Condition Monitoring of Materials and Structures

Many critical reinforced concrete structures and components undergo progressive damage because of the nature and magnitude of service loads, environmental effects, deterioration of material properties, and other factors. Depending on their locations, they may also experience major additional damage from natural disaster events such as earthquakes, hurricanes, high winds, and tornadoes. Continuous damage and structural assessment of these structures provides a mean for their possible timely repair and retrofitting for continuous use and operations. Research on brittle materials indicates that the microfracturing process (acoustic emission) is fractal in nature in both the temporal and spatial domains. The associated characteristics may serve as good indicators for continuous damage assessment and catastrophic failure prediction.

Recent advances in engineered wood composites are allowing us to further push the load limits of timber used in structural applications. In an effort to better understand failure processes we are applying acoustic emission techniques to examine progressive crack growth along the grain in clear wood specimens loaded for mode I fracture. We have developed a technique to estimate the energy released by an acoustic event. By applying basic fracture theory we can determine the bulk fracture energy of the specimen from load versus crack length data. A comparison of acoustic energy measurements with bulk fracture energy measurements gives us an idea of how much fracture energy is converted to acoustic energy. Results from the experiments presented here indicate roughly one sixth of the fracture energy is measured as acoustic emission energy. This comparison will allow us to apply the technique to laboratory tests of damage growth in larger structures.

Development of a fiber optic sensor for detection of acoustic emissions in civil structures is described. The main advantage of the sensor is its multi-point signal detection capability eliminating the need for intricate triangulation procedures normally associated in locating the source of acoustic emission (AE) signals. The fiber optic sensor is interferometric based, and a filtering system eliminates the noise associated with signals originating from sources other than the AE events. Preliminary experiments with concrete and steel specimens indicate that a single fiber optic sensor is capable of locating the source of the AE event.

Many fiber optic sensors have been proposed for both static and dynamic measurements. Recently the authors propose a double-side fiber optic transducer. Dynamic tests were carried out with the sensor fixed at the top of a steel bar mounted on a shaking table. The optical device supplied an output signal directly comparable with the signal given by the accelerometer placed on the top of the bar. This paper illustrates a new series ot tests conducted by arranging two transduction plates with a single V-groove on the two opposite faces of a steel column of a single degree of freedom structural system. The acquired data were compared with the responses of a capacitive accelerometer located on the structure girder and of strain gages glued on the steel columns.

The ability of an embedded optical fiber sensor to monitor strain distribution in a structural material depends on the bonding characteristic between the host material and optical fiber. A mathematical model was established in order to compute the effect of fiber optic coating characteristics and the gauge length on the strain transfer mechanism. The validity of the model was verified through the analysis of a finite element model representing the interface between the fiber core, coating and the host material. A strain modification factor is introduced in order to account for the shear lag and convert the fiber optic measured strains to the actual strains in the structure. The theoretical findings were verified through a series of experiments involving white light interferometry. The experimental investigation comprised of testing concrete cylinders in direct tension. The cylinders were instrumented with embedded fiber optic sensors as well as conventional extensometers for comparison. The strain modification factor was employed for the interpretation of strain in concrete from the fiber optic sensed data.

Improved performance of concrete structures requires better material processing techniques and early detection of defects. Poor service performance is often a result of improper material processing. Early development of damage in concrete structures is also related to insufficient strength development prior to loading the structure. An NDE technique that monitors the setting and hardening process of Portland cement concrete is presented in this paper. The technique is based on measuring the ultrasonic wave reflection factor (WRF) between the hardening concrete and a steel interface. This method can be used to assess the in situ properties of concrete at very early ages so that corrective measures can be taken. In addition, a nondestructive technique for determining the dynamic elastic material properties of concrete based on the vibrational resonance frequencies of finite length cylinders is presented. The accuracy of the proposed technique is verified through independent experiments on beam specimens.

