Structural Modal Parameters Identification Method Based On The Probability Histogram

Structural Modal Identification use measured data of the system to estimate the unknown modal parameters. The results can be the initial data of structural model updating, damage identification, damaged structure assessment, structural control, etc. This technology is more and more widely used in civil engineering in recent years, kinds of modal identification is close to complete. But it’s still hard to determine the order of system and eliminate spurious modal in practical engineering. To avoid these problems, a new method is proposed in this dissertation. Probability Histogram is combined some typical modal identification method in time-domain and frequency domain to realize automatic acquisition of structural modal. and the work presented in this dissertation contains the following aspects:Firstly, the basic principal and process of Poly MAX and SMI SSI are introduced. Probability Histogram combined these method are proposed. The modal parameters identified by system of different order are clustered into several categories by using certain similarity criteria, automatic acquisition of structural modal is realized. The index of Identification Probability is defined. The clustered modal parameters with large Identification Probability are chosen to be the final identification results. With calculation and discussion of the numerical examples, the probability histogram have adventage of the operability and anti-noise capacity over staple stabilization diagram.Secondly, composite of Natural Excitation Technique, Stochastic Subspace Identification and Probability Histogram, the NEx T-SSI-PH method is proposed. Choose a particular sensor as reference. Make the cross-correlation function between the reference and all sensors as the impulse response function to compose the Hankel matrix to do SSI-PH. Then cluster the results of SSI-PH with different reference. The effect of identification results are obtained to improve.Thirdly, composite of of Reference-based Stochastic Subspace Identification and Probability Histogram, the SSI/ref-PH method is proposed. Choose some sensors with preferred quality as references to compose the Hankel matrix to do SSI-PH. Then cluster the results of SSI-PH with different reference. The effect of identification results are obtained to improve.Fourthly, the SSI combined two stage Probability Histogram method is proposed. The second stage is to change coefficient of rows of the Hankel matrix to reconstruct the matrix. The effect of identification results are obtained to improve.