Identification of nonlinear dynamic systems using vibration measurements is crucial for efficient and reliable damage
detection in structural health monitoring and control system design. Because of the complexity of control devices, it is
usually difficult to model the nonlinear control devices with enough accuracy in a parametric form. In this study, a
multi-storey steel-frame model structure equipped with magneto-rheological (MR) dampers, which were employed to
introduce nonlinear phenomena to the model structure, was modeled with a neural network in a nonparametric way.
Corresponding to the availability of dynamic response measurements, two different network models were proposed to
predict the vibration response of the nonlinear model structure. Raw dynamic response measurements of the model
structure under a certain impulse excitation was employed to train the two neural network models and the generality of
the trained neural network models were validated in the form of forecasting the raw test dynamic response measurements
of the model structure under other impulse excitation conditions. Results show that two neural network models provide a
reliable way for the modeling of nonlinear dynamic structures and present a useful way for the control system design of
engineering structures equipped with nonlinear control devices.
Advances in optic fiber sensing technology provide easy and reliable way for the vibration-based strain measurement of
engineering structures. As a typical optic fiber sensing techniques with high accuracy and resolution, long-gauge Fiber
Bragg Grating (FBG) sensors have been widely employed in health monitoring of civil engineering structures. Therefore,
the development of macro strain-based identification methods is crucial for damage detection and structural condition
evaluation. In the previous study by the authors, a damage detection algorithm for a beam structure with the direct use of
vibration-based macro-strain measurement time history with neural networks had been proposed and validated with
experimental measurements. In this paper, a damage locating and quantifying method was proposed using modal macrostrain
vectors (MMSVs) which can be extracted from vibration induced macro-strain response measurement time series
from long-gage FBG sensors. The performance of the proposed methodology for damage detection of a beam with
different damage scenario was studied with numerical simulation firstly. Then, dynamic tests on a simply-supported steel
beam with different damage scenarios were carried out and macro-strain measurements were employed to detect the
damage severity. Results show that the proposed MMSV based structural identification and damage detection
methodology can locate and identify the structural damage severity with acceptable accuracy.
A neural networks-based structural identification method using absolute acceleration without mode shapes and frequency
extraction is proposed and validated with vibration absolute acceleration measurements from shaking table test of a
two-storey frame structure. An acceleration-based neural network modeling for acceleration forecasting and a parametric
evaluation neural network for parametric identification are constructed to facilitate the whole identification process.
Based on the two neural networks and by the direct use of absolute acceleration measurement time histories of the object
frame structure under base excitation, the inter-storey stiffness and damping coefficients of the frame structure are
identified. The identified results by the proposed methodology are compared with them by solving eigenvalues equation.
Results show that the structural stiffness and damping coefficients identification accuracy is acceptable and the proposed
strategy can be a practical tool for model updating and damage detection of engineering structures.
It is difficult to obtain dynamic response measurement of a whole structure in reality, development of structural
parametric identification methodologies using spatially incomplete dynamic response measurement is critical for
performance evaluation and the realization of infrastructure sustainability. A general structural parametric identification
methodology by the direct use of free vibration acceleration time histories without any eigenvalue extraction process that
is required in many inverse analysis algorithms is proposed. An acceleration-based neural network (ANN) and a
parametric evaluation neural network (PENN) are constructed to identify structural inter-storey stiffness and damping
coefficients using an evaluation index called root mean square of prediction difference vector (RMSPDV). The
performance of the proposed methodology using spatially incomplete acceleration measurements is examined by
numerical simulations with a multi-degree-of-freedom (MDOF) shear structure involving all stiffness and damping
coefficient values unknown. Numerical simulation results show that the proposed methodology is a practical method for
near real-time identification and damage detection when several seconds of spatially incomplete dynamic responses
measurements are available.
