Showing papers on "Structural health monitoring published in 1996"

Computation of structural flexibility for bridge health monitoring using ambient modal data

Abstract: Issues surrounding the use of ambient vibration modes for the location of structural damage via dynamically measured flexibility are examined. Several methods for obtaining the required mass- normalized dynamic mode shapes from ambient modal data are implemented and compared. The method are applied to data from a series of ambient modal tests on an actual highway bridge. Results indicate that for the damage case examined, the flexibility from the ambient mode shapes gave a better indication of damage than the flexibility from the forced-vibration mode shapes. This improved performance is attributed to the higher excitation load levels that occur during the ambient test.

High-frequency impedance analysis for NDE of complex precision parts

Abstract: A high-frequency, impedance-based structural health monitoring technique developed at the Center for Intelligent Material Systems and Structures (CIMSS) has been extended to the NDE of complex precision parts. Like many other non-destructive evaluation (NDE) techniques, this method relies on the detection of change in the dynamic properties of the structure when damage occurs. This NDE technique relies on the measurement of the electrical impedance at high frequencies (100 kHz - 275 kHz) with bonded non-intrusive piezoelectric actuator/sensors. Since the electrical and mechanical impedances of the bonded actuator/sensor are directly coupled, the proposed NDE method is able to detect incipient damage in the structure. Gears were chosen as complex precision parts for the experimental procedure because of their tight tolerances, high quality, and broad use. The goal is to show that incipient damage in the gear teeth, which are an extension of the cylindrical base structure, can be monitored. The most common types of damage in gears, i.e., abrasive wear and bending fatigue, were successfully detected. The impedance measurements before and after damage were converted into a scalar damage metric, which was used to detect the presence of damage when a threshold value was exceeded. Also, quality inspection was successfully demonstrated using the impedance-based technique.© (1996) COPYRIGHT SPIE--The International Society for Optical Engineering. Downloading of the abstract is permitted for personal use only.

Application of active structural health monitoring technique to aircraft fuselage structures

Abstract: Damage identification of complex structures can be performed using an Active Structural Health Monitoring System (ASHMS) utilizing an array of piezoceramic (PZT) sensor- actuators and an electromechanical impedance analyzer. Based on the theory of electromechanical impedance for surface-bonded collocated sensor-actuators, the system provides the means of implementing effective Non-Destructive-Evaluation health monitoring to a structure at any point in its life cycle. When integrated into a structure, the ASHMS can identify the location and extent of damage through a statistical analysis algorithm which compares the electromechanical admittance of the structure's current condition with the structure's `baseline' condition over a defined frequency range. This paper presents a model independent method of structure damage identification based on statistical analysis of the changes in the electromechanical impedance of the structural response. Using a small section of an airplane fuselage with free-free boundary conditions as the test structure, a prototype system has been successfully developed which automatically measures and collects the electromechanical admittance data from an array of transducers. The measurement system developed can apply large voltages to the PZT transducers, while accurately determining the spectral and harmonic information of the frequency range of interest resulting in an increase in the measurement sensitivity and the area monitored by a single transducer. The system can accurately measure and collect large amounts of structural information with a limited number of transducers in a very short period of time. The post-analysis of the measured structural variation can be performed with the advanced signal processing algorithm presented herein within seconds of data collection. The results show a remarkable accuracy of damage location identification for a complex structure. The basic research conducted so far indicates that the ASHMS and associated processing techniques have a promising potential for health monitoring of real complex structures.© (1996) COPYRIGHT SPIE--The International Society for Optical Engineering. Downloading of the abstract is permitted for personal use only.

Development of structural health monitoring techniques using dynamics testing

Abstract: Today`s society depends upon many structures (such as aircraft, bridges, wind turbines, offshore platforms, buildings, and nuclear weapons) which are nearing the end of their design lifetime. Since these structures cannot be economically replaced, techniques for structural health monitoring must be developed and implemented. Modal and structural dynamics measurements hold promise for the global non-destructive inspection of a variety of structures since surface measurements of a vibrating structure can provide information about the health of the internal members without costly (or impossible) dismantling of the structure. In order to develop structural health monitoring for application to operational structures, developments in four areas have been undertaken within this project: operational evaluation, diagnostic measurements, information condensation, and damage identification. The developments in each of these four aspects of structural health monitoring have been exercised on a broad range of experimental data. This experimental data has been extracted from structures from several application areas which include aging aircraft, wind energy, aging bridges, offshore structures, structural supports, and mechanical parts. As a result of these advances, Sandia National Laboratories is in a position to perform further advanced development, operational implementation, and technical consulting for a broad class of the nation`s aging infrastructure problems.