A novel concrete specimen was designed to facilitate the simultaneous action of mechanical loading and chloride penetration through concrete. The chloride penetration is evaluated by concentration profile in concrete specimens. The damages of concrete due to mechanical loading were characterized in two aspects: the applied loading level and loading repetition and the internal damage measured by travel time of ultrasonic signals. The test results of concentration profiles showed that with increasing loading level from zero up to 70% of compressive strength of concrete, both chloride concentration and penetration depth increased. The most drastic increase occurred in the loading range above 60% of compressive strength. The test results from ultrasonic tests are consistent with the analysis on concentration profile. This means that cyclic loading (low cycles) accelerates chloride penetration through concrete when the loading level is above 60% of the concrete strength.

This paper describes an empirical study of the effects of high amplitude (stress ratio) cyclic flexural loading on chloride permeability in concrete. Third point cyclic loading is applied to notched concrete beams at 1Hz, using a closed-loop control system based on deformation at the notch mouth (CMOD). The work addresses two types of fatigue: conventional constant load per cycle and constant deformation per cycle. Both compliance (CMOD/load) and chloride permeability (K_cl) using the "Rapid" test (ASTM C1202-94) are monitored with progressive deformation. The measured K_cl is a composite measurement through the 3.75" diameter, 2" thick cores recovered from beam sections bracketing the distressed area. Comparisons are made between the distressed K_cl and that measured on cores of "intact" concrete obtained at the non-loaded sections of the same beams. A relationship between chloride permeability and compliance (inverse of stiffness) is indicated, showing that fatigue cracking affects both mechanical and diffusive behavior.

This paper deals with the development, production and testing of different types of bridge decks utilizing FRP composites. The bridge decks discussed herein (concrete deck with FRP rebars; all FRP composite Hardcore Deck, H-Deck, Kansas Deck, MMM Deck and Cellular Deck) incorporate several different FRP composite bridge deck designs developed in the United States. A bridge utilizing FRP composite rebars in concrete deck is discussed first; followed by discussion of all FRP composite bridge decks. The information for each bridge system is summarized from referenced publications. Laboratory and field tests indicate that the bridge decks utilizing FRP composites are performing well. All FRP composite bridge decks are about five times lighter than concrete bridge decks. Other features of all FRP composite bridge decks are that they: (1) are well suited for modularization and mass production; (2) posess good energy absorbing capacity; (3) have enhanced fatigue and corossion resistant properties; and (4) require less erection time in the field since they are light-weight and use light equipment. At present, the higher initial cost of bridge decks utilizing FRP composites is an obstacle for their acceptance by bridge engineering community. The cost of bridge decks utilizing FRP composites will decrease as the demand increases, resulting in higher volume of production.

Development and testing of a bond-type anchorage system for specific use with Carbon Fiber Reinforced Polymer composite (CFRP) tendons have been described. The anchorage system employs expansive cement as the gripping agent. This was done in order to eliminate deformations associated with creep and thermal effects ordinarily associated with epoxy resin type binders in anchorage systems. Two different diameters (8 and 10 mm) of smooth CFRP rods were employed to examine the effectiveness of the anchorage system in post tensioning applications. The experimental program was designed in order to determine several important parameters including the expansive pressure, bond strength, and distribution of strain within the anchorage zone.

This paper presents the results of an investigation into the applicability of ultrasonic techniques for obtaining quantitative measures of changes in effective properties of glass-fiber reinforced polyester composite structural members manufactured by the pultrusion process. Immersion techniques were used to measure the through-thickness ultrasonic velocities, as a function of input frequency, in pristine specimens, as well as after the same specimens had been damaged; damage was induced by three-point bending. Specimens were evaluated in an unloaded condition, and distinct changes were observed between the measured velocities of the pristine sample and those after the introduction of damage. The measured changes were found to be greater than the experimental error associated with the test procedure.

