Recent surveys of plant engineers show a growing trend toward the utilization of digital sensors combined with greatly increased pressure for ease of use and ease of interfacing. With the current state of industrial fieldbuses these are somewhat contradictory objectives. Most fieldbuses were developed with the intention of replacing field wiring and thus focused on the physical and data link layers. When attention was turned to the application layer little emphasis was paid to the already well-developed world of distributed object computing. It is our belief that substantial progress could be made by adopting the common object request broker architecture as the basis for the application layer in all fieldbus network implementations. This has the considerable advantage that it shifts the focus from message packet structure to abstract description of sensor interfaces that are independent of the language and means of implementation.

Not Available

Optical fiber sensing techniques are ideal for applications where high-temperature, electromagnetic interference, or vibration cause traditional electrical sensors to become unreliable. Gold-coated, silica-based optical fibers can withstand temperatures up to 900 degrees C and sapphire fibers can be employed for temperatures as high as 2000 degrees C. We present dynamic strain and temperature measurements of ceramic matrix composite specimens using extrinsic Fabry-Perot interferometric fiber optic strain and temperature sensors. The extremely low-mass and rugged construction of the sensors will allow them to survive high- cycle, high-temperature fatigue testing.

Lamb waves are easily generated and detected using laser techniques. It has been shown that both symmetric and antisymmetric modes can be produced, using single-spot and phased array generation. Detection has been demonstrated with Michelson interferometers, but these instruments can not function effectively on rough surfaces. By contrast, the confocal Fabry-Perot interferometer can interrogate rough surfaces, but generally is not practical for operation below 300 kHz. In this paper we will present Lamb wave data on a number of parts using a robust, adaptive receiver based on photo-emf detection. This receiver has useful sensitivity down to at least 100 kHz, can process speckled beams and can be easily configured to measure both out-of-plane and in- plane motion with a single probe beam.

This paper discusses the principles of surface second harmonic generation (SurfS) technique and its application to the remote in-situ detection of low levels of surface corrosion and contamination. SurfS is a nonlinear-optical technique which is surface sensitive. As an optical technique, SurfS is nonintrusive, noninvasive and nondestructive. The SurfS technique has spectroscopic capability. The formation of metal-oxide or other species due to corrosion or contamination will lead to a change in the spectral signature. This provides a means to monitor the onset of corrosion or deposition of contamination. Since the SurfS technique is surface specific, it can examine any optically accessible interface. During corrosion or contamination-induced chemical changes at an interface, the chemical changes at the interface can be monitored as they occur. In this way, the technique can be used in an in-situ fashion.

Optical fiber corrosion sensors are being developed to address the high service costs associated with current structural maintenance procedures for civilian and military assets. A distributed optical fiber sensor system will help reduce the costs associated with corrosion damage and extend the lifetime of existing assets. Annual national losses in time, labor, materials and systems has been estimated in the billions of dollars. Additional costs arise from system downtime that results from disassembly procedures necessary to locate corrosion damage in remote locations. Furthermore, the potential to damage other system parts during maintenance is increased when disassembly and reassembly occurs. The development of on-line optical fiber sensors capable of detecting corrosion would eliminate a significant portion of the maintenance costs. We present recent test results using optical fiber long-period grating (LPG) corrosion sensors. With the appropriate coating, the sensors can be designed to detect water or metal ions in otherwise inaccessible regions of the aircraft. The LPG sensors can be designed with low temperature cross-sensitivity, multiplexed along a single fiber, and can be demodulated using a simple, low-cost spectrum analyzer.

The measurement and control of cleanliness for critical surfaces during manufacturing and in service operations provides unique challenges in aerospace. For re-usable propulsion systems, such as the solid rocket motors, the current thrust for environmentally benign processes has had a major impact on programs designed for maintaining quality in the production of bondline surfaces. The major goal is to improve upon our ability to detect and identify possible contaminants which are detrimental to the integrity of the bondline. This effort requires an in-depth study of the possible sources of contamination, methodologies to detect and identify contaminants, discriminate between contaminants and chemical species caused by environmental conditions, and the effect of particular contaminants on the bondline integrity of the critical surfaces. This presentation will provide an overview of several optical methods used to detect and identify contamination on critical surfaces, currently being performed by the Surface Contamination and Analysis Team at Marshall Space Flight Center. The methods under development for contamination monitoring include FTIR and Near-IR SPectrometry, UV Fluorescence, and Variable Angle Spectroscopic Ellipsometry. Comparisons between these methods and the current primary tool, optical stimulation of electron emission for on-line inspection will be presented. Experiments include quantitative measurement of silicone and Conoco HD2 greases, metal hydroxides, tape residues, etc. on solid rocket motor surfaces.

