A forecast model of the sinter tumble strength in iron ore fines sintering process

Abstract

Tumble Strength (TS) of iron ore sinter, affected by numerous factors, is considered as a vital performance to assess sinter quality for blast furnace (BF) iron-making. For the sake of providing a credible manipulative strategy of TS in sinter production, we built a mathematical model using artificial neural network, the so-called ANN technology, to predict the sinter TS. In building process of this model, the principal component analysis method (PCA) and the genetic algorithm method (GA) were introduced to optimize the ANN model so that we could obtain an accuracy model. Moreover, the sensitivity of possible influence factors was analyzed to pick up their detailed influence on TS. The calculating data demonstrate that the forecast precision of the model here is 95.1%, signifying that this model is available to describe the sinter TS basing on the initial values of numerous influence factors.