Research And Development On Fault Diagonisis System Of Distributor Gearbox On Blast Furnace

Distributor gearbox of blast furnace is an important device which operation security not only relates to the safety and production of blast furnace, but also indirectly affects the subsequent process like steel making, casting and hot rolling. As the impacts of heat, dust and other adverse environmental conditions, failures happened frequently on the gearbox, thus have a serious interference with blast furnace production and cause huge losses. But the present fault diagnosis method which relys on motor current monitoring and regular maintenance, It is very difficult to detect faults in early stage. In response to this demand, therefore we need to develop a intelligent gearbox fault diagnosis system of blast furnace based on vibration signal detection.In addition to the completion of basic monitoring and failure analysis capabilities, some studies on extraction of the early fault feature and reduction of interference with the diagnosis accuracy from on-site noise signal are also carried out in this thesis. The main research contents are as follow:The subject background and the significance and development of gearbox fault diagnosis technology are briefly described in the first chapter , as well as the research significance and content.Mechanical structure and transmission system of the distributor gearbox on blast furnace and difficulties of the project are detailedly analysed in chapter II, as well as the common fault forms and fault characteristics of gear and bearing.Condition monitoring system of the distributor Gearbox on blast furnace based on vibration signal detection is developed in chapter III, including selection of vibration acquisition points, hardware design of the circuit and selection of the components, as well as software development based on virtual instruments.The diagnostic program is proposed in the fourth chapter which includes the intelligent diagnosis based on neural network and analytical diagnosis based on signal processing in order to make use of the knowledge and experience of field experts. In order to detect early fault and achieve high diagnostic accuracy in noise conditions, intelligent fault diagnosis module extracts kurtosis parameter which is sensitive to faults in early stage in combination with the traditional wavelet package sub-bands energy parameter as diagnostic input, and uses distinct weights neural networks as the classifier. At the same time, fault analysis module is also developed.Diagnostic tests is conducted in the fifth chapter, certain roles for early fault diagnosis of kurtosis parameter as well as effective improvement of noise robustness of probabilistic neural network from a distinct weights method are verified.