The most established way to inspect surface roughness of machined surfaces is to measure with profilometers. These techniques are not applicable for on-line inspection. A few decades ago, the correlation between laser speckles of different light wavelengths and surface roughness was discovered. Since then several authors have been working on this topic. In measuring systems which are considered in this paper, the surface which is to be inspected is illuminated by a polychromatic laser beam. The scattered light is converted by an optical system, which ensures that the product of wavelength times focal length is constant. In the Fourier plane of each subsystem a CCD-array is installed. The CCD- data are captured by a frame grabber and stored for evaluation in a computer. One major problem in industrial processes where surface roughness measurements had ben tried was that the inspected metal workpieces may be tilted and, therefore, the direction of spectral reflection changes. In the method discussed in this paper an approximate value of the surface roughness can be obtained by determining the difference between the zero orders of the scattering patterns of the wavelengths. This difference can be approximated by the position of the maximum of the 2D- cross correlation function of related speckle patterns. The main benefit of the method described is the feasibility of measuring roughness during machining or other types of continuous or semi-continuous production processes. This is achieved by a synchronous detection of speckle patterns of the different wavelengths used. In the pilot project the shutter sped of the cameras is 1/10000 seconds, therefore, in this case only frequencies above 1kHz disturb the measuring results.

For the shape measurement of the cold rolled strips, the new rectangular fringe shift method is introduced. The image saturation always occurs in the image center because of the total specular reflection of the strip surface. The central area saturation is reduced by using the circular gaussian density filter. Least-square fitting is used to find out the phase information of the strip. Phase map is obtained from the eight quasi-rectangular strip patterns and the shape of the object surface can be expressed by unwrapped phase map of 256 gray levels.

This paper presents a novel non-destruction technique for stiffness measurement and defect detection in laminated composite plates using dry-coupled ultrasonic plate waves. The technique involves using a pari of wide-band transducers normally-placed on the plate surface to generate and receive the lowest order flexural mode (A₀ mode) at low frequencies. The stiffness measurement first experimentally determines the phase velocity of the A₀ mode, then reconstructs the stiffness matrix components D₁₁, D₂₂, A₄₄, and A₅₅ utilizing the higher order plate wave solution for laminated composites and a nonlinear least-square optimization method. Such stiffness measurements can be used for quality control/quality assurance of in-service composite structures or composite manufacturing process, and for verification of composite design. As well as the stiffness measurements, the time-of- flight and amplitude measurements are achieved by keeping the pulser-receiver distance and pulsing frequency constant, and scanning the transducers over the plate surface. Since the plate wave measurements are based on the direct first arrival of the flexural mode, they are not limited by the thinness of the specimen as the conventional c-scan. Examples of stiffness measurement and defect detection are given, including cases of impact damage, delamination and debonding in composite structures. Finally the main hardware and software features of a recently-developed F-Scan system are also described.

A sensor for detecting poor interlaminar bonding between a topmost ply and composite substrate was developed and validated. The sensor was designed for the specific application of real time process control for in situ consolidation processes, in which the composite structure is local heat and pressure are applied by a moving head as each new composite tow is laid down upon the existing composite substrate. The sensor is designed to follow the consolidation head and provide real time information on the quality of bonding between the topmost play and substrate. The sensor design consists of two piezoelectric transducers oriented in a pitch catch arrangement. The transmitting transducer is angled so as to produce surface waves in the composite panel. These waves interact with defects at the top ply/substrate interface. The received signal is analyzed for wave speed, attenuation and energy dissipation. Experimental validation of he sensor is presented for thermosetting matrix composites. Several composite panels were made under various manufacturing conditions so as to create poor interlaminar bonding at the topmost ply. Interlaminar bonding was then evaluated by destructive and nondestructive measures. There was good agreement between the destructive and nondestructive measures of top ply bond strength.

When an imperfectly conductive structure is illuminated by an electromagnetic wave, its response, on the magnetic component, is comparable to a first order integrator, if the skin effect is negligible at the wave frequency. Using this model, we have developed a method and a magnetic prove allowing to build 'electric images' of the defects present in carbon epoxy composite materials, by measurement of magnetic field transmission in these materials. After a presentation of the theory of electromagnetic scattering in a non perfectly conductive materials from M.A. Bethe and K.F. Casey, we demonstrate that it is not necessary to perform the magnetic field attenuation measurement by transmission; we can proceed on one face of the sample, by reflection, obtaining the same result. We present the experimental set up and the magnetic probe developed, we show several images achieved on samples having induced defects and we compare these image with the images achieved by a classical ultrasonic method on the same specimens. Furthermore we present images produced by integration in a structure of a 'sensor-network' based on the principle used by the probe.

