EFFECTS OF DAMPING, DURATION, AND SIGNAL-TO-NOISE RATIO ON COVARIANCE-BASED STOCHASTIC SUBSPACE IDENTIFICATION

Author

Pramila Shrestha, Southern Illinois University CarbondaleFollow

Date of Award

8-1-2022

Degree Name

Master of Science

Department

Civil Engineering

First Advisor

Tezcan, Jale

Abstract

AN ABSTRACT OF THE THESIS OFPramila Shrestha, for the Master of Science degree in Civil Engineering, presented on June 16, 2022, at Southern Illinois University Carbondale. TITLE: EFFECTS OF DAMPING, DURATION, AND SIGNAL-TO-NOISE RATIO ON COVARIANCE-BASED STOCHASTIC SUBSPACE IDENTIFICATIONMAJOR PROFESSOR: Dr. Jale Tezcan The modal parameters play an important role in determining the dynamic characteristics of a structure which are further used in analyzing and designing various civil engineering structures. Many challenges are faced by the civil engineers in obtaining the input data due to the size and complexity of the structures. It is not feasible to apply artificial forces in the structures and uneconomical to conduct laboratory tests for such huge structures. Various output-only methods have been developed so far to identify the modal parameters which use the response of the structure under ambient vibrations. The Stochastic Subspace Identification (SSI) method has been designed as an advanced and powerful method that is driven by output-only data. During the last two decades, the SSI method has become increasingly popular for its computational efficiency and accuracy. Further, compared to classical approaches, SSI has a better ability to identify the closely spaced modes. This method can identify the modal damping ratios and mode shapes along with the modal frequencies.In this study, Covariance Driven Stochastic Subspace Identification (SSI-COV) method has been used to study the modal parameters, i.e., the eigenfrequencies, modal damping ratios, and the mode shapes of a five-story building. The analytical simulation is carried out in MATLAB. The simulated data is used to enable how well the approach performs under various levels of measurement noise, damping, and duration. The traditional stabilization diagram is used to manually compute the number of modes which has been verified using the automatic identification of modes using the clustering stabilization diagram. The identified values are then compared with the true values to study the effects of different levels of measurement noise, duration, and damping using the SSI-COV method.