Within the past few years a number of structures have had various types of sensors either embedded into them or surface attached onto them. Regardless of the sensor type (electrical, chemical, fiber optic, etc) the measurement has, almost exclusively, entailed somehow connecting a data acquisition system to the sensor, or network of sensors. Usually this has meant developing some reasonable method of running wires or optical fibers throughout the structure. The developments in secure, inexpensive wireless communication systems provides a method of reducing, if not eliminating, the cabling burden for instrumenting a large civil structure. Wireless communication schemes applicable to this task, represented by components meeting standards 802.11 and P1451.x, as well as Bluetooth, may be configured in such a way as to provide a seamless path for acquiring the sensors' readings and (potentially) even placing them on an automated Website. These standards, as well as realistic applications of interrogating embedded sensors, are presented.

A major contributor of the problems encountered by bridges is the deterioration of the riding surface. It would be very beneficial to locate problem areas under the surface, such as delaminations, before the riding surface spalls out. In this study, we use the conventional chain drag method and compare it with the ground penetrating radar (GPR) in determining subsurface characteristics of concrete decks. The chain drag method was applied on laboratory specimens to determine different frequency characteristics under controlled conditions such as specimens with air filled cracks, with water filled cracks and with steel reinforcement. A GPR system from the West Virginia Department of Transportation (WVDOT) was used to scan three bridge decks. The characteristics of signal patterns from these three different bridge decks were studied to determine concrete bridge deck thickness, and where steel reinforcement and delamination exist. The findings from the GPR analysis were compared with that from the bridge deck inspection reports generated by the WVDOT using the chain drag method.

Modal frequencies derived from experimental motion measurements have been shown to be useful measures for indicating the presence of perturbations in structures. The sensitivities of modal frequencies to different types of structural change can be quantified through the use of finite element models, and the comparison of experimental modal frequencies with FEM models is now an accepted practice for the initial analysis of experimental motion data. However, most structures are directly exposed to the environment and are thus subjected to variations in temperature, humidity, insolation as well as other influences which may mimick actual structural change. The extent literature appears to indicate that the sensitivities of the mode frequencies to temperature are of the same order as those predicted to occur for small but significant structural modifications. In view of this observation, there is a great need for further data which can be used to develop statistically reliable temperature corrections and to establish long-term repeatability of the corrections. To this end, this paper presents preliminary correlations of the modal frequencies with seasonal temperature variations during a seven month period of observation of a large-span pre-stressed segmental concrete bridge, during a time when outside air temperatures ranged from —15 (C) to 33 (C), which may be useful for this purpose.

Optical instrumentation for structural testing is convenient and less expensive for installation and data acquisition. This paper presents a method using optical devices for structure health monitoring by global diagnosis. The method uses a probabilistic approach to diagnosis-decision making. It is experimented in the laboratory respectively using a coherent laser radar system and a CCD high-resolution camera. Results show that this method is promising for field application.

This paper discusses implementation and modeling details for a simple semi- active control system which uses an electro-rheological (ER) damping wall as an actuator. The ER damping wall follows the principle of viscous damping walls. The control system requires measurements that are available directly from the ER damper itself: namely a relative velocity and an absolute acceleration. An absolute velocity is obtained from the accelerometer via an analog integrator which was designed for this application. The integrator is made from a cascaded low-pass filter and a differentiator, is stable at zero frequency, and has a perfect ninety degree phase lag at the first resonant frequency of the structure. The control rule is implemented with four op-amps, five transistors, and a high-voltage switch. Modeling details for this system are emphasized in the second part of the paper. The ER damping wall is a highly non-linear device, however its behavior can be described via a state-variable model that has only one non-linear term. The structure to be controlled exhibits linear behavior, although linear behavior is not presumed in the control rule synthesis. Natural frequencies, damping ratios, and mode shapes are obtained from a frequency-domain curve-fit. A technique to determine the damping and stiffness matrices from the modal data requires that the mass matrix is known a-priori. This method is used to create a state-variable model in physical coordinates. By comparing simulations of the numerical model to data from the closed-loop control experiment, aspects of the ER device behavior most salient to the performance of the closed-loop system are identified.