An Automatic Control System Of Prediction Compensation For E.A.F Pole

This paper comes from a project which is to design and produce a new automatic control system to replace an I.A.F system runned by Guangzhou Iron Factory.The older I.A.F system was produced by U.S.A. Its maintain cost is too high and its running effect is not satisfied enough.The new systerm is an artificial one.Tested by many experiments,it is justified that it can improve many important characteristics and strengthen manegement.The paper presents a new E.A.F(electric arc furnace)adaptive control method based on a neural network's compensation by predicting.The neural network is used here to predict the condition of E.A.F,then to control an actual system by compensating the outputing of routine controller.The dynamic modification of the weighting value of a neural network by improved DBD algorithm serves as an adaptive method.Hence the system can adapt itself better to the variations in load and external interference.which has yielded a satisfied control quality.