Review of Progress in Quantitative Nondestructive Evaluation

Our keynote talk today has several objectives: First, we want to call attention, as some of us have before, to the interdisciplinary nature of NDE science and technology and some approaches for fostering R&D in such a situation. Next, we want to describe the objective of the DARPA, Air Force (and now Navy) core program for developing a science base for NDE and how it has evolved during the past two years. Some changes have indeed taken place; we feel that they were both necessary and evolutionary. Many of you are probably familiar with these changes by now but there may be some residual concerns or questions in your minds. Since I was the initiating influence behind most of them, it is appropriate for you to hear me say what they are and what they aren’t, and to have an opportunity to question us. Finally and most importantly, we want to enlist your participation in the difficult task of identifying exploratory development programs—and we will try to define this term—which will benefit from the growing science base that all of us are helping to develop. We hope to stimulate the interaction between people such as you and the consumers of the evolving NDE technology in order to identify reduction-to-practice possibilities that should be pursued.

A challenge has been presented to the scientific community to apply “first principles” to the understanding, prediction, and control of nondestructive evaluation (NDE) processes and applications. The success of the program is evident in the attention of new researchers and in the diversity of scientific specializations that have been directed to NDE problems. Critique of the program has been in awareness of existing technology and in focus of resources. An approach based on “lessons learned” is suggested for meeting continuing challenges and projected challenges of the future. “Lessons learned” from NDE reliability assessment programs are reviewed. Quantitative NDE performance as a function of signal-to-noise ratio is discussed. Focus of future efforts on signal-to-noise improvements in production and maintenance environments is proposed.

This paper presents results of Monte Carlo simulation of the Retirement-for-Cause (RFC) engine maintenance system as developed by Pratt and Whitney Aircraft and the U. S. Air Force. The Retirement-for-Cause concept is addressed, conventional Monte Carlo modeling techniques are explained, and an alternative approach developed at Pratt and Whitney is presented. Next, a simplified non-ideal Non-Destructive-Evaluation (NDE) model with fixed probabilities of Type I and Type II errors is described and simulation results obtained using this model are presented and discussed. An appendix presents a survey of various methods used to model NDE.

The damage tolerance approach to structural safety is based on the predicted growth of the “largest” crack that could be present in a structure at the start of a usage period1–4. The length, a_NDE of this potential crack has generally been determined by correlating crack size with the reliability of the non-destructive evaluation (NDE) system employed during structural inspections. Since many factors other than crack length influence detectability, NDE reliability as a function of crack length must be expressed as a probability of detection (POD) and must be estimated from a demonstration experiment whose results are non-deterministic.

Current nondestructive evaluation techniques generally do not produce identical indications when applied to flaws of the same length. The chance of detecting a given crack length depends on many factors, such as the location, orientation and shape of the flaw, materials, inspectors, inspection environments, etc. As a result, the probability of detection (POD) for all cracks of a given length has been used in the literature to define the capability of a particular NDE system in a given environment. Some POD curves are shown in Fig. 1 for various laboratory inspection techniques. Many other POD curves can be found, for instance, in Refs. 1–3.

Nondestructive detection of flaws in jet engine components is an essential part of assessing the integrity of the parts, and optimal design of experimental configurations is essential for high reliability detection. Due to the high cost of experimental evaluation, it would be highly desirable to accomplish this through the use of computer simulation of the detection/measurement process. This paper presents a first generation simulation model for the problem of ultrasonic detection of flaws in turbine rotor components. Included is the derivation of the model, the results of limited experimental validation of the software, and an example of its use to model the detection of penny shaped cracks in the plate-like geometry of a turbine rotor component web region.

A measurement model has been developed to describe ultrasonic measurements made with circular piston transducers in parts with flat or cylindrically curved surfaces. The model includes noise terms to describe electrical noise, scatterer noise and echo noise as well as effects of attenuation, diffraction and Fresnel loss. An experimental procedure for calibrating the noise terms of the model was developed. Experimental measurements were made on a set of known flaws located beneath a cylindrically curved surface. The model was verified by using it to correct the experimental measurements to obtain the absolute scattering amplitude of the flaws. For longitudinal wave propagation within the part, the derived scattering amplitudes were consistent with predictions at internal angles of less than 30°. At larger angles, focusing and aberrations caused a lack of agreement; the model needs further refinement in this case. For shear waves, it was found that the frequency for optimum flaw detection in the presence of material noise is lower than that for longitudinal waves; lower frequency measurements are currently in progress. The measurement model was then used to make preliminary predictions of the best experimental measurement technique for the detection of cracks located under cylindrically curved surfaces.

The frequency dependent ultrasonic attenuation and grain noise were determined for samples of three alloys often used in jet aircraft engine turbine discs: Waspalloy, IN-100, and Ti-6246. In addition to propagation of longitudinal waves, also shear waves were considered. The frequency dependence was extracted from broadband echos received through a low attenuation buffer. A key feature of the results for IN-100 is the presence of a low concentration of micropores which appear to influence the scattering of ultrasound and therefore the attenuation and material noise values. Of the three alloys, Waspalloy was found to have the highest attenuation value and Ti-6246 the lowest.

The purpose of this paper is to describe an integrated model for assessing the performance of a given ultrasonic inspection system for detecting internal flaws, where the performance of such a system is measured by probability of detection and other related quantities. The integrated model incorporates much of the work described in Thompson (1982), Gray and Thompson (1982), Thompson and Gray (1982), Addison and Elsley (1982), and Tittmann and Ahlberg (1982), The major challenges to these investigators were to properly account for the effects of real part geometries on sound propagation and to measure noise spectra due to various noise mechanisms. The results of these efforts could be incorporated into a model which computes a signal-to-noise ratio for any given transducer configuration and flaw state. The choice of an optimal transducer configuration might then be guided by such calculations.

This paper describes theoretical and experimental work directed toward finding the optimum probe dimensions and operating frequency for eddy current detection of half-penny surface cracks in nonmagnetic conducting materials. The study applies to probes which excite an approximately uniform spatial field over the length of the crack at the surface of the material. In practical terms, this means that the probe is not smaller than the crack length in any of its critical dimensions.

Eddy current signals obtained from variations in the probe liftoff are in general much larger in amplitude than the useful flaw signals. Small flaw signals can, however, be detected in the presence of liftoff noise if a large enough phase angle exists between them. Figure 1(a) shows how this phase discrimination can help in liftoff noise suppression. Here, the oscilloscope traces the complex impedance of the probe. The impedance plane has been rotated so that the liftoff noise lies entirely in the horizontal channel. Now if we choose to look only at the signal in the vertical channel of the scope, or the Q channel (in phase quadrature with liftoff), there will be no liftoff noise. This, however, is not a very realistic picture. Figure 1(b) is obtained when we try to detect much smaller flaws (in this case a closed crack of 20 mils in aluminum). We see that the trace of the liftoff noise has a curvature and that there are also fluctuations along the Q channel axis. Both of these effects eventually limit the detectability of small flaws. Since this contribution of liftoff to the Q channel is in practice larger than circuit noise, we define the detection figure of merit for an EC probe as 1$$D = \frac{{(\Delta {Z_f})\sin \beta }}{{{{(\Delta {Z_{\ell O}})}_Q}}}.$$

Recent research programs have concentrated mainly on developing techniques to characterize surface breaking cracks and very little has been done towards characterizing subsurface flaws in conducting materials. Presented in this paper are the results of some initial theoretical work aimed at the development of a reliable eddy current technique to detect and characterize defects in engine disk bolt holes under a 0.05″ stainless steel sleeve. The change in impedance of an absolute eddy current coil with and without ferrite core, and the distribution of eddy currents around a second layer crack with and without a thin insulating film between the two conducting layers have been predicted numerically. The overall system development goals and methods to accomplish them are outlined briefly.

