Time-frequency Analysis With FPGA And Its Application In Fault Diagnosis

In the actual production, using vibration signal analysis is one of the methods that realize the real-time monitoring and fault diagnosis on machinery equipments condition. Because Fourier transform can not express the time detail information of related frequency components, hence, in order to obtain the more accurate characteristic defect frequency, it is necessary to use time-frequency to analyze the non-stationary signals in the practical project. In the traditional hardware implementation, signal processing module mainly takes Application Specific Integrated Circuit or Digital Signal Processing chip as the core. For the former, its external circuit is complex, besides, its expansibility is very poor, the latter needs large number of redundant operations, specific external interfaces and RAM, which limit the processing speed, so it can not meet the requirement of real-time analysis. For solving these problems, the thesis researches into time-frequency analysis with Filed-Programmable Gate Array (FPGA), the details are studied as follows:First of all, according to the theory of time-frequency, the thesis discusses several traditional time-frequency methods, such as short-time Fourier Transform, Wavelet Transform, Wigner-Ville Distribute and Hilbert-Huang Transform, studies their realization algorithm, which establishes the theoretical basis for the implementation of time-frequency FPGA-based.Next, the thesis bases on the basic theory of FPGA and its related techniques, realizes the Fast Fourier Transform of the practice signal based on the design flow of FPGA, which provides the thought and method for time-frequency with FPGA, meanwhile, proves the correctness of the analysis results.Then, the thesis researches into STFT Goertzel recursive algorithm and its filter structure, in the premise of selecting FPGA device of Altera and its development tool Quartus II, takes the implementation of STFT as an example, puts forward the implementation of time-frequency FPGA-based, and uses FPGA to realize time-frequency. The design scheme has the advantage of flexible programming, and it over comes the lack of implementing slowly, which exists in the software processing. Finally, the thesis uses time-frequency with FPGA to process the rolling bearing fault signal, and the design can realize its fault diagnosis availably. Meanwhile, the analysis results show the correctness of the design scheme.Through the further research, the thesis gives the design of time-frequency with FPGA, and uses the design to analyze the simulation signal and the practical signal. According to the analysis results, the design is reliable and accurate, and it has better real-time. The study has certain reference value for setting up small-scale fault diagnosis system and developing fault diagnosis instruments.