Hierachical intelligent optimization blending system based on production indices for lead-zinc sintering process

In lead-zinc sintering process, various kinds of lead-zinc concentrates and returning powder from ineligible sintering production are blended and then sintered in updraft sintering machine to produce agglomerate for imperial smelting process. It is required to determine a proper mixture ratio for these materials blended to assure sintering production indices such as agglomerate composition and sintering permeability. Considering the relations between the mixture ratio and the production indices, an intelligent integrated model is firstly constructed to predict agglomerate compositions, which consists of Expertiseand-Mechanism-based model and Supervised Distributed Neural Networks. Based on the composition prediction model and reasoning rules for mixture ratio modification, a hierachical intelligent optimization strategy with expert reasoning is proposed to determine an optimal mixture ratio, which includes multiobjective optimization for the first blending process and area optimization for the second blending process. Practical running results show that the qualified rates of agglomerate compositions are increased by about 7% and the fluctuation of sintering permeability is reduced by 7.0 % through proper blending process according to the optimal mixture ratio, which effectively stabilizes the imperial smelting process.