Calculations of the change in probe impedance produced by a flaw in or beneath the surface of a conducting material are necessary for the analysis of eddy-current flaw-detection systems. However, probe sensititivy alone does not completely determine the flaw-detection capability of such a system; the effect of noise and clutter in the system must also be considered. The objective of this work is to develop a statistical detection model for an eddy-current system, and to use this model to calculate probability of detection as a function of probability of false rejection (false alarm) for the purpose of determining the optimality of a particular system. The target flaw chosen for these calculations is a 0.015-in.-long, 0.007-in.-deep surface crack in a bolt hold in a turbine disk made of material such as IN-100, which has a conductivity of about 106 mho/m.

The inspection of butt-welded stainless-steel pipe joints in nuclear power plants is routinely performed using ultrasonic non-destructive evaluation methods. Field experience, based on conventional ultrasonic signal amplitude criteria, indicates that a large number of indications are recorded. Most of these are not due to cracks, but are inherent in the geometry of the specimen. Discrimination between crack and geometrical/weld (malignant vs. benign) indications is principally based on operator experience, variations in signal amplitude, and the location of the reflector. Field experience and round-robin tests show that indication discrimination is a very time-consuming process. Besides, significant variations in performance exist due mainly to operator experience, fatigue, concentration, and conventional signal amplitude evaluation criteria.In response to this problem, this paper describes an artificial intelligence methodology and results for classification of intergranular stress-corrosion cracking (IGSCC) from geometrical/weld reflectors in austenitic stainless-steel pipes. This algorithm was developed using the protocol method of artificial intelligence heuristic programming and, as such, can provide answers comparable to those supplied by well-trained technicians during the flaw discrimination process, i.e., discrimination between crack and geometry/weld ultrasonic signals. Preliminary results show that this approach yields a better than 90-percent index of performance.

High performance weldments for critical service applications require 100% inspection. Balanced against the adaptability of the ultrasonic method for automated inspection are the difficulties encountered with nonhomogeneous and anisotropic materials. This research utilizes crystals and bicrystals of nickel to model austenitic weld metal, where the anisotropy produces scattering and mode conversion, making detection and measurement of actual defects difficult.Well characterized samples of Ni are produced in a levitation zone melting facility. Crystals in excess of 25 mm diameter and length are large enough to permit ultrasonic measurements of attenuation, wave speed, and spectral content. At the same time, the experiments are duplicated as finite element models for comparison purposes.Finite element models permit easy description of boundary conditions, geometry, and loading. Direct integration of the wave equation is done with the Newmark-Beta and Wilson-Theta Methods. The usual problem with the large number of degress of freedom can be alleviated with the use of Guyan reduction.Two-dimensional comparisons showing mode conversion and a plate with a flaw are made. The continued development of this computational tool should increase understanding of quantitative ultrasonic inspection.

Accompanying the requirements for higher quality welds in structural parts, there is a growing demand for more rapid and automatic methods for their nondestructive inspection. Conventional X-ray methods are inherently slow and are difficult to automate. Furthermore, they often present a safety hazard that adds mass and bureaucracy to the application of the method. Ultrasonics, on the other hand, can be very rapid, is easily made automatic, and does not present any safety problems. Its main drawbacks are the requirement for a carefully aligned transducer, a plumbing system to supply liquid couplant, and an educated operator to maintain the alignment and coupling throughout the scan of the weld line. Since electromagnetic acoustic transducers (EMATs)1 eliminate the need for a couplant fluid, they would appear to offer a major improvement for weld inspection technology.

A new type of electromagnetic-acoustic transducer (EMAT) has been developed that may be particularly suitable for use as an element of ultrasonic arrays. The new transducer can generate and receive compact ultrasonic pulses that exhibit a component of polarization parallel to the free surface. In the plane of symmetry that is normal to the free surface and bisects the EMAT (the sagittal plane), the ultrasonic signals generated by the new transducer are SH waves. In addition, the new transducer can efficiently receive ultrasonic signals from a very wide range of direction in the sagittal plane. This property is required to realize very long synthetic aperture lengths, which are needed to maximize the transverse resolution of ultrasonic inspection systems. The focusing performance of different linear synthetic array configurations using the new EMAT is compared analytically with that of a linear end-fire system using per iodic-permanent-magnet (PPM) EMATs that have been used in the past in weld inspection. The advantages and inherent limitations of such systems are examined using analytical and numerical methods. Particular emphasis is placed on the analysis of arrays that can be focused obliquely with respect to the free surface and the potential usefulness of such systems in weld inspection.

At the 1980 Review of Progress in Quantitative NDE the authors presented a paper outlining the NDE techniques then under development for fatigue crack detection and monitoring in welded structures (1). The present paper describes the progress made since then in applying the techniques to welded pressure vessels.

Ultrasonic waves are attenuated as they propagate past the tip of a crack due to the reflection of the energy at the crack face and diffraction at the crack tip. Crack closure modifies the situation since partial transmission can occur at points along the crack face where asperities come in contact. This phenomenon is important in defining the ability to nondestructively detect closed cracks and in developing a more detailed understanding of the closure phenomenon itself. Modified compact tension specimens were used to investigate the effects of partial crack closure on focussed, through-transmission ultrasonic signals. Data obtained from fatigue cracks in 7075-T651 A1 provides evidence for a gradual transition from a fully closed crack condition at the crack tip to an essentially fully open condition at a distance of a few mm from the tip, with additional localized contact along the length of the crack. This interpretation of the data was aided by a two-dimensional, quasi-static model for ultrasonic interaction with a partially contacting interface. The model relates width and separation of asperity contacts to the frequency dependence of the ultrasonic reflection and transmission. These measurements were supplemented by tests in which water infiltrated into the crack opening. The frequency spectra of the ultrasonic transmitted signals for this case were used to estimate the average COD at various points along the crack length.

It is well known that partial contact of two rough crack surfaces will lead to transmission of an acoustic signal across the crack, thus giving rise to a reduced probability of detection (POD). To explore the effects and consequences of such partial contact, impression experiments—using small spheres—have been performed to determine the effects of contact area on the amplitude transmitted. The results have been compared with a theory described elsewhere in these Proceedings. Based on the experimental results it will be speculated that the residual stress field responsible for the crack closure may be calculated based on a determination of the size and separation of the contact areas.

