The Research Of Experimental Modal Analysis Methods Based On Ambient Vibration

Experimental modal analysis based on ambient vibration is mainly to deal with the engineering problems from with we can hardly achieve the input signals, such as the operational modes of a flight vehicle during the flight or the dynamic behavior of a large-scale structure like an offshore platform, a bridge as well as a skyscraper on which we can't exert a heavy kick.There is a focus on application of the stochastic subspace identification method in this article. For its defect in application of engineering, it also studies auto-cross spectrum density method which is classical in engineering, and it can be used as a supplement.At first in this article, by looking into plenty of articles at home and abroad for reference, it gives an overview of ambient vibration modal analysis. It expounds the great importance of ambient vibration modal analysis to engineering, introduce the development course of the methods in this field as well as its current status. For the next, it studies the theory of auto-cross spectrum density method and stochastic subspace identification method. As a classical and practical method in ambient vibration experimental modal analysis, auto-cross spectrum density method replace the frequency response function with power spectrum and achieve its mode shape from transmissibility between the response signals and reference signal. By deducing the theory of this method, it comes to a conclusion that auto-cross spectrum density method is much the same as peak picking method based on frequency response function. And it also talks about the virtues and defects of the method. Stochastic subspace identification method is the major in this article, and this method is based on the theory of projection of row space, and also uses QR decomposition, SVD decomposition as well as least squares estimate to state space matrices of the discrete system so as to achieve the parameters of the dynamic behavior and it can achieve an accurate result. It deduces the formula of stochastic subspace identification method in details and writes out the program, and later verifies the accuracy and anti-interference by simulating a 2-order system in different conditions.