No abstract

The optimization model is proposed on the basis of the Litster's granulation model for the prediction of the granulation effectiveness of the sinter feed. Formulas are proposed to calculate parameters for the adhesive layer of the granules, including its mass‐averaged particle size and the water saturation. Based on small‐scale granulation and sieving experiments, a functional correlation is established that relates the granules properties to the distribution coefficient curve, confirming the validity of the model. The prediction of model is discussed with adhesive layer boundary conditions of −0.25, −0.5, and −1.0 mm, and when the boundary is −0.5 mm, the relative error is minimal. The optimization of the distribution coefficient curve based on the binomial interpolation method can make the model achieve a hit rate of 92.9%. Based on this model, the granulation effectiveness of the mixtures could be evaluated.

An optimization model for the fuel distribution of the material layer with mechanism models and algorithms is presented, which can reduce the fuel ratio of ore blending and enhance the fuel content in the upper layer to improve the homogeneity of the sinter quality. The actual fuel distribution of the material layer is analyzed through the granulation model and the theoretical fuel distribution for each unit is obtained using the numerical model. Then, the fuel particle size composition and segregation characteristics are optimized via the particle swarm optimization algorithm to bring fuel distribution close to the theoretical value, with a sum of their absolute values of 0.025. In comparison with the initial conditions, the fuel particle size composition conforms to a normal distribution with increased coarse particles, and the deviation of the granules from the vertical direction of the sinter layer decreases. Through this optimization method, the sinter bed with the ideal thermal profile can be achieved, namely uniform sintering along with a reduced coke ratio.

No abstract