This paper reviews some recent work on the detection and sizing of closed internal fatigue cracks by ultrasonic techniques. Major emphasis is put on the diffraction of shear waves at the crack tip. Both fully open as well as partially closed cracks were considered. The effect of crack closure stress on back- scattered (pulse-echo) shear waves was studied with the aid of an A1 compact tension specimen. Noticeable changes with crack closure stress were documented for the structure of both the time- domain and frequency-domain representations. The techniques acquired with this specimen were applied to the study of a 50 μm radius semi-circular crack internal to a diffusion bonded Ti-alloy plate. Improved signal processing techniques were employed to detect the crack and to distinguish it from an artificial surface crack. The probability of detection, assumed to be proportional to the signal-to-noise ratio, was measured as a function of crack interrogation angle and crack closure stress to provide data on optimum probability for detection and sizing. Vigorous research efforts on good models for closed cracks in specific materials and environments are needed to refine the techniques of detection probability.

This section of the Proceedings is devoted to the problem of technology transfer and each of the other papers describes what I would call a formal approach to the problem. That is, they describe organizations with specific charters to perform technology transfer (such as the EPRI NDE Center) or specific contracts to bridge the gap between the laboratory and the production floor (such as the RFC program). My purpose is to explain how it REALLY happens in the majority of cases. Specifically, I will describe the key role played by individual entrepreneurs and the personal chemistry that is needed to spark the separate steps along the treacherous trail of technology transfer.

In the world of technology today, scientists and engineers are concerned about having the results of their work proceed through development to application. This concern exists because of a funding-source pressure to be relevant. One response to this pressure is to become a method evangelist and another is to transfer emphasis from research to technology development. However, a way is available to avoid this unfortunate circumstance and that is to focus on appropriate technology.

A combined finite element and analytical method is presented here for analyzing scattering of time harmonic horizontally polarized shear (SH) waves by material and geometric irregularities in an isotropic linearly elastic infinite plate. All the irregularities are assumed to be contained in a bounded region. The problem of scattering is solved by replacing this region with a finite element mesh. A nodal force-displacement relation is developed to satisfy the continuity conditions along the boundaries separating the inner finite-element region from the exterior regular region. The method is illustrated by solving the problem of scattering of SH waves by a surface breaking crack. The crack is taken to be normal to the surface of the plate. The reflection and transmission coefficients are computed for zeroth, first, and second incident wave modes. The validity and accuracy of the results are checked by satisfaction of the energy conservation principle and the reciprocity relations.

Of strong practical interest in the subject of fracture mechanics and related structural design and damage are the surface-breaking and sub-surface cracks. Indeed the scattered-wave fields generated by the interaction of incident surface or body waves with these cracks would be expected to yield most of the important in-formation about the geometries of the cracks. It follows, in the subject of quantitative non-destructive evaluation (QNDE) there is considerable interest in scattering by surface-breaking and sub-surface cracks, as important steps in solving the inverse problems of obtaining the crack geometries from the scattered wave fields. However, having the solutions to the corresponding direct problems is a prerequisite to the inverse problems.

In the last ten years several numerical methods have been developed for the solution of elastic wave scattering problems that have found application in quantitative flaw definition. Before the development of these methods, due to the complexity of Navier’s equation which governs wave motion in an elastic continuum, numerical results were available only for circular cylinders and spheres. The elastic wave equation is separable only in polar and spherical coordinates. For other geometries, three types of numerical methods have been developed. They were all originally developed for acoustic and electromagnetic problems governed by the scalar and vector wave equations respectively.

This paper provides an overview of an ongoing program to utilize acoustic emission methods to detect fatigue crack growth in aircraft structure during operation. Experimental methods including instrumentation, test specimens, newly-developed sensors, and experimental procedures used in acquiring acoustic emission waveforms are described. Fretting noises are produced at the interfaces between 7075 aluminum plates and a steel bolt during fatigue cycling of three-part joint specimens. A two-channel waveform recorder and analyzer records the experimental data for subsequent pattern recognition analysis. The observed signals cluster into three distinct groups on the sinusoidal load curve. By varying experimental procedures, these groups of signals can be associated with crack growth, fretting, and crack opening.Another apsect of the project is the development and evaluation of two new sensor designs having improved bandwidths. Calibration information and performance of these sensors in fatigue test monitoring are discussed. It is possible to obtain well-defined data sets from joint specimens that can be used in pattern recognition analysis. The results of such analysis is given in a companion paper. The future direction of this project in leading to in-flight AE monitoring of aircraft is also presented.

Computer pattern recognition has been used to identify and separate acoustic emission (AE) signals that are similar in appearance but are due to different sources. Simulated joint specimens were tested in the laboratory in which a fatigue crack was grown from the edge of a central loading pin hole. The hardened steel loading pin produced fretting AE by its contact with the 7075 T651 aluminum plate specimens during cyclic loading. The fatigue crack produced AE due to crack growth and to crack face rubbing during load cycling. The AE signals detected at two transducers mounted on opposite sides of the loading pin hole, at 2 in. and 4 in. from the fatigue crack, were digitally recorded at a 5 MHz digitization rate. The waveform features that were extracted from these AE signals and used in the pattern recognition were derived from the frequency spectral content of the waveforms. Better than 90% separation of crack growth from crack face rubbing was achieved using frequency features of the waveforms from either transducer separately. Better than 95% separation of fretting from crack growth or crack face rubbing, separately or combined, was achieved using the ratios of the spectral energies detected at the two transducers.

The objective of the work described in this paper is to develop signal analysis techniques that can automatically discriminate be-tween non-critical acoustic emission (AE) from crack growth and acoustic noise signals, such as fretting, of fasteners. The ultimate application of this work is for in-flight AE monitoring of critical aircraft structures.Fatigue crack growth experiments were performed with center notched plate specimens and simulated joint specimens of 7075-T651 aluminum. The experimental conditions were controlled such that acoustic signals were obtained from crack growth, crack interface rubbing, and from fastener fretting.This paper reports the results of pattern recognition analysis of the signals using autocorrelation lags and statistical measures of the signals and their power spectra as features. The goal of the pattern recognition analysis was to isolate crack growth AE signals from the other acoustic data. The results indicate that autocorrelation lags are the most important features for discriminating these signals.

This paper describes Acoustic Emission Linear Pulse Holography which combines the advantages of linear imaging and acoustic emission into a single NDE inspection system. This unique system produces a chronological linear holographic image of a flaw by utilizing the acoustic energy emitted during crack growth.Conventional linear holographic imaging uses an ultrasonic transducer to transmit energy into the volume being imaged. When the crack or defect reflects that energy, the crack acts as a new source of acoustic waves. To formulate an image of that source, a receiving transducer is scanned over the volume of interest and the phase of the received signals is measured at successive points on the scan.The innovation proposed in this paper is the utilization of the crack generated acoustic emission as the acoustic source and generation of a line image of the crack as it grows.A thirty-two point sampling array is used to construct phase-only linear holograms of simulated acoustic emission sources on large metal plates. The phases are calculated using the pulse time-of-flight (TOF) times from the reference transducer to the array of receivers. Computer reconstruction of the image is accomplished using a one-dimensional FFT algorithm (i.e., backward wave).Experimental results are shown which graphically illustrate the unique acoustic emission images of a single point and a linear crack in a 100 mm × 1220 mm × 1220 mm aluminum plate.

