Hybrid intelligent forecasting method of the laminar cooling process for hot strip

To overcome the difficulties of frequently varying operating conditions of laminar cooling processes and of measuring the strip temperature in the cooling process online, a hybrid intelligent forecasting approach of the strip temperature was developed, which combines mathematic and hybrid intelligent methods. The proposed approach is based on the hybrid multi-intelligence technology, where the RBF neural networks, CBR and fuzzy logic reasoning have been used to obtain the parameter estimates, with which a desired prediction on the coiling temperatures has been obtained together with the cooling temperature curve in the cooling process. A number of tests using industrial data have been conducted where desired numerical results have been obtained. It has been shown that the proposed algorithm has a high potential of being used to realize an effective control of the whole process. © 2007 IEEE.