Study On Calcination In Precalcing Kiln Based On Artificial Neural Network

The calcination of Rotary Kiln is a key step of cement production ,it makes directly influence on the quality of cement clinker .But the process is so complex that there is no traditional method which can describe it precisely .In the paper, two new methods -Numerical Heat Transfer and Artificial Neural Network are introduced .Through the combination of experiment and the methods, some important principles is found .Finally , a new method is used to describe the process of calcinations precisely.Through great deal of experiment and theoretic analysis of its results, some principles about thermal field is obtained. And this proves Finite Element Method can be used to describe temperature field. But a whole Rotary Kiln is a very complex system and the factors which influence the process of calcinations is nonlinear. As a result, it is almost impossible to use traditional method to describe the process. But this is just strongpoint of Artificial Neural Network, so it is used to simulate the calcination Rotary Kiln.The paper have a analysis on all influence factors about cement quality, omit some influence factors which have little effect on it, and set up the Artificial Neural Network model about cement clinker with all of influence factors. There is a analysis in the paper, have the stabilization of f-CaO of cement clinker as the goal, through the analysis about ten factors. Such as the quantity of raw materials in calciner, the decomposition rate of raw materials, the quantity of coal in calcinet, the quality of coal, ventilating volume of calciner, second air temperature, rotate speed of Rotary Kiln, KH of raw materials, SM of raw materials, IM of raw materials etc. At last, it have six factors of them as influence factor about f-CaO and set up the Artificial Neural Network model about cement clinker with all of these six influence factors. These six factors are the quantity of raw materials in calciner, the quantity of coal in calciner, rotate speed of Rotary Kiln, KH of raw materials, SM of raw materials, IM of raw materials. The model have a nicety forecast on cementproduction in factory. In order to compare and analyze, there is a compare about the result of regress base on the theory of least square method and Artificial Neural Network. The study show that: for the model on nonlinear, Artificial Neural Network have strongpoint at that.Matlab is a very useful math calculation tool software. It give us advantage to set up model by using artificial neural network toolbox of Matlab. This model is base on the data of production, it have strongly guidance on the practice.