The relation of the occurrence of airframe acoustic emissions to aircraft manoeuvre are reported for Avro CF-100 upper forward wing trunnions. Periods of excessive noise are found when the airframe load is changing during entry to and exit from sustained-G manoeuvres. During constant-G periods, the airframe noise level is reduced by a factor of more than one hundred. These quiet periods provide a suitable signal-to-noise level for the in-flight detection and monitoring of slow, stable crack growth in common airframe materials, even in a noisy load transfer component such as the wing trunnion studied here. Simultaneous in-flight acoustic emission measurements in symmetrically-located airframe components are also reported. The ratio of the number of recorded event counts in a cracked component to that in an uncracked component during the same flight is found to increase linearly with the crack face area for through crack lengths in the range 0–5 mm.

The optical transduction of acoustic emission signals offers many advantages over piezoelectric devices. These include high bandwidth, no modification to the signal as well as providing contactless measurement. The major difficulties associated with optical devices are stability against low frequency vibrations and the generally complex nature of an optical interferometer. This paper describes the attempts to miniaturize a Michelson interferometer while at the same time overcoming some of the stability problems associated with these devices.Active stability of an interferometric transducer with dimensions of ∿ 5cm (2″) cube has been achieved over 8 fringes of red light at 100Hz and 4 fringes at 300Hz. The range of active stabilization of the interferometer is limited by the frequency response of the large amplitude piezoelectric element and the filter characteristics of the feedback electronics. A sensitivity of 0.5Å (0.5 × 10-10m) has been achieved.

The scattering of Rayleigh surface waves by a variety of defects has been experimentally studied by the use of ultrasonic spectroscopy in the frequency range of 1 – 10 MHz. The defects are: (1) a through edge crack, both normal and inclined; (2) a cavity; (3) four different fluid inclusions, namely, carbon tetrachloride, water, mercury, and glycerine; (4) a solid inclusion; and (5) a shell. The transmission coefficient, A_T, defined as the total transmitted field normalized with respect to the incident field, is calculated from the FFT of the two signals. (The receiver is located sufficiently far away from the defect so that it only senses the far-field.) Whenever possible, the phase difference caused by the presence of the defect, ⌽, is also measured. The spectroscopic measurements were verified by the more accurate (albeit far more time consuming) tone-burst method. The classical problem of a quarter-space was used as a test case. The experimental results were found to be in excellent agreement with some of the recent analyses.It is concluded that ultrasonic spectroscopy can be used as an efficient NDTE tool. In particular, the “signature” of a variety of inclusions is shown to be quite distinct, a fact that should prove very useful from the viewpoint of the inversion problem.

The method of equivalent inclusion, originally developed by Eshelby1,2 was first applied to study dynamic behavior of composite laminates by Wheeler and Mura3 where they neglected the presence of the mass density mismatch. Fu4,5 gave a similar formulation for the problem of a pair of ellipsoidal inhomogeneities embedded in an infinite elastic medium subjected to time-harmonic waves. A complete formulation and presentation of analytic and numerical results for a single ellipsoidal inhomogeneity are subsequently given by Fu6 following his work on the ultrasonic determination of fracture toughness, References (7) and (8). The eigenstrain approach is employed and the eigenstrains are expanded as a geometric series in position vector.

Ultrasonic theories generally predict a scattering amplitude which relates a spherically spreading, far-field scattered wave to an incident plane wave. In ultrasonic immersion measurements, the frequency and angular dependences of the scattering amplitude are convolved with those of the transmitting and receiving transducers and the propagation through the liquid-solid and solid-liquid interfaces. This paper presents a set of approximate corrections for these effects for the cases of angle beam inspection through planar, spherically curved or cylindrically curved surfaces. The primary parameters in the correction are the function D, which corrects for the diffraction effects occurring during a transducer calibration experiment, and the function C, which describes the on-axis pressure variation of the beam. Values of C and D are available in the literature for the case of a piston transducer radiating into an infinite fluid medium. The major portion of this paper is concerned with the extension of those results to the aforementioned two media problems in which mode conversion, refraction, diffraction, and focussing all play interrelated roles. Results of preliminary experiments to test the corrections are also included.

The high specific strengths attainable from composite materials can be negated by pores between the plies in the material. Classical acoustic attenuation measurements can indicate the presence of porosity if all other material properties remain constant. Even though attenuation measurements accurately predict the amount of porosity, its spatial distribution cannot be inferred from these data. To make accurate predictions of structural performance, current micromechanical models of composite materials require accurate estimates of the volume fraction of porosity and its spatial distribution. This work is an attempt to model the acoustic interactions that occur between plane compressional waves and the fiber arrays and pores in order to develop an ultrasonic technique to estimate the spatial distribution and volume fraction of porosity.

The distortion and displacement of a finite-aperture acoustic wave incident from a fluid onto a fluid-solid interface at or near the Rayleigh critical angle is phenomenon which has received considerable attention recently1–5. Physically, resonant generation of Rayleigh-like waves and rapid reradiation of the surface-wave energy into the fluid are responsible for the characteristic displacement and distortion of the acoustic beam. These effects typically lead to a bimodal reflected acoustic field, composed of a remnant specular reflection summed coherently with the radiating surface wave. When conditions of beam width and wavelength are favorable, a portion of the field distribution shows a strong reduction in amplitude, due to the phase cancellation of component fields. The sensitivity of this null zone to surface condition (including the presence of coatings) suggests that a NDE technique may be based on this effect. The situation is illustrated schematically in Fig. 1. An acoustic beam incident at the Rayleigh angle suffers displacement and distortion with most of the power contained in the shaded regions. The suppressed specular reflection is indicated by the dashed lines.

A code for calculating ultrasonic fields has been developed by revising the thermal-hydraulics-structures code STEALTH.1,2 This code may be used in a wide variety of situations in which a detailed knowledge of a propagating wave field is required. This paper describes the preliminary results of the code applied to a realistic experimental situation.

The scattering from rough surfaces and cracks in the high frequency regime is analyzed via a scattering formula based on the reciprocity relation. Scattering from the smooth cracks is investigated first to rederive the “flash point” concept by Fourier transform methods. Based on this analysis, an inversion procedure is proposed for obtaining the characteristic function of the crack, which, for the case of rough cracks, gives information about the roughness as well as the dimensions and shape of the crack. The theory is applicable to both 2-D and 3-D scattering problems, as well as surface wave scattering from surface breaking cracks. Elastodynamic ray theory predicts that scattering from cracks can be described in terms of discrete source points on the contour of the crack.1 These points are generally called the “flash points”, and their positions depend on the transmitter and receiver locations as well as the crack shape. For instance, for 2-D scattering problems, (or for deep surface breaking cracks under surface wave excitation), the two edges of the crack act as the flash points.

Most ultrasonic NDE experiments and their theoretical models deal with perfectly smooth interfaces, but true materials generally exhibit rough interfaces. As an approach to include the ultrasonic scattering which occurs on the different interfaces along the beam path, the reflection factors of acoustic waves diffracted by periodic surfaces is investigated theoretically and experimentally by looking at the frequency dependence of the reflected signal. Mode conversion bulk and surface waves are shown to be the result of strong coupling between the incident wave and the geometry of the grating. As a consequence, the geometrical parameters of the interface can be obtained to within 5%.

