Researchand Implementation Of Rolling Bearing Fault Diagnosis Algorithm Application Based On FPGA

With the development of science and technology, mechanical equipment developstowards the orientation of precision. The equipment should be maintained at times orthe machine might break down and lead to needless loss as a result. To settle thisproblem, we had better real-time monitor the running status especially the early failure,which can’t be done by people. It is important to real-time monitor the equipmentrunning status by utilizing instruments. Many novel methods and technologies areimplemented into mechanical equipment fault diagnosis. Most results of mechanicalequipment fault diagnosis theoretical research results while some others are based onPC and DSP processor, which do not has the advantage of processing speed. Therefore,the FPGAis used as processor in this thesis for its obvious high speed.The main purpose of the research in this paper is to exploit rolling bearing faultdiagnosis algorithm. FPGA realization method based on Hilbert transform and FFTtransform are studied. Multiple series of bearing failure data verify the reliability of thealgorithm.Firstly, the rolling bearing fault mechanism is introduced. Several kinds of rollingbearing failure modes and failure frequency characteristics as well as time domain ofthe vibration signal, frequency domain and modal analysis method are analyzed.Secondly, wavelet transform and Hilbert transform and fast Fourier transform arestudied. Based on Mallat algorithm, a wavelet transform with high computing accuracyis realized by FPGA. Then Hilbert transform is implemented in time domain byutilizing FIR filter and FPGA. The algorithm is validated after simulation. Algorithmstructure from the discrete Fourier transform (DFT) to the fast Fourier transform (FFT)is expatiated.4096point FFT is implemented and validated by a circuit with a core ofsingle path delay feedback switch (SDF). High speed of these three algorithms in FPGAimplementation provides real-time performance guarantee for the entire rolling bearingfault diagnosis algorithm.Finally, the reliability of the algorithm is verified through the comparison ofmultiple sets of data. Process fault bearing data of case Western Reserve University in the United State including the bearing inner ring, outer ring, ball three sets of data andcompare them with MATLAB calculation result. It can be deduced that the error is verysmall thus veracity and reliability of the algorithm is validated. Good result obtained byprocessing the bearing inner ring, outer ring and ball data with this algorithm illustratesthe reliability of the fault diagnosis algorithm as well.