Automated Operational Modal Parameter Extraction With Stochastic Subspace Identification

Operational modal analysis(OMA)is a significant technique to obtain the dynamic behavior of structures.Advantages,such as effective operation,low cost,make OMA become an indispensable tool for product development,validation,usage and maintenance in aerospace,automobile and other field.The core objective of OMA is to obtain the accurate structural modal parameters.However,on the one hand,the mathematical poles generated by the over-specified of model order and the noise poles caused by measurement noise disturb the determination of the modal parameters hugely during the process of the identification,on the another hand,picking up final parameters manually is not only time-consuming,laborious,but also influenced easily by subjective factors,which lead to the final result unreliable and inaccurate.To surmount the deficiencies,an automatic modal parameters identification algorithm based on stochastic subspace identification is proposed.Firstly,in this paper,Cov-SSI,Comb-SSI,Data-SSI were applied to 5 degrees of freedom mass-spring-damping model and rectangle plate.Moreover,the obtained modal parameters were compared based on stabilization diagram and FDD method from 4factors,natural frequency,damping ratio,shape and computational efficiency.The results showed that the natural frequency,damping ratio,shape of the two models identified by the three methods were highly consistent.However,in computational efficiency,Cov-SSI was the fastest,Data-SSI was the slowest.Subsequently,a spurious poles rejection algorithm based on reconstructed Hankel matrix was presented,and applied to the original poles of the 5 degrees of freedom mass-spring-damping model and rectangle plate.By comparing with the results obtained by traditional stabilization diagram,conclusion could be drawn that the spurious poles were nearly eliminated by the algorithm.Moreover,the rejection results was better than those of the traditional stabilization diagram technique.Finally,an automatic modal parameter identification method in light of hierarchical clustering was established and conducted to obtain the modal parameters of the 5degrees of freedom mass-spring-damping model and rectangle plate.The results showed the method could peak up the modal parameters of the two models accurately.Then,operational modal analysis was performed to a break disc with doublet modes to obtain the modal parameters utilizing the automatic modal parameter identification algorithmbased on stochastic subspace identification.Comparing with the results obtained by B&K OMA software,a conclusion could be drawn that the method could identify the modal parameter of the break disc accurately.Therefore,the technique suited for cyclic symmetric structure.