Time-Varying System Identification of High Voltage Switches of a Power Substation with Slide-Window Least-Squares Parameter Estimations


This paper is aimed at identifying the time-varying parameters and ultimate behavior of high voltage switch structures based on a series of full-scale shake table tests with harmonic excitations. Each structure involves a mechanical device for switch-on and switch-off, a friction-based switch, and three porcelain pillars. To identify the structural properties over time, a novel slide-window least-squares estimation method is developed. Each time-varying parameter is firstly approximately expressed by a simple polynomial or exponential function with time in a short slide-window. The time-invariant coefficients of the polynomial or exponential function are then estimated using a least-squares method. Finally, the time-varying parameters can be simply calculated from the estimated polynomial or exponential function. The proposed method is validated by simulated one- and two-story buildings with three kinds of time-varying parameters (stiffness varying abruptly, gradually, and periodically) under earthquake excitations. The application of the proposed method to the tested switch structures demonstrated that the time-varying fundamental frequency of the structures decreased from 7.5 to 6.5 Hz near resonance, which is consistent with the shake table test observations under an excitation of 1.27 and 2.54 mm in stroke. During the shake table tests, all switch structures failed at the bottom of the mechanical device under cyclic loading.


Civil, Architectural and Environmental Engineering

Keywords and Phrases

Earthquake Excitation; Fundamental Frequencies; High Voltage Switches; Least Squares Methods; Least-Squares Parameter Estimations; Time Varying Parameter; Time-Varying System Identifications; Exponential Functions; Parameter Estimation; Polynomials; Time Varying Control Systems; Least Squares Approximations

International Standard Serial Number (ISSN)


Document Type

Article - Journal

Document Version


File Type





© 2013 IOP Publishing Ltd., All rights reserved.

Publication Date

01 Jun 2013