Sensor Placement for Damage Detection in Nonlinear Systems Using System Augmentations

Abstract

Currently, most sensor placement methodologies are focused on maximizing the controllability and observability of the monitored structure. Recently, there have been several sensor placement techniques proposed for damage detection. The work herein provides an integrated sensor placement and reduced-order health assessment approach that can be applied to both linear and nonlinear structures. The method uses the idea that damageable regions (hot spots) of the system are often known in advance and, therefore, the modes that are sensitive to changes in these hot spots should be the ones exploited for damage detection. Generally, sensors are placed near the hot spots. However, if that is not possible or additional sensors are being used, then a generalized effective independence distribution vector method is applied for the remaining sensors. The partial eigenvector information obtained from measurements is expanded to the full (model) space using the knowledge that damage is limited to the hot spots of the system. Modalbased damage detection methods, such as minimum rank perturbation theory, can then be used to solve for the damage. Alternatively, damage identification by hot spot projection is a novel option for damage detection presented herein. Nonlinear systems are handled by forming (higher dimensional) augmented linear systems that follow the same trajectory of the nonlinear system when projected onto the physical (lower dimensional) space. The sensor placement methodology for nonlinear systems is similar, but it requires that sensors be placed at the location of all nonlinearities as well as the hot spots. Damage can be detected using the multiple augmentations generalized minimum rank perturbation theory, previously developed by the authors, or by damage identification by hot spot projection. Numerical simulations of the methodology are presented for linear and nonlinear five-bay frame structures.

A multi-objective optimization approach to optimal sensor location problem in IGCC power plants

“…Singh and Hahn (17) developed a technique for sensor placement that could be applied to stable nonlinear systems such as binary distillation columns. D'Souza and Epureanu (18) developed an integrated approach for sensor placement and damage detection assessment that can be applied to both linear and nonlinear structures.…”

“…Different methodologies have been applied for solving optimal sensor placement problems. Some commonly used methods include optimality criteria based on the error covariance matrix (19)(20)(21)(22)(23) and observability matrix or Gramian (18,24). Importance has been given to relevant aspects such as measurement cost, sensor failure, redundancy and process information (25)(26)(27)(28).…”

Clustering of Sensor Locations Using the Effective Independence Method

“…In particular, sensor placement is considered one of the most important pretest steps in experimental/operational modal analysis [1][2][3][4]. Structural health monitoring techniques also incorporate optimal sensor placement (OSP) methodologies for different purposes: system identification, structural damage identification, and finite-element updating [5][6][7][8][9].…”

Energy-aware wireless sensor placement in structural health monitoring using hybrid discrete firefly algorithm

“…Numerous studies of the OSP problem have been conducted using a variety of placement techniques and criteria. Examples include the effective independence method , the effective independence method with driving point residue , the minimization modal assurance criterion (MAC) method , the kinetic energy method , the QR decomposition method , the information entropy‐based method , the damage detection‐based method , and the method minimizing the deviations of measured reduced model from the full model . These approaches focus on the significance of measurement data for structural condition evaluation; the performance of networks is disregarded.…”