Rotor acoustic monitoring system (RAMS): a fatigue crack detection system

Abstract: The Rotor Acoustic Monitoring System (RAMS) is an embedded structural health monitoring system to demonstrate the ability to detect rotor head fatigue cracks and provide early warning of propagating fatigue cracks in rotor components of Navy helicopters. The concept definition effort was performed to assess the feasibility of detecting rotor head fatigue cracks using bulk- wave wide-bandwidth acoustic emission technology. A wireless piezo-based transducer system is being designed to capture rotor fatigue data in real time and perform acoustic emission (AE) event detection, feature extraction, and classification. A flight test effort will be performed to characterize rotor acoustic background noise and flight environment characteristics. The long- term payoff of the RAMS technology includes structural integrity verification and leak detection for large industrial tanks, and nuclear plant cooling towers could be performed using the RAMS AE technology. A summary of the RAMS concept, bench-level AE fatigue testing, and results are presented.© (1996) COPYRIGHT SPIE--The International Society for Optical Engineering. Downloading of the abstract is permitted for personal use only.

Health Monitoring Studies on Composite Structures for Aerospace Applications

Abstract: This paper discusses ongoing work to develop structural health monitoring techniques for composite aerospace structures such as aircraft control surfaces, fuselage sections or repairs, and reusable launch vehicle fuel tanks. The overall project is divided into four tasks: Operational evaluation, diagnostic measurements, information condensation, and damage detection. Five composite plates were constructed to study delaminations, disbonds, and fluid retention issues as the initial step in creating an operational system. These four square feet plates were graphite-epoxy with nomex honeycomb cores. The diagnostic measurements are composed of modal tests with a scanning laser vibrometer at over 500 scan points per plate covering the frequency range up to 2000 Hz. This data has been reduced into experimental dynamics matrices using a generic, software package developed at the University of Colorado at Boulder. The continuing effort will entail performing a series of damage identification studies to detect, localize, and determine the extent of the damage. This work is providing understanding and algorithm development for a global NDE technique for composite aerospace structures.

Theoretical modeling of wave localization due to material damping

Abstract: A structural health monitoring technique developed at the Center for Intelligent Material Systems and Structures (CIMSS) relies on the high frequency impedance characteristics of the structure to qualitatively detect incipient damage. An important feature of this technique is the limited PZT actuator-sensor sensing area, which has been attributed to energy dissipation through material damping and nonconservative joints. In this paper, a study of the structural damping effect at high frequency on the localized sensing region is presented. A theoretical model of the energy dissipation, using a wave propagation approach and the correspondence principle is derived to obtain the specific damping capacity. For this purpose, an infinitely long bolted beam has been simultaneously analyzed using numerical and experimental procedures to obtain a quantification of the energy dissipated through the structure. The results of this study are presented in part II of this paper.© (1996) COPYRIGHT SPIE--The International Society for Optical Engineering. Downloading of the abstract is permitted for personal use only.

Identification of broadband acoustic emission signals

Abstract: Aircraft and rotorcraft health monitoring systems (HMS) must be capable of identifying structural damage(cracks) when the damage is very small. Typically, small is small enough that classical finite element modeling methods are inadequate. Monitoring the structure at very high frequencies and using a combination of relativistic,and absolute measures of the monitored signals can be used to provide a measure of the structures health. Passively monitoring the structure for waveforms, generated by crack growth is a promising method for incorporation intoa health monitoring system. This paper will discuss research into the implementation of such a system. The issue of signal de-noising using adaptive filters prior to waveform identification will be addressed, and results presented. Keywords: structural health monitoring, smart structures, crack detection, acoustic emission, adaptive filters. 1 INTRODUCTION There are fleets of aging aircraft and helicopters that are being operated well beyond their intended service lives.At present there is a need for structural health monitoring systems that can autonomously monitor particular

Damage detection of a damped structure with neural networks

Abstract: In the future, is envisioned that passively damped structures and structural health monitoring techniques will be applied to the same class of structures. However, damage detection research has not been performed on damped structures because of the mathematical complexities associated with highly damped systems. A neural network is presented for detecting damage of a structural system damped with viscoelastic material dampers. The damage characteristic (location and severity) are predicted using driving point impedance functions as the inputs to a four layer backpropogation network. While the neural network adequately predicts damage on a simple structural system, it may become unreasonably large for structural systems with a large number of degrees of freedom.

