| Hydrocyclone is a kind of widely used separation equipment. It can realize many tasks such as the classification of solid particles, the clarify of liquid, condense of slime, abstersion of solid particles and so on. The separation procedure of hydro cyclone is very complex, the relationship between affective factor and index factor is a multiple-nonlinear system, it is difficult to describe their relationships with simple linear mathematical facility.Artificial Neural Networks is a kind of structure to simulate human being's neural system; it discovers the nonlinear relationship in data. a great deal of processing units makeup a nonlinear dynamic system, obsessing many merits such as self-adapting, self-studying, fault-tolerant and nonlinear function approximating, it can erect models of complicate modulus. Therefore, this paper applies with this new technology, Ann, to realize the prediction of hydro cyclone's performance index.The paper regards F150 hydro cyclone as research object, apply with some iron-mineral separation data which are collected in a hydro cyclone company, and erect a neural net model to predict the indexes of hydro cyclone. In this model, original data need be pre-processing. Compare models which are erected in different conditions, and confirm the structure of neural net is 5-11-5, the arithmetic is traingdm. Then the paper provides the main codes compiled with the toolbox of Ann.In order to provide users a convenience platform for prediction of hydrocyclone, the paper applied the technology of matlab and Delphi, erect a predictive windows for prediction, then illustrate the procedure of transfer. In order to validate the veracity of predictive system, this paper select 5 group of industrial data, and compare the predictive indexes with industrial data, and draw a conclusion that the predictive effect is good, and, the arithmetic traingdm has a good ability to be popularized.In a word, the paper realized the prediction of F150 hydrocyclone, and the predictive system is value to iron separation in industry. |