An important problem of many complex systems is that of assessing the reliability in the minimum probability of survival, from the reliabilities of the components. If system performance can be represented by a network flow diagram, the system probability of failure can be expressed in terms of path probabilities of failure, each of which is a function of component failure probabilities.

The weakest link in the inspection process is the subjective interpretation of data by inspectors. To overcome this troublesome fact computer based analysis systems have been developed. In the field of nondestructive evaluation (NDE) there is a large class of inspections that can benefit from computer analysis. X-ray images (both film and fluoroscopic) and acoustic images lend themselves to automatic analysis as do the one-dimensional signals associated with ultrasonic, eddy current and acoustic emission testing.Computer analysis can enhance and evaluate subtle details. Flaws can be located and measured, and acceptance decisions made by computer in a consistent and objective manner.This paper describes the interactive, computer-based analysis of real-time x-ray images and acoustic images of graphite/epoxy adhesively bonded structures.

In order to ensure a minimum strength for the low density fibrous tiles to be used in the thermal protection system of the Space Shuttle an NDE test that uses sonic velocity measurements to predict strength has been developed. The empirical correlation between strength and sonic velocity which is the basis of this test, is shown to be consistent with a previously developed micro-mechanical model. The model is reviewed and is shown to describe the fracture behavior of these fibrous materials regardless of their density or testing direction. Measurement of the density and sonic velocity in these materials allows Young’s modulus to be calculated and as low modulus materials tend to have low fracture toughness and strength, they can be eliminated by this test.Appropriate accept/reject criteria can be developed statistically from the strength or fracture toughness correlations but for use in design, other factors such as the size-dependence of strength, the stress fields encountered during flight, and the property variations within tiles should also be incorporated.

An initial study of a technique proposed for the nondestructive testing of metal matrix composites is the subject of this paper. These composites are manufactured in the form of approximately 1/2-mm-diameter “precursor” wires. Larger structures are fabricated by diffusion bonding of lay-ups. Reliable nondestructive quality control indicators of wire integrity have not yet been developed although a number of possibilities are being examined.1 Testing of the precursor wires is difficult because current manufacturing processes produce wires that may be entirely satisfactory but that vary in cross-sectional geometry, in surface properties, and sometimes in the amount of matrix material that is present. Techniques based on observations of wire resistance, surface emissivity, and sound emission signatures are difficult to interpret because of these characteristics. Wire imaging using x-ray or neutron techniques is also difficult because large lengths of wire must be examined with a resolution in the plane of the wire exceeding 50 line pairs per millimeter.

The objective of the work was to nondestructively examine the effects of thermal cycling on properties (e.g., torsional velocity and electrical resistance) of graphite/aluminum (Gr/A1) precursor wire. Precursor wires were periodically heated at selected temperatures for various periods of time; the exact heating period being dependent on the test temperature. Wires were tested in the as-received condition and after each thermal excursion. Changes occurred in both the torsional velocity and the resistance; although at different rates and to different extents. Subsequent longitudinal velocity tests run on the heat treated sections of the wires showed little if any changes in the longitudinal velocities.

The nondestructive evaluation of metal matrix composite precursor wires is being pursued by measurements of the attenuation and velocity of both torsional and longitudinal ultrasonic pulses propagating along the wire axis. The sound waves are generated by non-contacting electromagnetic transducers. Continuous scans of the attenuation and velocity are made along wires of any length by use of a four-transducer arrangement. The attenuation and velocity have been related to physical wire properties which are important for nondestructive evaluation and characterization.

The objective of the subject program was to apply nondestructive evaluation (NDE) methods to assess the integrity of FP/Mg composites. The material investigated was ZE41A magnesium alloy reinforced with FP (aluminum oxide) fiber. Twenty-one specimens (three specimens of each of six flawed and three unflawed specimens) were evaluated using ultrasonic scanning, wave propagation velocity, wave attenuation coefficient, and x-ray radiograph inspection techniques. The results for two of the 21 specimens are included herein.After the NDE inspections were completed, a representative specimen from each of the seven groups was sectioned and micrographs were made for comparison with the NDE records. It was found that ultrasonic scanning using a 15 MHz compression wave, focused transducer operated in the pulse-echo mode generating an analog C-scan gave the best pictorial results. The wave attenuation and wave propagation velocity measurements were found to be consistent with the ultrasonic C-scans, but x-ray radiography was useful only at locations of gross material defects.

A review is presented of work performed in our laboratory on the nondestructive inspection of metal matrix composites. In order to obtain damage representative of that which occurs in service, the specimens were mechanically loaded to intermediate load levels below that which causes final, catastrophic failure. Various nondestructive techniques were used both during and after the applied loadings to follow damage initiation and progress.

In ultrasonic inspection one frequently encounters situations in which flaws lie too close to surfaces (or interfaces) for the applications of the conventional theory of scattering in an infinite host medium. One approach, pursued by Varadan, Pillai, and Varadan, is to rework scattering theory with a Green function suitably modified for the presence of surfaces. Our approach uses a Fourier elastodynamics formalism in which the scattering process is represented by the generalized transfer function for a slab of material containing the flaw. A sufficiently simple elastodynamic system can be built up of separately characterizable subsystems defined by intervals on the z-axis. Fortunately, it turns out that, in the case of propagating waves the generalized transfer function can be simply related to the scattering amplitude, the representation of scattering in the conventional theory. In the case where evanescent waves are present, the generalized transfer function can also be simply related to the scattering amplitude in which the incident and scattered direction $$\mathop e\limits^{{ \to _1}}$$ and $$\mathop e\limits^{{ \to _S}}$$ have been analytically continued from real to complex forms. In some cases it may be a good approximation to ignore completely the effects of evanescent waves. An integral equation has been derived whose solution gives the scattering from a system composed of a flaw and a nearby surface or interface. The computational results pertain to the scattering from spherical voids and inclusions in plastic at various distances from the plastic-water interface with the incident wave propagating through the water.

In recent years, attention has shifted from the ultrasonic NDE of bulk flaws to flaws in structural components that involve nearby surfaces. Very often it is convenient to exajnine the part in a water tank, locating the transmitting/receiving transducers in water. This paper is concerned with a theoretical study of a near surface flaw in an immersion tank experiment. The analytical problem to be studied is the scattering of ultrasound by an obstacle that is located near a solid-fluid (water) interface. The distance ‘d’ of the center of the flaw from the interface ranges from ‘a’ to ‘15a’ where ‘a’ is a typical radius of the flaw.

In ultrasonic NDE measurements the detection of subsurface flaws is of practical importance, especially flaws too far from the surface to be detected by eddy current methods and yet close enough to the surface for the flaw-surface interaction to be important. In this paper we report experimental results of ultrasonic scattering measurements of subsurface flaws in the presence of a fluid-solid interface and compare these results with theoretical calculations of subsurface flaw scattering. Comparison of the absolute value of the scattering amplitude in terms of frequency, flaw-to-surface distance, ultrasonic mode and scattering angle will be made for an oblate spheroidal void in the interior of bulk titanium and for a spherical inclusion near the surface of a thermoplastic sample. Results of applying the one-dimensional inverse Born algorithm to the sizing of near-surface flaws are also reported.This work was sponsored by the Center for Advanced Nondestructive Evaluation, operated by the Ames Laboratory, USDOE, for the Air Force Wright Aeronautical Laboratories/Materials Laboratory and the Defense Advanced Research Projects Agency under Contract No. W-7405-ENG-82 with Iowa State University.

