Data Acquisition For Vibration Signal Of Bearing And The Research For Methods Of Fault Detection

With the development of intelligent technology, fault diagnosis of mechanicalequipment has received increasing attention. An intelligent as well as integrated systemis needed for monitoring the operating state of mechanical equipment especially whenthey are running under a closed environment whose components are quite inaccessibleto the technical people. In most cases, faults occur due to the abnormal running of themechanical components inside the equipment. Bearings play a very important role inmechanical equipment. However, they are prone to be out of failure and suffer fromrelatively higher failure rate. Besides that, informative signal of faulty bearings is hardto detect and the fault condition is thus difficult to determine especially when in earlystage. While the bearings failure may result in abnormality or even breakdown of thewhole system fast. Therefore, it is necessary and imperative to monitor the running stateof rolling bearings to ensure the safe operation of the mechanical equipment.The main works of this paper are listed as follows:(1) The design of the hardware.After the selection of sensor type, the power supply module, the communication moduleand the storage module are designed respectively. The acquisition device of vibrationsignals based on STM32microprocessor has been accomplished in this paper.(2) Thefaults diagnosis of vibration signals. This paper introduces the signal processingmethods in time domain, frequency domain and time-frequency domain respectivelyand presents a comparative analysis of their performance. Traditional Winger-Villedistribution (WVD) method is utilized in this paper and the simulation results show itsadvantages. In the meanwhile, its disadvantages are pointed out. When dealing withmulti-component signals, the WVD method will lead to the presence of cross-terms,resulting from interactions between signal components (auto-terms). These cross-terms may cause an erroneous interpretation. To solve this problem, an improved WVDmethod based on Short Time Fourier Transform (STFT) is proposed here. In thismethod, a very short STFT time window moves along time to slice the vibration signalso as to eliminate the uninformative cross-terms while remaining the informativeauto-terms. The STFT-WVD joint distribution will remain the satisfactorytime-frequency resolution as well. Finally, the proposed method is used to analyze thesimulation signals and the bearing vibration signals in practical engineering and theexperimental results show its feasibility and effectiveness.