Electrical impedance tomography (EIT) is recently demonstrated to be viable for damage localization over a spatial area. The algorithm reconstructs the spatial conductivity distribution within a defined boundary via boundary voltage measurements. To solve this inverse problem, a finite element model (FEM) conforming to the interrogated geometry is required. Previous studies on identifying a center crack’s propagation suggests that an FEM-updating strategy may help identify both the existence of a crack and the plastic zones formed around the crack’s tips. In this paper a data-driven algorithm is applied to automatically update the FEM. The selforganizing map algorithm is adopted to categorize the reconstructed conductivity data, tracing the boundary of the crack to be updated as material-absence. The EIT results from the updated FEM model are able to identify damage location and damage severity with desired accuracy.
Carbon nanotube (CNT)-embedded polymer solution can be inkjet-printed into a thin sheet consisting uniform morphology and consistent electrical properties. When subjected to a loading scheme, the thin film’s inherent electrical property changes in tandem with the deformation. This unique property makes CNT thin films the appropriate candidate for strain sensing applications. Recent studies on characterizing the gage factor of CNT-embedded thin films are limited to learning the materials resistance change along the loading direction only. However, research interests on strain measurement of a structure have shifted from point-based interrogation to spatial strain-state monitoring. In this study an attempt to characterize its anisotropic resistivity was carried out. The resistivity-strain constitutional relation of an inkjet-printed CNT thin film is established based on theories for semi-conductive materials. The 2D elastoresistivity properties were characterized via the Montgomery method. It is observed that the change in resistivity in both directions are exhibiting linear trend to their strains in the same direction, but the thin film is more sensitive toward compressive strains. The final result of this study has inspired future research on fully characterizing the thin film’s elastoresistivity under different loading situations, and the way to characterize shear elastoresistivity shall also be reconsidered.
Electrical impedance tomography (EIT) has been recently applied as a structural health monitoring (SHM) technique to many different kinds of structures. In short, EIT is an algorithm that reconstructs the spatial conductivity response of a conductive body using only voltage measurement along its boundaries. For a conductive structure with its electrical properties being sensitive to damages and/or strains, mapping the distribution of its conductivity allows one to obtain its corresponding damage and/or strain distribution. To date, the EIT inverse problem has been solved using different techniques. This study compared the performance of two different approaches using four evaluation criteria. The first technique is based on EIDORS, which is an open-source EIT solver based on the maximum a posteriori (MAP) approach. It can rapidly, using a one-step linear approach, evaluate the relative impedance change of a given region when a baseline measurement (i.e., the response collected under its initial state) is provided. The second approach is a two-step iterative shrinkage thresholding (TwIST) method that compresses a signal’s sparsity in preserving sharp edges of an image. Both methods were evaluated using a 16-electrode 2D square shape with a simulated “point” damage at different locations. The evaluation results suggested that TwIST outperforms MAP in terms of the resolution and accuracy of the reconstructed results, and MAP wins over TwIST in causing minor shape deformation and less ringing. Results from both methods exhibit position-dependency. These results are significant in promoting EIT becoming a powerful technique for in situ health monitoring.
KEYWORDS: Data modeling, Structural health monitoring, Algorithms, Stochastic processes, Matrices, Systems modeling, System identification, Wind energy, Process modeling, Wind turbine technology
Structural health monitoring (SHM) relies on collection and interrogation of operational data from the monitored
structure. To make this data meaningful, a means of understanding how damage sensitive data features relate to the
physical condition of the structure is required. Model-driven SHM applications achieve this goal through model
updating. This study proposed a novel approach for updating of aero-elastic turbine blade vibrational models for
operational horizontal-axis wind turbines (HAWTs). The proposed approach updates estimates of modal properties for
spinning HAWT blades intended for use in SHM and load estimation of these structures. Spinning structures present
additional challenges for model updating due to spinning effects, dependence of modal properties on rotational velocity,
and gyroscopic effects that lead to complex mode shapes. A cyclo-stationary stochastic-based eigensystem realization
algorithm (ERA) is applied to operational turbine data to identify data-driven modal properties including frequencies and
mode shapes. Model-driven modal properties are derived through modal condensation of spinning finite element models
with variable physical parameters. Complex modes are converted into equivalent real modes through reduction
transformation. Model updating is achieved through use of an adaptive simulated annealing search process, via Modal
Assurance Criterion (MAC) with complex-conjugate modes, to find the physical parameters that best match the
experimentally derived data.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.