Research On Mechanical Fault Diagnosis Method Based On Compressed Sensing

Rotating machinery is an important production equipment in modern industrial production.It is of great engineering significance to use condition monitoring technology to monitor the running condition of machinery and equipment in real time,find early faults in time and take effective measures to solve the faults,which can improve production efficiency and avoid major shutdown loss.This paper aims at a series of problems such as large amount of data,data transmission,analysis and storage produced by continuous condition monitoring of equipment.In this paper,motor and bearing,the key components of rotating machinery,are taken as the research objects and the methods based on sparse representation and compressed sensing are adopted to carry out the research.The following aspects are mainly studied:1)Described the background and significance of the subject research,described the research status of compressed sensing theory in detail,and summarized the research achievements of compressed sensing in the field of fault diagnosis.2)The theory of compressed sensing is introduced from three aspects of signal sparse representation,compressed sampling and signal reconstruction,and the sparse structure preservation property of signal is expounded.3)Taking industrial robot as the research object,aiming at the problem that the low sampling period of servo controller cannot provide high-frequency information for the state recognition of high-precision servo control system,the sparse signal representation and sparse structure preservation property were studied,and a high-resolution signal reconstruction method based on sparse structure preservation was proposed.Based on the sparse structure preservation principle of homologous signals,the objective function of high-resolution signal reconstruction is established to realize the efficient reconstruction of low-resolution signals.The effectiveness of the reconstruction method is verified through experiments and simulations,which provides support for the subsequent fault diagnosis and health assessment of industrial robots.4)Taking rolling bearings as the research object,the bearing compression feature extraction and fault diagnosis of acoustic emission signals were studied.Compressed sensing and particle swarm optimization were combined to enhance the weak information of compression features.The accuracy and effectiveness of the proposed method were verified by experiments.5)The bearing condition monitoring system was developed based on LabVIEW,and the research method in this paper was applied to the system to realize the functions of data acquisition,storage,analysis and trend prediction.Meanwhile,the compression feature enhancement method proposed was applied to the state trend function to realize the evaluation of bearing condition.