Real-time structural health monitoring system with fiber optic sensors

Abstract: A new technique of the real-time structural health monitoring system with fiber-optic sensors is reported. The fibers are etched equidistantly to fabricate the sensing ranges and are embedded in composite material perpendicularly to build the 2D fiber-optic sensor array. This paper discusses the methods of the etching fiber and the embedding fiber. The output light signals of the fibers that sense the strain distribution of composite material are processed by a computer and shown on the screen. The detecting results are given.© (1996) COPYRIGHT SPIE--The International Society for Optical Engineering. Downloading of the abstract is permitted for personal use only.

Dynamic learning rate neural network training and delamination detection

Abstract: In this paper, a new learning procedure, called dynamic learning rate steepest descent (DSD) method, is proposed for training neural networks. Based on a simple steepest descendent (SSD) method, the proposed method improves the learning convergence speed significantly without sacrificing the computational effort, memory cost, algorithm simplicity, and computational locality in the standard layered error back propagating training algorithm. Through numerical experiments, the current method is shown to have much stronger learning ability than that of standard constant learning rate steepest descent method (SSD) and accelerated steepest descendent method (ASD). The numerical experiments also indicate that the current method is robust to the initial learning rate, which is critical in the standard steepest descent method. It is also proven to be efficient. The CPU time increases, due to the extra operations in the DSD algorithm, is negligible. The DSD method is then used to train a neural network for direct identification of composite structure delamination through structural dynamic responses. The result indicates that neural network can be used for real-time flaw detections and advanced structural health monitoring. * Graduate Research Assistant 1 Professor, Aerospace Engineering, Member AIAA Copyright © 1996 American Institute of Aeronautics and Astronautics, Inc. All rights reserved. INTRODUCTION Damage detection based on structural dynamic response and system identification techniques has been a very active research subject in the past several years. Traditionally in the detection methods, an intact system model is first formulated numerically or experimentally as a reference system. The in-service response of the system is then detected and compared with the response of the intact model. The discrepancies are used to infer the health condition of the system. Since the damage information inference is usually a complicated inverse process and the system identification technique is usually time-consuming, the conventional damage detection methods can not be applied to real-time system health monitoring. The neural network technique offers a potential for real-time health monitoring. Instead of inferring the damage information through the reconstruction of the system on the basis of the dynamic response, the neural network technique is able to map the dynamic response to the damage information directly. A very important task in the neural network technique is to train the network so that the network can perform the mapping correctly. Error backpropagation algorithm,"'' which is based on the steepest gradient descent method, has

Statistical system identification and applications to seismic response of structures

Abstract: A pragmatic and versatile statistical system identification framework is presented and applied to seismic response records of structures. The framework is based on the interpretation of probability as a measure of plausibility and on Bayesian statistical inference. Various classical system identification techniques can be derived and viewed as the special cases of the framework. However, the framework can provide a more informative interpretation of the identified optimal model. When the number of sampled input and output data from structures is large, useful asymptotic approximations of the analytical results are available. These asymptotic approximations are incorporated into the framework by introducing the definitions of system identifiability and model identifiability. New asymptotic approximation results are derived for the system un-identifiable case. From the viewpoint of asymptotic approximations, the system identification problem is a non-trivial global optimization problem. Two generalized trajectory methods, the homotopy scheme and the relaxation scheme, are presented which can be combined to provide a very robust numerical procedure for global optimization. Both methods can also be applied to find the roots of a set of nonlinear algebraic equations. Structural model updating is useful because it can be applied to structural health monitoring and is also desirable since the theoretically based stiffness matrix of a structure can be improved by using the measured structural response data. However, no well-accepted solution to this difficult problem has emerged primarily because it is an ill-conditioned and non-unique inverse problem. A single-stage structural model updating approach using the least-squares prediction-error system identification method and a substructuring technique is proposed and applied to simulated and real structural response data.

Dynamic Pretest on Ground of Steel Girder Section Units of Tsing Ma Bridge

Abstract: Publisher Summary This chapter discusses the dynamic test of a group of steel girder section units of Tsing Ma Bridge. The test is one of pretests for the research on the Wind and Structural Health Monitoring System (WAHMS) of Tsing Ma Bridge under the Highway Department of Hong Kong. In consideration of very short time period before open to traffic and after its completion, and the importance of local vibration modes of the steel girder of Tsing Ma Bridge to detect its damage, four girder section units are selected and tested on the ground. The test includes ambient test and impulse response tests on each unit. The test identifies up to 18 vibration modes, including the wrapping and distortion of cross section of closed profiles of the girder and local modes of each section unit that are mostly coupled modes of the units with their support pile foundations. Nonlinearity in the vibration responses of the combined systems of a unit and a pile foundation is identified.