It is the main function of quantitative NDE to detect and to evaluate defects. Some of the most dangerous defects are cracks, especially cracks on or near surfaces. These cracks can be found by scattering ultrasonic waves from them (either bulk waves or surface waves), but up to now there is no theory (at least in the most interesting low-to-intermediate frequency region) which has been implemented to compute scattering from surface or near-surface cracks in 3d. The purpose of the present report is to explain, via a simple scalar example, the principles of a general boundary-integral-representation method which has been used1 to calculate scattering of waves of all polarizations by a 2d surface or subsurface crack. The method is developed for bulk defects and cracks in a slab as well as in a half-space, and is straightforwardly applicable to 3-dimensional problems as well as to 2d ones.

The sensitivity of the propagation of an elastic wave to changes in the microstructural details of a material is well known.1 In particular, numerous experiments have shown that the attenuation of the wave is sensitive to the inclusions, voids, cracks, grain boundaries, twin boundaries, interphase boundaries, magnetic domain walls, dislocations, substitutional impurities of a material. For attenuation studies in metals, ceramics and polycrystals, three formulas, each for different wavelength regimes, are generally used in the quantitative interpretation of experimental results.1–3 If λ is the wavelength of the elastic wave and <D> is the average grain diameter, then in the Rayleigh regime (λ≫D), α = A₁<D>3λ4, in the stochastic regime (λ≃D), α = A₂<D>λ2, and in the diffusive regime (λ≪D), α = A₃/<D>-1. By fitting the data to these formulas, one tries to infer <D>.

Reported here are some preliminary computer simulation results on the feasibility of using oblique incidence ultrasound for the detection and estimation of porosity in composites. In the oblique incidence approach, the composite is illuminated at off normal angles in such a manner that the reflected returns from the fibers are in directions away from the illuminating transducer. Since the scattered returns from porosity tend to be more omnidirectional than the fiber returns, there is a larger received signal in the presence of porosity.

The purpose of the work described in this paper is the development of an ultrasonic measurement technique which provides a convenient way to detect dilute porosity conditions in materials and to extract certain properties of the flaw distribution which are important in failure prediction. Use has been made entirely of ultrasonic backscatter measurements; thus, the technique differs considerably from other investigations which lead to porosity determinations in that no reliance is placed upon either attenuation measurements or precise ultrasonic velocity measurements [1,2]. The technique thus possesses a distinct advantage for practical implementation, i.e., it is a “one-sided” measurement which does not require ultrasonic echo returns from an opposite face of the sample in order to be useful. At present, the work is limited to dilute porosity concentrations. Reasons for this limitation will become clear in the paper. With additional effort it is expected that this limitation can be removed and the work extended to larger concentrations.

It is well known that when a finite ultrasonic beam of a given spatial distribution is incident at the Rayleigh angle to a liquid-solid interface, the spatial distribution of the reflected field may be altered significantly. The “energy redistribution” is due to the interference between the specularly reflected beam and a surface wave which has leaked back to the water. The “shape” of the reflected field depends on the so-called Schoch displacement (which is characteristic of the interface) and on the width of the ultrasonic beam. It has also been observed that significant energy is scattered back to the transmitter at the Rayleigh angle. Experimental results will be presented on the evaluation of the parameters effecting the back-scattered amplitude. The backscattered Rayleigh angle phenomena are also applied to measured surface wave velocities of anisotropic materials such as casts and welds.

Calculation of attenuation and dispersion spectra for a general anelastic body may be posed as an inversion problem. We observe the time-dependent strain due to “instantaneous” changes in stress in order to characterize the anelastic response. This experimental technique reaches a range of frequencies which is lower than that used in resonance bar experiments, but we can make only indirect measurement of the viscoelastic properties. We employ the most general anelastic model which states that the observed compliance is due to the summed effect of an arbitrary spectrum of mechanisms. The analysis requires the solution of Fredholm integral equations of the first kind. It is well known that this problem is ill-conditioned so that any numerical scheme will have to involve some smoothing to obtain accurate solutions. The present work employs Butler’s method of constrained regularization which takes advantage of the fact that the solution is positive and uses data dependent smoothing. This work indicates that the imposition of the positivity constraint makes the computation of the solution much better conditioned. Computations with the method of constrained regularization employing near-optimal smoothing demonstrate its superiority over the method of Shapery for obtaining accurate solutions when the data are very noisy.

The problem of inverse ray tracing in a homogeneous anisotropic elastic solid is considered, with specific application to crack sizing. The data is assumed to be in the form of travel times of diffracted ultrasonic signals between transducers positioned on an exterior surface of the body. Both pulse-echo and pitch-catch data are considered. First, it is assumed that the wave speeds are unknown and must be obtained as part of the inversion procedure. The specific problem of locating a crack tip in a two-dimensional geometry is investigated. It is found that travel time data on the exterior surface suffices to locate the crack tip only if the material is isotropic. If the material is anisotropic, we must be able to move the source and/or receiver in the direction normal to the surface. The same problem is considered with the source and receiver positioned in a surrounding isotropic material, e.g., a water bath. It is shown that the ray inversion is now possible only if the solid is isotropic, the problem being underdetermined for an anisotropic solid. Numerical results are presented for a synthetic experiment in which a finite crack is present in some anisotropic elastic solids. Next, the problem is considered when the speeds are known a-priori. It is shown that a crack edge can be mapped by a local approximation procedure.

We study the recently derived reflection coefficient for plane waves in a liquid that are incident on the liquid-solid interface of a solid half space which consists of a single layer of one elastic material bonded to a substrate of a different material. Plots of the magnitude of the reflection coefficient versus the incident angle are presented for several sets of material parameters and values of frequency f and layer thickness d. The use of the results presented for the study of nonspecular reflection of bounded acoustic beams is of primary interest. We therefore seek to identify all the critical incidence angles for nonspecular reflection.We also investigate, in particular, the surface wave propagation for the case of a stiff layer on a soft half space, and we find that the purely propagating mode cuts off with increasing fd (f is the frequency and d the layer thickness) when its speed reaches approximately the shear wave speed of the substrate, as reported in the literature. However, as fd increases further, a leaky mode appears that approaches the Rayleigh wave for the layer. This leaky mode is also associated with nonspecular reflection for large enough fd.

This paper introduces the Maximum Entropy method of resolving underdetermined objects (flaws or inclusions) by a physicists’ brand of nonlinear processing of the image data. We survey three areas of research: (1) synthetic aperture imaging to resolve three dimensional flaws, with the aid of selective back projection, (2) scattering from anomalies according to the inhomogeneous Fredholm integral equation of the second kind, and (3) ultrasonic flaw characterization by the boundary integral equation method. A simple example is offered to illustrate the potential resolving power of the ME method for problems in area (2). We present some criteria for effective ME inversion.

