Integrated neural-network-based method for predicting synthetic permeability in lead-zinc sintering process

In order to deal with the time variance and strong nonlinearity of lead-zinc sintering process, an integrated method of predicting synthetic permeability based on neural networks (NNs) and a particle swarm algorithm has been developed. In this paper, the concept of the exponent of synthetic permeability, which reflects the state of the permeability of the process, is first explained. Next, NNs are used to establish time-sequence-based and technological-parameter-based models for predicting the permeability. Then, an integrated structure based on a fuzzy classifier is described and used to construct an intelligent integrated model for predicting the permeability that combines the two models just mentioned. Finally, the results of actual runs show the method to be both effective and practical.