A substructural identification methodology by the direct use of acceleration measurements with neural networks is proposed. The rationality of the substructural identification methodology employing a substructural acceleration-based emulator neural network (SAENN) and a substructural parametric evaluation neural network (SPENN) is explained. Based on the discrete time solution of the state space equation of the substructure, the theory basis for the construction of SAENN and SPENN is described. An evaluation index called root mean square of prediction difference vector (RMSPDV) corresponding to acceleration response is presented to evaluate the condition of object structure. The performance of the SAENN for acceleration forecasting and SPENN for parametric identification is examined by numerical simulations with a substructure of a 50-DOFs shear structure involving all stiffness and damping coefficient values unknown. Based on the trained SAENN and the PENN, the inter-storey stiffness and damping coefficients of the substructure are identified. Since the strategy does not require structural modes or frequencies extraction, it is computationally efficient, thus providing a possibly viable tool for structural identification and damage detection of large-scale infrastructures.
In this paper, comparisons are made between the performances of two kinds of distributed sensors, Electric Time Domain Reflectometry (ETDR) cable sensor that is based on the propagation of electromagnetic waves in an electrical cable and Brillouin Optical Time Domain Reflectometry (BOTDR) optical sensor that is based on the propagation of optic pulses and Brillouin scattering that occurs when light is transmitted through the optic fiber. A cable sensor was mounted near the surface of the 80% scale beam-column reinforced concrete assembly that was loaded cyclically until the shear failure occurred. The embedded depth was 0.5 inches. At the same time, a fiber optic sensor was mounted on the surface of the assembly with two installation procedures called Point Fixation (PF) Method and Overall Bonding (OB) Method to measure the strain distribution. Both BOTDR and ETDR sensors were subjected to tension and compression in one loading cycle. Strain distributions obtained from the ETDR and BOTDR sensing systems under different cycle loadings were compared with each other. They were also compared with those measured from the traditional strain gauge.
A structural parametric identification strategy based on neural networks algorithms using dynamic macro-strain measurements in time domain from a long-gage strain sensor by fiber optic sensing technique such as Fiber Bragg Grating (FBG) sensor is developed. An array of long-gage sensors is bounded on the structure to measure reliably and accurately macro-strains. By the proposed methodology, the structural parameter of stiffness can be identified. A beam model with known mass distribution is considered as an object structure. Without any eigenvalue analysis or optimization computation, the structural parameter of stiffness can be identified. First an emulator neural network is presented to identify the beam structure in current state. Free vibration macro-strain responses of the beam structure are used to train the emulator neural network. The trained emulator neural network can be used to forecast the free vibration macro-strain response of the beam structure with enough precision and decide the difference between the free vibration macro-strain responses of other assumed structure with different structural parameters and those of the original beam structure. The root mean square (RMS) error vector is presented to evaluate the difference. Subsequently, corresponding to each assumed structure with different structural parameters, the RMS error vector can be calculated. By using the training data set composed of the structural parameters and RMS error vector, a parametric evaluation neural network is trained. A beam structure is considered as an existing structure, based on the trained parametric evaluation neural network, the stiffness of the beam structure can be forecast. It is shown that the parametric identification strategy using macro-strain measurement from long-gage sensors has the potential of being a practical tool for a health monitoring methodology applied to civil engineering structures.
A neural networks based modeling of structural parametric identification strategy with the direct use of dynamic measurements in time domain is developed. Without any eigenvalue analysis, the structural parameter of stiffness and Rayleigh damping coefficients can be identified simultaneously. A shear frame structural model with known mass distribution is considered as an object structure. First a reference structure with assumed structural parameters is chosen. The assumed reference structure has the same degree of freedoms and topology with the object structure. An emulator neural network is constructed and trained to identify the reference structure by the use of dynamic measurements under dynamic excitation. The trained emulator neural network can be used to forecast dynamic measurements of the reference structure with enough precision and decide the difference between the dynamic measurements of other assumed structure with different structural parameters and those of the reference structure. The root mean square (RMS) error vector is presented to evaluate the difference. Subsequently, corresponding to each assumed structure with different structural parameters, the RMS error vector can be calculated. By using the training data sets composed of the structural parameters and the corresponding RMS error vector, a parametric evaluation neural network is trained for the purpose of forecasting the structural stiffness and the Rayleigh damping coefficients.
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.