Research Of Adaptive Decomposition Algorithm In The Fault Diagnosis Of Rolling Bearing

Big data era has arrived,the data is increasingly complex and changeable,nonlinear,non-stationary data analysis has gradually become a research hotspot in the field of statistics and information.As a branch of modern science,data analysis is widely used in mechanical engineering,electronic information,environmental engineering and intelligent manufacturing.With the continuous development of Chinaundefineds manufacturing 2025 strategy,machinery and equipment play an important role in industrial development.Rolling bearing is one of the key parts in mechanical equipment,and its running state is directly related to the operation of the whole machine.Because the adaptive decomposition algorithm has better data drive and self-adaptability to the data analysis,so the self-adaptive decomposition algorithm has better data-driven and self-adaptive characteristics.The self-adaptive decomposition algorithm is used to study the vibration signal of rolling bearing.The detection of bearing state and fault diagnosis and recognition can effectively reduce the fault incidence,which can ensure the normal operation of machinery and equipment.It is of great theoretical and practical significance to ensure the normal operation of mechanical equipment and avoid major accidents.This paper takes rolling bearing vibration signal as the research object and fault identification as the starting point,the method of combining adaptive decomposition with random forests algorithm is proposed to analyze the vibration signal data of rolling bearing.The main contents are as follows: Firstly,the theory of adaptive decomposition algorithm is summarized systematically,including empirical mode decomposition,ensemble empirical mode decomposition and variational mode decomposition.Secondly,according to the collected vibration data of rolling bearing,descriptive statistical analysis is carried out.Four kinds of data,such as normal state of rolling bearing,inner ring fault,outer ring fault and roller fault are decomposed by MATLAB software,and the adaptive data are decomposed respectively.Several intrinsic modal functions are obtained after decomposition.Especially in variational mode decomposition,threshold method is proposed to determine the number of decomposition functions,which makes the number of decomposition more scientific and reasonable.Then,the mean,variance,range,variation coefficient,fluctuation index,energy entropy and information entropy of each intrinsic modal function are calculated by SAS software,and the attribute feature sets of different types of data are constructed.Finally,three adaptive decomposition methods are combined with random forests algorithm to realize rolling bearing fault diagnosis.The empirical results show that the combination of variational mode decomposition and random forests algorithm is better than the other two algorithms,and can be used to diagnose rolling bearing faults more effectively.This provides a scientific basis for ensuring the normal operation of machinery and equipment and improving the economic benefit of enterprises,and widens the application field of adaptive decomposition method.