OPERATIONAL MODAL ANALYSIS USING CORRELATION-BASED ARMA MODELS

This paper develops a correlation-based ARMA model for modal parameter identification. Compared to the traditional ARMA algorithm, the correlation-based ARMA algorithm employs the least mean squares (LMS) to determine the optimal order of the ARMA model, and adopts the correlation function of the dynamic response as the input signal. Using LMS, the accuracy of the identified results is optimized at a low computational cost, and the results are also insensitive to noise since the correlation function is used. Numerical simulations show that this algorithm is capable of identifying rather close modes with great accuracy and time efficiency even when highly noisy signal is present. The correlation-based ARMA algorithm is proven to be an efficient, reliable and robust tool for the output-only modal identification.