Design Of Rolling Bearing Fault Diagnosis System Based On Adaptive Resonance Demodulation

In today’s prosperous railway transportation business in China,the safe and long-term operation of locomotives is receiving more and more attention.In the process of locomotive running,if there is a fault condition,but it fails to alarm and release the fault in time,it may cause serious consequences,and in extreme cases,it may even lead to a major accident that destroys the car and kills people.As one of the key components of locomotive operation,whether the rolling bearing can operate normally and safely is one of the important factors for the safety and stability of the whole locomotive.If real-time diagnosis of rolling bearing failure can be realized,the probability of locomotive accidents can be greatly reduced.Based on this,this paper starts from the research point of rolling bearing fault diagnosis,selects the fault diagnosis technology with bearing vibration signal as the analysis data according to the causes of the common fault of rolling bearing in locomotive,carries out in-depth analysis for the vibration signal acquisition and fault feature extraction of rolling bearing under the hardware environment of FPGA,and completes the design research of hardware system.The main contents of the research in this paper are as follows.First,the overall scheme of the system is determined.Starting from the common failure forms of rolling bearings,the overall framework of the system in this paper is designed according to the causes of their appearance and the vibration mechanism of different failures.The main programming software used is Vivado,a development tool from XILINX,USA.The main control chip of the system is selected according to the system requirements and related specifications.The software development adopts the programming method of Verilog language programming and IP core use,thus the system is not only convenient to develop the program,but also can meet the real-time requirements for accurate acquisition and processing of bearing vibration signals.Secondly,for the shortcomings of poor adaptive capability of traditional resonance demodulation algorithm,this paper proposes the use of adaptive resonance demodulation algorithm based on LMS adaptive filter as the core algorithm used as vibration data processing in the design scheme of this paper.The problem of poor adaptive capability of the resonance demodulation algorithm is solved by using the LMS adaptive filter to adaptively noise reduce the original bearing vibration signal,and then extracting the fault characteristics according to the resonance demodulation algorithm.Thirdly,the FPGA design and implementation of the data acquisition module and the adaptive resonance demodulation module of the system are presented.In the data acquisition module,the selected sensor and sampling chip are introduced,and the hardware circuit connection,sampling frequency selection and program flow design of the sampling module are analyzed;in the adaptive resonance demodulation module,the selected LMS-based adaptive resonance demodulation algorithm is implemented in hardware,and the LMS adaptive filter and the resonance demodulation algorithm are ported on the Vivado development software respectively.The hardware implementation of the selected LMS-based adaptive resonance demodulation algorithm is carried out,the LMS adaptive filter and resonance demodulation algorithm are ported on Vivado development software,and the two algorithms are combined to realize the hardware adaptive resonance demodulation system.Finally,the data acquisition module and adaptive resonance demodulation module of the system were integrated to build the final rolling bearing fault diagnosis system,and the system characteristics of the overall system were tested.The test results showed that the system designed in this paper saved the consumption of resources and obtained a high system response speed;after that,the system was downloaded to the FPGA board used in this paper and experiments were conducted on the bearing test bench.Through the experiments,it is proved that the rolling bearing fault diagnosis system can be able to determine whether there is a fault in the bearing and can get the data containing the bearing fault frequency information for subsequent fault diagnosis.