The Inverse Born Approximation (IBA) to the elastic wave inverse scattering problem is known to give highly accurate results for the shape of complex voids. In this paper we present an argument demonstrating that the IBA is, in fact, exact for determining the size, shape and orientation of a wide class of these scatterers given infinite bandwidth and unlimited aperture information. Essentially, our argument demonstrates how the IBA algorithm picks out the singular contribution to the impulse response function and correctly relates it to the shape of the scatterer. Some specific examples will be used to illustrate the more intuitive aspects of the discussion.

This paper presents results of a study on an inversion transformation ℐ_A for backscattering of ultrasonic waves from an obstacle embedded in a solid. For a rigid sphere the resulting function A(t) is derived which is related to the cross-sectional area intercepted at any given time by a transverse plane moving across the obstacle in the same direction as the incident wave. The maximum value of A(t) is the total backscattering cross section. From this study emerges a proposed inversion algorithm for cavities. The use of the inversion algorithm is demonstrated for “exact” theoretical data obtained from Opsal as well as experimental data collected on an ellipsoidal cavity and a double cavity. Associated with the inversion of the experimental data is an iterative technique which optimizes the construction of A(t) and therefore the estimates for the size and shape of the scatterer.

The Born approximation has been widely employed as a basis for determining flaw sizes using individual pulse-echo waveforms together with the assumption of an ellipsoidal flaw geometry. A major difficulty in implementing such algorithms has been the determination of the time delay corresponding to the flaw centroid. However, both the time delay calculation and the flaw size determination itself can be performed in an optimal fashion using statistical estimation techniques with an appropriate error model. We will discuss the application of these techniques to an automated flaw-sizing algorithm requiring a minimum of operator input, and will compare the results obtained by this method with those obtained by previous operator-intensive methods.

A computationally tractable inversion algorithm has been developed for the case of the scattering of longitudinal elastic waves from an inclusion. It is assumed that the material properties of the inclusion are known a priori but that the boundary geometry is unknown — in fact the boundary could belong to two or more separate inclusions. It is further assumed that the material properties of the inclusion are sufficiently close to those of the host that the Born approximation can be employed.

Considerable progress has been made in recent years in the development of signal processing algorithms for use in ultrasonic non-destructive evaluation which yield the size, shape, and orientation of a flaw. This kind of flaw information is necessary in order that failure predictions of materials and components can be made from non-destructive tests. The signal processing algorithms that have been developed for ultrasonics are based upon both direct and inverse approximate solutions to the elastic wave scattering problem, and cover various ranges of the parameter ka where $$k = \frac{{2\pi }}{\lambda }$$ is the wave number of the ultrasound and a is a flaw size dimension. In order to use these algorithms effectively in the determination of flaw parameters, it has been found necessary to obtain measurements of the flaw at several viewing angles. At this time, there is no ultrasonic transducer available which permits this to be done efficiently and conveniently in the long and intermediate wavelength end of the spectrum. This region has been shown to be quite rich in flaw information and is appropriate to ultrasonic NDE in many practical applications (e.g., thick wall sections).

The purpose of this paper is to evaluate the technique of scanning photoacoustic microscopy (SPAM) for the detection of fatigue cracks in metal alloys, and to describe an experimental arrangement for SPAM measurements on the inner surface of a cylindrical bolt hole. The experimental technique is based upon the physical mechanism of thermal wave imaging and has been described in detail at previous1, 2 Reviews of Progress in Quantitative NDE and elsewhere.3 In this paper we will also present some results of theoretical calculations for thermal wave scattering from closed, slanted cracks which intersect the surface of an opaque solid, and compare these results with our experimental data.

Although the use of thermoacoustic emission in ultrasonic non-destructive testing and in medical imaging has been discussed by Van Gutfeld, 1 among others, its application is principally confined to the use of laser beams and surface or thin-layer emissions. In this paper we propose a more general thermoacoustic method of non-destructive material testing in which deeply penetrating radiation generates thermoacoustic signals thoughout the volume of a material. These ideas are the outgrowth of early work by one of the authors, T. Bowen, in the use of thermoacoustic emissions in cosmic ray detection2 and experimental checks have been carried out in particle accelerator beams.3 Their application in the field of medical imaging has recently been discussed in two articles; one theoretical4 and the other presenting experimental results obtained using pulsed rf current on tissue phantoms and a human subject.5 In these articles it was shown that many forms of pulsed heating radiation, e.g., non-ionizing, ionizing, acoustical, etc., will induce thermal stress with the resulting acoustic emission yielding image information. It was also shown that the received acoustic pressure waveforms are proportional to the gradient of a function which depends on local thermal properties (coefficient of thermal expansion and specific heat) as well as the local density and the energy absorbed per pulse.

A simple electrochemical technique is described, which images and quantitatively measures the distribution and severity of fatigue damage in aluminum alloys. The technique is based upon (i) the creation of microcracks in a surface anodic oxide film during fatigue of the underlying metal, and (ii) the detection of these microcracks by contacting the surface with a gel electrode. When a voltage pulse is applied, current passes through the fatigue—induced microcracks in the oxide film, and an image of the sites of current flow is retained in the surface of the gel. The capabilities of the technique are illustrated by measurements on 6061-T6, 7075-T6 and 2024-T4 aluminum. The electrochemically formed images correlated directly with scanning electron micrographs of the specimens. Hairline fatigue cracks ≥10 μm long are easily imaged, while the charge flow during the formation of the image is a quantitative measure of the crack length. The accumulation of fatigue deformation prior to the appearance of a fatigue crack is also detected, and in this regard the sensitivity of the gel electrode exceeds that of a scanning electron microscope. The distribution of fatigue deformation may be mapped as early as 1% of the fatigue life, and the charge flow to the regions of most severe damage increases systematically with fatigue cycling as the density of microcracks in the oxide increases. The simplicity of this electrochemical gel electrode method renders it directly applicable to field investigations, and provides a new tool for quantitatively assessing the distribution of fatigue damage.

This article describes the evaluation of a new geophysical technique used to map fractures between boreholes: electromagnetic geotomography used in conjunction with salt water tracers. An experiment has been performed in a granitic rock mass. Geotomographic images have been generated and compared with borehole geophysical data: neutron logs, acoustic velocity logs, caliper logs and acoustic televiewer records. Comparisons between the images and the geophysical logs indicate that clusters of fractures were detected but single fractures were not.

The application of generic flaw sizing techniques to specific components generally involves difficulties associated with geometrical complexity and simplifications arising from a knowledge of the expected flaw distribution. This paper is concerned with the case of ultrasonic flaw sizing in turbine engine rotor components. The sizing of flat penny shaped cracks in the web geometry will be discussed and new crack sizing algorithms based on the Born and Kirchhoff approximations will be introduced. Additionally we propose a simple method for finding the size of a flat, penny shaped crack given only the magnitude of the scattering amplitude. The bore geometry is discussed with primary emphasis on the cylindrical focussing of the incident beam. Important questions which are addressed include the effects of diffraction and the position of the flaw with respect to the focal line. The appropriate deconvolution procedures to account for these effects will be introduced. Generic features of the theory will be compared with experiment. Finally, the effects of focused transducers on the Born inversion algorithm are discussed.