A photorefractive optical lock-in is discussed in relation to ultrasonic vibration modal analysis of inertial confinement fusion (ICF) targets. In this preliminary report, the method is used to analyze specimens with similar response characteristics to ICF targets with emphasis on both the displacement and frequency resolution of the technique. The experimental method, based on photorefractive frequency domain processing, utilizes a synchronous detection approach to measure phase variations in light scattered from optically rough, continuously vibrating surfaces with very high, linear sensitivity. In this photorefractive four-wave mixing technique, a small, point image of the object surface is made to interfere with a uniform, frequency modulated reference beam inside a Bismith Silicon Oxide crystal. Optical interference and the photorefractive effect of electronic charge redistribution leads to the formation of a refractive index grating in the medium that responds to the modulated beams at a frequency equal to the difference between the signal and reference frequencies. By retro-reflecting the reference beam back into the crystal, a diffracted beam, counter-propagating with respect to the original transmitted beam, is generated. Using a beamsplitter, the counter-propagating beam can be picked-off and deflected toward a photodetector. The intensity of this diffracted beam is shown to be a function of the first-order ordinary Bissel function, and therefore linearly dependent on the vibration displacement induced phase modulation depth (delta) , for small (delta) ((delta) < 4 (pi) (xi) /(lambda) < < 1) where (xi) is the vibration displacement and (lambda) is the source wavelength; analytical description and experimental verification of this linear response are given. The technique is applied to determine the modal characteristics of a rigidly clamped disc from 10 kHz to 100 kHz, a frequency range similar to that used to characterize ICF targets. The results demonstrate the unique capabilities of the photorefractive optical lock-in to detect and to measure vibration signals with very narrow bandwidth and high displacement sensitivity. This level of displacement sensitivity is particularly important in detecting changes in vibrational mode shapes and frequencies that might be associated with asymmetries in ICF targets.

Fiber optic grating sensors have been embedded into a 2.5 cm thick composite panel and used successfully to measure the manufacturing process, impact testing and bend tests until failure. This paper describes tests performed and includes a representative set of test results.

Residual stress is induced in fiber composite materials during the cure process because the thermal expansion coefficient of the fiber is generally much lower than that of the polymer matrix. The two materials are 'locked' together at the cure temperature. Then, as they cool, the matrix attempts to contract more than the fiber leading to tension in the matrix and compression in the fiber. This can lead to the formation of microcracks parallel to the fibers in thick composite piles or yarns. The magnitude of residual stress can be reduced by modifying the cure cycle; however, optimizing the cure cycle requires a complete understanding of the state of cure throughout the composite. This is a complex problem -- especially in thick composites. Pilot studies have been performed placing Bragg gratin sensors in glass fabric preforms and monitoring the response of the grating during resin infusion and cure. The typical response shows the initial thermal expansion of the Bragg grating, a rapid contraction of the grating as the resin gels, slower contraction during cure, and thermal contraction at the composite thermal expansion coefficient during cool down. This data is then sued with micromechanical models of the fiber/matrix interaction during cure to establish material parameters for cure simulation. Once verified, these cure simulation methods will be used to optimize tooling design and cure cycles in composite components.

On-line ultrasonic monitoring of polymer co-extrusion and gas-assisted injection molding are presented. During the co- extrusion of high density polyethylene and Santoprene ultrasonic sensors consisting of piezoelectric transducers and clad ultrasonic buffer rods are used to detect the interface between these two polymers and the stability of the extrusion. The same ultrasonic sensor also measures the surface temperature of the extruded polymer. The results indicate that temperature measurements using ultrasound have a faster response time than those obtained by conventional thermocouple. In gas-assisted injection molding the polymer and gas flow front positions are monitored simultaneously. This information may be used to control the plunger movement.

Composite process/product development relies on both process monitoring information and nondestructive evaluation measurements for determining application suitability. In the past these activities have been performed and analyzed independently. Our present approach is to present the process monitoring and NDE data together in a data fusion workstation. This methodology leads to final product acceptance based on a combined process monitoring and NDE criteria. The data fusion work station combines process parameter and NDE data in a single workspace enabling all the data to be used in the acceptance/rejection decision process. An example application is the induction welding process, a unique joining method for assembling primary composite structure, that offers significant cost and weight advantages over traditional fasted structure. The determination of the required time, temperature and pressure conditions used in the process to achieve a complete weld is being aided by the use of ultrasonic inspection techniques. Full waveform ultrasonic inspection data is employed to evaluate the quality of spar cap to skin fit, an essential element of the welding process, and is processed to find a parameter that can be used for weld acceptance. Certification of the completed weld incorporates the data fusion methodology.

