Research On Fault Diagnosis Method Of Large-scale Mill

Mill machinery is one of very wide equipment for industrial application, because of the complexity of its structure, and the bad work condition, Fault diagnosis Requirements and difficulty of the mill must be higher than conventional equipment. With the continuous evolution of vibration detection, signal processing and other related technology, it has became an important research direction of fault diagnosis based on the vibration signal detection, processing and analysis. The EMD method is a new signal processing method, after proposed, it has been developing rapidly, and has been widely applied in the field of fault diagnosis.Based on EMD method, this paper proposed several time-frequency analysis method, and utilized these methods in real fault diagnosis of mill,identified the fault accurately, and achieved good results. Firstly, Introduced in detail the theory of EMD method, including the concept of instantaneous frequency and the intrinsic mode functions in EMD theory, and introduces the process of EMD, then made improvement against the end effects problem and false modal problem of EMD, and verified with two simulation signals. Secondly, Based on EMD method, proposed the Energy Operator Demodulation Method, Frequency weighted Time-Frequency Entropy method, and Local Hilbert spectrum analysis method, For the energy operator demodulation method, through the comparison with Traditional Hilbert demodulation method, verified the superiority of the method; For frequency weighted time-frequency entropy method, is improved base on time-frequency entropy method; For the local Hilbert spectrum analysis method, including the Hilbert spectrum and Hilbert marginal spectrum, and has made the improvement to it; For the above three methods, respectively using the simulation signal and simulation experiments show that these methods are effective in signal analysis. Finally, Using the above three methods based on EMD method diagnosis the fault of mill, including the gear, bearing in the speed reducer, and the mill grinding roller, three parts. The research results show that, the Energy Operator Demodulation Method, Frequency weighted time-frequency entropy method,and Local Hilbert spectrum analysis method, Based on EMD method, have good effect in the fault diagnosis of each part of mill.