Almutawa, Jaafar (King Fahd University of Petroleum and Minerals)

Diagnostic Subspace Identification for the Dynamical Structural Models

Scheduled for presentation during the Contributed session "Beams and Flexible Structures" (WeAT6), Wednesday, October 23, 2013, 11:55−12:15, Room 134

6th Annual Dynamic Systems and Control Conference, October 21-23, 2020, Stanford University, Munger Center, Palo Alto, CA

This information is tentative and subject to change. Compiled on April 26, 2024

Abstract

Recently more attention has been given to the application of the so called subspace system identification methods to the dynamical structural models. In fact, the dynamical structural models can be written in the form of stochastic state space model. This paper propose several diagnostic techniques for the state space model fitting in the subspace system identification algorithms framework by deleting observations from the data and measuring the change in the estimates of the parameters. This method is considered for measuring the influential subsets in the state space model. We generalize the Welsch statistics in order to be applicable to the state space model to measure the effect of adding more variables to the model. Furthermore, a new algorithm to detect the outliers (damage detection and health monitoring) for the structural model has been developed. Moreover, we propose a Cook’s distance to identify the influential outlying cases. It also shown that the diagnostics based on the innovations variance are much clearer and more sensitive than those for the coefficients. A Monte Carlo simulation of the vibrating structure model demonstrated the effectiveness of the proposed algorithms and there ability to check the validity of the model, detect outliers.