Phase Space Reconstruction Based On Improved CAO Algorithm And ACFPA-ELM Bearing Fault Diagnosis

Mechanical bearings often wear out due to wear and tear during continuous operation,and failures often occur.Therefore,a more complete system for fault diagnosis must be developed.Based on the analysis of the current status of bearing fault research,this paper proposes a research theory based on the improved CAO algorithm phase space reconstruction and ACFPA-ELM bearing fault diagnosis.Firstly,the various parameters of bearing failure are introduced.And on this basis,the method of solving the optimal embedding dimension for the CAO algorithm for writing parameters is improved.Combining the symbolic analysis method to obtain the maximum joint entropy.Through several sets of numerical simulation experiments and comparative analysis.It is verified that this method can better find the best embedding dimension.Then,on the basis that the reconstruction parameters of the phase space can be accurately determined,the basic theory of singular spectrum is introduced.Take it as the characteristic quantity of the system and analyze its anti-wave ability and anti-noise ability.On this basis,a theoretical analysis of its characteristics.From the perspective of functional analysis,its characteristic space and noise platform are explained.Then it analyzes the vibration signal emitted by the industrial rolling bearing when it runs under different structural damage.So verifying the validity of its theory.Next,this paper proposes a classification and identification method based on improved flower pollination algorithm to optimize ELM,and introduces the structural composition,advantages and disadvantages of ELM;On this basis,this article also proposes tent chaos search.This chaotic mapping method based on reverse learning can improve the distribution quality of the initial gametes.Through comparative experiments,the efficiency and accuracy of the proposed neural network classification of ACFPA-ELM in a single hidden layer are confirmed.Finally,the experimental data of rolling bearing faults of the American university experimental platform and the measured data released by Baogang SP1580 rolling mill are applied to the method studied in this paper to obtain the parameters of phase space reconstruction.Perform chaos singular spectrum analysis to extract the characteristic quantity of the chaotic signal of the bearing.Using our optimized flower pollination algorithm to optimize ELM to diagnose and identify its mechanical faults.Through comparative analysis,the results show that the method proposed in this paper meets our expectations.