Soft Sensor of the Material Flow from the New Suspension Preheater Kiln

摘要

Because it is unable to do online measure on the material flow from the kiln in NSP cement production line, a soft sensor model is established by using RBF neural network. As output variables, the material flow from the kiln which obtained by the clinker instantaneous that according to the grate cooler stable state under the condition of clinker. RBF neural network soft sensor model has two dimension of input and one-dimensional of output which we get from the study on NSP cement rotary kiln normal conditions. We use the data which collect from the cement production site to test and compared the reliability and the prediction accuracy of the model. Experimental results showed that the measured result and the prediction model result are close and the precision can reach to 97.5%. From the compared experiment it can find that the RBF neural network soft sensor model is superior to the BP neural network soft sensor model in approximation ability, classification ability, learning speed and so on. Therefore, RBF networks can be used as the kiln material flow of soft measurement model.