The detection of longitudinally oriented defects in steel plate using ultrasonics has been widely reported. Ultrasonic methods are capable of detecting extremely small volume flaws in strip steel, but are limited because of the need to maintain fluid couplant between the transducer and steel strip. At a minimum, this couplant requirement slows the test speeds considerably, can introduce errors in test results, and, in many cases, prevents the test from being performed at all. The purpose of this paper is to present the results of the investigation of EMAT generated Lamb waves for the volumetric inspection of steel strip and subsequent on-line system performance. The strip steel industry has described a manufacturing problem of internal inclusions in their strip steel product for use in the automotive/appliance industry which is manifested after the rolling operation. The 'pencil pipe', a non-metallic inclusion introduced during the continuous casting process, is not detected prior to the roll, and after rolling it is too late to recover. A major midwestern US steel company considers this defect to be their number one quality problem. A method of detecting these inclusions prior to rolling was needed and is the basis of this development. The objective of this evaluation was the selection and implementation of EMAT generated Lamb wave modes that could be used for on-line detection of pencil pipe defects in strip steel before the strip is rolled to its final thickness. In addition, different Lamb waves modes were used to discriminate between the internal pencil pipe and non- deleterious surface scratches.

Laser-ultrasonics is a very attractive technique for in-line process control in the semiconductor industry as it is compatible with the clean room environment and offers the capability to inspect parts at high-temperature. We describe measurements of the velocity of laser-generated Lamb waves in silicon wafers as a function of temperature using fiber- optic laser delivery and all-fiber interferometric sensing. Fundamental anti-symmetric Lamb-wave modes were generated in 5 inches < 111 > silicon wafers using a Nd:YAG laser coupled to a large-core multimode fiber. Generation was also performed using an array of sources created with a diffraction grating. For detection a compact fiber-optic sensor was used which is well suited for industrial environments as it is compact, rugged, stable, and low-cost. The wafers were heated up to 1000 degrees C and the temperature correlated with ultrasonic velocity measurements.

An application of machine vision applied to the analysis of ultrasonic images formed using the time-of-flight diffraction (TOFD) method on incomplete weld geometries is described. The rationale of the work being to identify weld defects as soon as they are produced, thereby reducing the costs of any subsequent repairs. The analysis uses TOFD scans as input to a filtering and 'window' based variance operator for the segmentation of suspect defect areas inside the weld region. A suite of pc based software and a high temperature TOFD data acquisition system have been benchmarked through a series of demonstration trials on both 80mm thick carbon steel submerged arc welded testpieces, and 25mm thick carbon steel tungsten inert gas welded testpieces. The range of intentionally implanted defects, from root cracks to lack of side wall fusion, were detected with an overall accuracy of 79 percent on a data set of 174 defects on scans performed at 10-90 percent weld completion.

This paper describes an integrated nondestructive inspection framework that has been developed, under DARPA sponsorship, to meet the joint needs of optical design, high quality manufacturing and operational integrity monitoring that is both pragmatic and affordable. These needs stem form fiscal and regulatory drivers that mandate affordable hazard analysis and risk mitigation in both commercial and defense sectors. The framework features standardization of data formats for easy data exchange, automated data processing via the Internet, geographically remote data archiving and killer application interconnects, all tied in to user tailored conventional workstation interfaces. The effectiveness of this framework is demonstrated by an exemplar that is based on investment casing of gas turbine blades exploiting x-ray CT inspection, reverse CAD image processing and multimodal data fusion.

Film has been the primary x-ray acquisition medium for many years and is still widely used for NDE radiography. In recent years new technologies have been developed for x-ray image capture that offer the promise of improvement over film by taking advantage of digital systems. The availability of a direct digital detector, high speed computers, digital networking, and large capacity storage brings new capabilities to the radiographer in filmless radiography. These capabilities provide images that can be stored in mass digital storage systems offering longer storage life and faster retrieval, image processing software can be used to enhance images and provide computer assisted interpretation, and high quality digital images that can be transmitted from remote sites to interpretation centers. Amorphous silicon image sensors, developed by dpiX, A Xerox New Enterprise Company, offer an improved method of acquiring digital x-ray images. Amorphous silicon image sensor technology provides the opportunity to have large format size similar to x-ray film, high resolution, and a compact package for ease of use in NDE applications. This technology can also be used to replace x-ray examination of objects on a conveyor belt. This paper presents a description of amorphous silicon image sensor technology and a system developed to demonstrate the capabilities of this technology. This system called the FlashScan 20 has an 8 X 10 inch x-ray image acquisition area with a pixel matrix of 1536 X 1920 127 micron pixels.

