Study On Intelligent Fault Diagnosis System Of The Cold Rolling Mill

With the rapid development of the steel industry, the use of cold rolling mill becoming more and more widely. Cold rolling mill is not a simple rotary machine, it has many special problems, so the diagnosis is very difficult. Although has experienced decades of development, the knowledge of the fault mechanism and diagnosis method for cold rolling mill is not mature enough, It is difficult to accurately diagnose with the traditional diagnostic methods. With artificial intelligence techniques become more sophisticated, the application in the field of fault diagnosis is more and more widely. Intelligent diagnosis technology to replace the traditional diagnostic method for diagnosis of cold rolling mill has become an inevitable trend of development.In the paper, The structure and principle of1700Cold Rolling Mill in "No.1Production Line" of Ansteel Cold Rolrd Works is taken as research object, and get its fault diagnosis type and reason of diagnosis on the working stand, reducer and main motor. Using time-domain analysis and wavelet packet analysis method on the vibration signal feature extraction, Combined with the field monitoring of rolling force tension, and strip speed on its running state monitoring and diagnosis.Analysing the theory of BP Neural Network and its applications in fault diagnosis field when building the fault intelligent diagnosis system, building the sample set for neural network, test set and simulation training basing on dealing with the vibration signals of Cold Rolling Mill. In the process of system implementations, using Visual C++6.0to get human-computer interface design, and using MATLAB to achieve Neural Network training, testing and simulation, Making full use of them the interactivity to finish system design. The testing proves to be the system reaches the actual requirements, and the system gets the better accuracy.