Conditions identification model based on LLNFM and RBR in cement raw meal calcination process

In cement raw meal calcination process, there are three conditions (i.e., easy calcination condition, difficult calcination condition, and abnormal condition), however it is difficult to be estimated in time by operators. To solve this difficult problem, a prediction model has been proposed by combing local linear neuro-fuzzy model (LLNFM) with rule-based reasoning (RBR). The LLNFM was applied to the model to predict the output temperature of the preheater C5 using input variables. Rule-based reasoning decided conditions according to predicting output valve. The proposed model has been successfully applied to calcination process of Jiuganghongda Cement Plant in China, and the application results showed its effectiveness.