Multi-channel output-only identification of an extendable arm structure under random excitation: A comparison of parametric methods

The problem of multi-channel output-only identification of the time-varying dynamics of an extendable prismatic arm structure under random excitation is considered. The identification is based on three simultaneously measured non-stationary vibration response signals obtained during a single experiment in which the arm extends from its fully retracted to its fully extended position. Non-stationary Functional Series Vector Time-dependent AutoRegressive (FS-VTAR) and AutoRegressive Moving Average (FS-VTARMA) methods are used, while critical comparisons with corresponding Pseudo-Linear Regression (PLR-VTAR and PLR-VTARMA) methods employing recursive models are also made. The superiority of the Functional Series methods, and particularly of the FS-VTARMA method, in capturing the time-varying dynamics is demonstrated.