Based On Iron Ore Sintering Basic Characteristics Of Sinter Quality Forecast Model

The sintering basic characteristics of iron ore is one of the important indicators ofquality iron powder, which is an important factor influencing the quality of sinter.Aiming at the characteristics of the iron and steel enterprises of iron ore supply changesfrequently in our country, a forecasting model of sinter quality was established whichbased on iron ore sintering basic characteristics is to solve the problem that how to rapiddetermination of sintering ore blending and sinter quality prediction under conditions offrequent changes in raw material.Firstly, this paper analyzes the utilization status of China’s iron ore, research statusof iron ore sintering basic characteristics, sintering ore blending study and the neuralnetwork is used in the field of iron and steel, and proposes the main research method isartificial neural network.Secondly, the history and the basic working principle of artificial neural networkswere introduced, by expounding the basic knowledge of neural network and the basicnetwork model, analyses and discuss the characteristics and applications of artificialneural networks.And then discussed how to build a predictive model which based on BP neuralnetwork in detail, explore the main technical that to establish a BP neural networkprediction model, mainly including the selection of training samples and pretreatmentinputting and output neurons, choosing the number of hidden interneuron’s, determined,the initial weight, threshold function, training function, and activation function, at last tobuild a reasonable network model.Finally to establish a predictive model of iron ore sintering basic characteristics anda prediction model of sinter quality. The prediction results show that the hit rate ofassimilation reached90%, the hit rate of liquid phase fluidity and bonding strengthrespectively achieves75%and70%; For the sinter quality of forecasts, conversion andyield strength, respectively reached80%, RDI+3.15reached85%, The prediction resultsto reaches the desired objectives. This method also proved the relationship between thesintering basic characteristics of mixed ore and sinter quality that in a certain error range,the predictive model can be used for the prediction of sinter quality.