Singular Spectrum Analysis: A Solution for Modal Parameter Identification from Force-embedded Acceleration Data of Linear Dynamic Systems

This study addressed a problem when identifying modal parameters of a linear dynamic system resulting from forced acceleration response. The problem is attributed to an input force imposed on its impulse acceleration response. With no knowledge of the input force, use of the acceleration response can thus pose a problem on the identification accuracy of the modal parameters. This paper is aimed at proposing a method that can solve the problem of the force-embedded acceleration data in the time-domain identification. The best solution for the problem is to anticipate an effective separation technique that can separate the unknown input force from the forced acceleration data. In this regard, a new robust and adaptive approach based on subspace analysis using the singular spectrum analysis is proposed. The use of the singular spectrum analysis decomposes a time series into its component parts and then reconstructs the time series by leaving noise components behind. The performance of the singular spectrum analysis to the acceleration response was demonstrated via simulation of a single-degree-of-freedom linear dynamic system excited by white Gaussian noises. Results show that the use of the singular spectrum analysis can effectively diminish the noises in the correlation-related function of the forced acceleration response.