The high penetrating powers of neutrons and of high energy gamma rays are utilized for the non-intrusive, non- destructive elemental characterization of materials. Neutrons produced in microsecond pulses interact with nuclei in the material, leading to the emission of gamma rays that have energies unique to each element. The pulsing of neutrons with a frequency of a few kHz allows the measurement of gamma rays resulting from a variety of neutron induced nuclear reactions. The elemental content of the material is deduced from the analysis of the gamma-ray spectra acquired during the neutron pulse and during the quiescent period between neutron pulses. A large number of elements from hydrogen to uranium can be identified and quantified within a few minutes of interrogation. This method called pulsed fast-thermal neutron analysis (PFTNA), can be used for on-line NDE, as well as for in situ characterization of materials. Devices based on PFTNA can be constructed in a stationary mode for on-line analysis, or in a transportable mode for in situ analysis. Examples on the utilization of the method include on-line coal analysis, on- line quality control in cement plans, measurement of oxidation of opaque metallic structures.

This paper illustrates the importance of residual stress characterization in welds and post weld processes. The failure to characterize residual stresses created during welding and/or post weld processes can lead to unexpected occurrences of stress corrosion cracking, distortion, fatigue cracking as well as instances of over design or over processing. The development of automated residual stress mapping and the availability of portable and fast equipment have now made the characterization of residual stresses using x-ray diffraction practical for process control and optimization. The paper presents examples where x-ray diffraction residual stress characterization techniques were applied on various kinds of welds including arc welds, TIG welds, resistance welds, laser welds and electron beam welds. The nondestructive nature of the x-ray diffraction technique has made the residual stress characterization of welds a useful tool for process optimization and failure analysis, particularly since components can be measured before and after welding and post welding processes. Some examples presented show the residual stresses before and after the application of post weld processes such as shot peening, grinding and heat treatment.

The control of fiber spacing is a difficult challenge in the manufacturing of composite materials. This paper describes an analytical approach coupled with a nondestructive evaluation method to analyze the effects of fiber spacing on the material properties of a composite material. Results of a finite element analyses are presented to quantify the effects of fiber spacing in unidirectional metal-matrix composites. Computed tomography (CT) data of unidirectional metal-matrix composite samples provide information on fiber locations for the analysis of the fiber distribution within the composite. Image processing methods are developed to extract fiber centers form the CT data. The processed CT data are used to produce a rectangular grid of finite elements which model the composite cross-section and where the stiffness matrix for each element is based on the local fiber volume fraction. The finite element results how that in some cases, stresses in the composite can be as high as 56 percent greater than the average stress and thereby set up stress concentrations which can initiate yielding and/or damage at loads well below those that would be calculated using average stress considerations only.

To provide real-time x-ray imaging, industry relies almost entirely on the combination of the x-ray image intensifier and the high-performance television camera. Although these devices have ben pushed to remarkable degrees of performance, they remain complex electro-optical assemblies with significant built-in errors, instabilities and degradation mechanisms. We describe a replacement for these system utilizing as a sensor a large array of amorphous silicon photodiodes and thin-film switching transistors. Specifically, the equipment described is a replacement for a 9-inch dual-mode x-ray image intensifier with a high performance 2000-line digital tv camera capable of operating in both real-time video radioscopic and high-performance radiographic modes.

This paper describes a magnetoelastic torque transducer nondestructive to the measured shaft. A sensing sleeve added to the measured shaft can not only measure the torque of the rotating shaft, but also be nondestructive to the shaft. Moreover, the grid areas are cut on the sleeve, which enables the common coil structure to replace the classic multipolar structure, so that the dynamic error reduces by a factor of 10. Our prototype has been sued to detect the torque of oil rig.

An application of machine vision applied to the analysis of radioscopic images of incomplete weld geometries is described. The rationale of the work is to identify weld defects as soon as they are produced, thereby reducing the costs of any subsequent repairs. Existing methods of weld and defect identification are compared, leading to the development of filtering and 'window' based variance operator for segmentation of suspect defect areas inside the weld region is described. The software and radioscopic imaging system have been benchmarked through a series of demonstration trials on both 80 mm thick carbon steel submerged arc welded testpieces, and 25mm thick carbon steel tungsten inert gas welded testpieces. The range of intentionally implanted defects, from root cracks to lack of side wall fusion, were detected with an overall accuracy of 87 percent, and classified in terms of defect size, shape, and position within the weld region.