy of some variables testedon line lag seriouslybehind thecontrol strategy. Secondly, there are not enough effective online prediction models toguide the automatic control of gravity separation process. Therefore, this papercarries out research on online prediction of the gravity separation parameters in termsof heavy medium separation and coarse coal slime separation process.Based on the actual ash content in raw coal andash of on-line ash admeasuringapparatuscombined with the measurable factorswhich influence the error betweenthem, real-time raw coal ash content model established by using statistical analysismethods is regarded as the basis of online prediction of dense medium separationprocess.By correlation analysis between the ash and the cumulative production ofeach density grade in the raw coal, the raw coal ash content and density gradecumulative rate modelwith good correlation aredeveloped.The production rate ofdensity grade with bad correlation is solved by using the empirical model of densitycurve in the washability curve so as to establish the real-time cumulative model ofeach densitygrade. Stepwise regression analysis is carried out to cumulative ash offloat in each density grade, density grade and raw coal ash content to develop themodel of the real-time accumulated ash model in each density grade together withthecumulative production rate model of each density grade to form the real-time modelof float and sink data of raw coal.Combination of raw coal real time ash model andraw coal real-time float and sink modelproduceprediction method of productstructure, separation density, and circulating medium density and finally thecirculation medium density can be obtained by ash content measured by ashadmeasuring apparatus in order to predict the product structure andpracticalseparation density. The predicted circulation medium can be used to control thedense medium system.Online prediction of coarse coal slime separation process is mainly concernedthe prediction of coarse clean coal ash. CSS coarse clean coal ash is influenced by the properties of feed material and operation parameters of the CSS separator.Therefore, the prediction model of CSS coarse clean coal ash is also affected byvarious factorsand present a strong coupling. It is difficult to build accurateprediction model by mechanism analysis. The principal component analysis, supportvector regression machine and genetic algorithm are applied to deal with theexperimental data from the field in this paper. ACSS coarse clean coal ash predictionmodel based on the GA–SVMR is developed torealize the real-time prediction ofcoarse clean coal ash through operation parameters of the CSS separator.