Study On Gas Prediction Methods In Branch Plants Of Main Process In Iron-Steel Enterprise

In iron-steel enterprise, it is important that gas generation and gasconsumption is predicted scientifically in order to improve the level ofenergy saving, change the present state of balancing gas generation andgas consumption based on the experience, make the reasonable use planof gas and reduce the loss of gas emissing.Main factors effecting on gas generation and gas consumption ofbranch plants have been found through the detail researches on thegas-generation mechanism and the consumption characteristics in branchplants of main process in iron-steel enterprise. The indexes influenced bythese main factors have been found in energy-balance table of iron-steelenterprise. Relative-degrees between gas generation and indexes andbetween gas consumption and indexes in energy-balance table have beencalculated by the gray relation method. Main indexes have been foundafter ordering the relative-degrees. Considering the different number ofmain indexes effecting gas generation and gas consumption and theirdifferent influencing degree in branch plants, prediction methods of gasgeneration and gas consumption in branch plants of main process iniron-steel enterprise have been put forward. The linear regression analysismethod can be adopted in the branch plants that have the simple relationand the small number of influencing indexes and otherwise the BP Neuralnetwork method can be adopted.Taking the gas system of Xiangtan Iron & Steel Group Co. Ltd. as anexample, accuracy and practicability of the gas prediction method havebeen studied. The linear regression method has been adopted and theregression equations have been setup to predict gas generation of the cokeoven and gas consumption in the coke-oven plant, iron plant, bar mill,wire rod mill and the second wire rod mill. The BP neural networkmethod has been adopted and the neural network models have been setupto predict gas generation of the blast furnace and the rotary furnace andgas consumption in the sinter mill and the steel mill. Researches showthat the errors between the prediction values and the actual ones are small and the gas prediction method is accurate and can meet the demand ofproduction.