Early Fault Diagnosis And Life Prediction Of Rolling Bearing

Rolling bearing is one of the core components of rotating machinery,which is widely used in the field of industrial engineering,so its normal operation is one of the important links to ensure the safe and efficient operation of machinery and equipment.However,due to the complexity of its working environment,long-term high load,high fatigue environment,coupled with not regular repair and maintenance checks,rolling bearings in large machinery and equipment has become one of the most prone to failure components.Because timely fault diagnosis monitoring and regular life prediction of rolling bearings has also become an essential link,for improving the safety of machinery and equipment,improve work efficiency,to ensure the stability of equipment has a vital role and significance.The paper focuses on the analysis and study of rolling bearings,the realization of fault diagnosis and life prediction,through the extraction of indicators that can react to bearing fault characteristics,and then life prediction,and a good indication of the evolution of rolling bearing degradation trends,the specific work content is as follows:(1)Introduces the principle of denoise of CEEMD(complete overall experience modal decomposition)algorithm,and carries out experiments to simulate the effectiveness and superiority of the CEEMD algorithm along.(2)The principle of wavelet threshold denoise is introduced,and according to the actual needs,the improved wavelet function is used,combined with the CEMD algorithm and the ant colony algorithm optimization to improve the wavelet threshold to combine noise removal.(3)This paper introduces the life prediction model to achieve effective life prediction,carries out simulation experiments,compares the three algorithms of GA,PSO and GA-PSO to find excellent performance,and uses GA-PSO to optimize the extreme learning machine.(4)Combined with the data of PRONOSTIA experimental platform to complete the experimental verification,pre-processing the selected data,combined with the extreme learning machine to complete training and prediction,to achieve the prediction of bearing life,combined with specific indicators to verify the validity and accuracy of the model.