Research On Fault Detection Of Rolling Bearing Based On Wireless Sensor

Wireless communication technology is developing at a high speed,and modern society is also a society with fierce competition in science and technology.Under the influence of science and technology and modern industry,mechanical equipment is gradually becoming larger and more integrated and sophisticated,which is of great significance to the cost performance of products,the protection of environmental resources and the improvement of economic benefits.However,this trend has led to different machinery.The relationship between devices and the different components of the same mechanical device becomes more complicated and coupled to each other to form an inseparable whole.This puts higher requirements on the real-time and accuracy of mechanical equipment fault monitoring.Once a small component in a large equipment fails during operation,it will have a great impact on the entire production line,and economic losses are immeasurable and even have catastrophic consequences.At the beginning of the twentieth century,bearings were widely used as the core components of rotating machinery,but this also led to its main source of mechanical failure.This paper mainly uses the bearing as the detection component to diagnose the fault of the large machine,and accurately and effectively detects the bearing fault type and design the real-time monitoring software system.In the early fault diagnosis of rolling bearings,the fault characteristics are extremely weak,and it is easy to be disturbed by noise,so it is difficult to identify the fault.In this paper,a time-frequency diagram based on improved wavelet transform and a convolutional neural network(CNN)rolling bearing are proposed.A new method for weak fault diagnosis.Firstly,the acceleration vibration sensor is used to collect the data of the rolling bearing to obtain the sound vibration signal,and then the continuous wavelet transform is performed on the vibration signal to obtain the time-frequency diagram.The wavelet coefficients corresponding to each frequency are filtered by the autocorrelation algorithm from the time-frequency diagram.In addition to noise interference and extracting periodic fault components;finally,the Hilbert transform is used for envelope demodulation to obtain the fault feature frequency,and the processed time-frequency map is input as the feature map,and the early fault category is judged by training the CNN model.The method can automatically extract the time-frequency domain characteristic parameters and exhibit high accuracy and stability in machine fault detection.This paper designs a complete rolling bearing fault detection system software based on LABVIEW virtual instrument technology.The software mainly includes configuration acquisition parameters,startup data acquisition and storage module,signal time-frequency analysis module,CNN model import and result display module.The system can accurately identify the type of fault,and has practicality,stability and portability.