Two methods to map crack edges by the use of arrival times of edge-diffracted signals have been briefly reviewed. They are a global triangulation method and a local crack-edge mapping technique. The local mapping technique, which generally has the greater accuracy, has been applied to both synthetic and experimental data. The effect of a uniform error in the data has been investigated. The experimental data were obtained at the Rockwell International Science Center for a doorknob specimen with a plane crack of elliptical shape at its center. Crack mappings of very satisfactory accuracy were achieved.

This paper describes a new approach for classifying NDE waveforms. Using this approach a set of matched filters is constructed one for each category of waveform, based on parameters from autoregressive models. The method offers advantages in terms of hardware implementation over conventional pattern recognition approaches. Feasibility is shown using computer generated data. Results are then presented for real data from acoustic emission experiments.

Several techniques for characterizing flaws and inclusions using ultrasonic scattering information have been developed in recent years. These algorithms assume a noise-free medium and are sensitive to perturbations in the acquired spectra. However, surface roughness and volumetric porosity effects alter the available data.In order to determine the effects of surface roughness typical of aluminum castings on inversion accuracy, a sequence of experiments was performed. Ultrasonic backscattering data were acquired from spheroidal defects in flat, smooth surfaced, diffusion bonded titanium samples. Next, the scattering spectra were perturbed using theoretically or experimentally determined transmission spectra obtained from rough surfaced cast aluminum samples. Inversion procedures were applied and results analyzed.

One of the most important effects of complex part geometry is that the available entrance and exit angles for ultrasound are limited. We will present a study of the Inverse Born Approximation in which we have data for incident (and exit) directions confined to a conical aperture. Modeling the direct problem by the Born Approximation, we obtained analytical results for (1) a weak spherical inclusion, and (2) a penny shaped crack (modeled by an oblate spheroid). General results are: (a) the value of the characteristic function γ is constant in the interior of the flaw, but reduced in value; (b) the discontinuity at the boundary of the flaw occurs over the “lighted” portion of the flaw; (c) this discontinuity is contrasted by a region where γ is negative; and (d) new non-physical discontinuities and non-analyticities appear in the reconstructed characteristic function. These general features also appear in numerical calculations which use as input strong scattering data from a spherical void and a flat penny shaped crack in Titanium. The numerical results can be straightforwardly interpreted in terms of the analytical calculation mentioned above, indicating that they will be useful in the study of realistic flaws. We conclude by discussing the stabilization of the aperture limited inversion problem and the removal of non-physical features in the reconstruction.

The objective of the ongoing work reported here is to optimize pulsed eddy current instrumentation for the detection of small flaws, specifically 250 × 125 pm crack-like defects in low conductivity materials such as titanium 6-4. Our approach parallels that employed in the optimization of continuous wave eddy current systems1,2 in that we seek first to develop a mathematical model of a typical flaw detection system, and then use the model as a guide in the design of probe instrumentation and in the selection of signal processing methods to enhance flaw detectability.

The classical work of Dodd and his coworkers at the Oak Ridge National Laboratory deals with the analysis, design and optimization of eddy-current probe coils wound around an air core. Many applications, however, require that the magnetic field produced by the probe coil be “shaped” or confined to certain regions of space, especially at higher frequencies, and this necessitates the use of highly permeable core materials, such as ferrites.

Preliminary studies performed during a previous investigation at SRI showed that a modified floppy-disk tape head operating at 100 kHz is a very sensitive crack detector. For example, tests on 0.25-in.-long fatigue cracks in aluminum produced signals an order of magnitude larger than those obtained using a commercial 100-kHz coil probe. The floppy-disk probe contains ferrite material, but its construction is different from other ferrite-containing eddy-current probes.

Presently, the rotary wing-head and hub subassemblies of the Army’s Black Hawk helicopter require almost complete disassembly to inspect failure critical threads of the main spindle. Even with direct access to the threads, detection of fatigue cracks in the thread roots is very difficult using visual and penetrant methods. Therefore, the purpose of this project was twofold: (1) to demonstrate an improved nondestructive inspection method for the spindle threads applicable to routine teardown maintenance, and (2) to determine the feasibility of performing safety-of-flight inspections on the spindle with only minimal disassembly.Recent projects funded by the Air Force have shown that the electric current perturbation (ECP) method is capable of detecting very small surface fatigue cracks in gas turbine engine disks1 and second layer defects in relatively thick structural wing sections.2 Based on these results, the ECP method was evaluated for its capability to inspect the spindle thread roots not only by scanning the outside diameter (crest of the threads), but also by scanning the hollow spindle bore under the threads and inspecting through the wall thickness for flight-critical cracks. With an ECP probe located on the crest of the threads, high sensitivity to very small defects in the thread roots was achieved and thumbnail shaped EDM slots as small as 0.53 mm long by 0.23 mm deep by 0.064 mm wide were detected. Inspection from the bore requires only that the rotary wing be removed so that a probe can be inserted into the spindle bore. Since this inspection is performed through the spindle wall, sensitivity is reduced and only larger defects are detectable. From the bore, detection of a thumbnail shaped EDM slot measuring 7.75 mm long by 2.21 mm deep by 0. 102 mm wide was successfully demonstrated.

The problem of detection and location of a small flaw inside a conducting cylinder using an eddy current coil coaxial with the cylinder has been addressed. The electric field at an arbitrary axial and radial position inside the conductor has been obtained from a previous solution of the boundary value problem. An expression for the change in complex impendance due to a small flaw located within a conducting body has been derived and is shown to be a function of the electric field at the position of the flaw. For the case of a degenerate point flaw this expression is further simplified by using just the value of the electric field at the position of the centroid of the flaw. The overall impendance is shown to be a function of the ratio of the radii of the loop and cylinder and of the conductivity of the material. The expression for the change in complex impedance has been bactored into two terms, one dependent on the axial location of the flaw, and the other on the depth of the flaw. The axial location of the flaw is seen to affect only the magnitude of the change in impedance; whereas the depth of the flaw is seen to affect both the magnitude and phase of the change in impedance. Plots of the complex change in impedance as a function of the axial location and depth of the flaw have been provided to illustrate its functional dependence on these parameters.

The electric current perturbation (ECP) method1–4 consists of inducing or injecting an electric current flow in the material to be examined and then detecting localized perturbations of the magnetic flux associated with current flow around material defects such as cracks or inclusions. Empirically, ECP data has shown strong correlations among certain signal features and crack size characteristics, and thus promises to be a useful method for quantitative NDE. To aid in the further development of the method, the objectives of the work reported in this paper are (1) to develop a mathematical model of the ECP flux distribution for a half-penny crack, (2) to determine the degree of validity of the model through comparisons with experimental data, and (3) to develop a detailed theory of sizing relationships for half-penny cracks.

The work described in this paper arises from a program for the detection and measurement of surface cracks in metals carried out at University College London. The instrument which was developed for the purpose, the Crack Microgauge, employs the acpd (alternating current potential difference) method. An alternating electric current at a frequency of 6 kHz is applied to the specimen, and the instrument measures the voltage between the probe terminals which are applied to the surface of the specimen. By examining the variation of the voltage readings with position on the surface and, in particular, the jump in readings obtained when the probe crosses the crack, the crack can be detected and features of its geometry deduced. The correlation between instrument readings and information about the crack geometry must be made by use of a theoretical model of the electromagnetic field produced in the crack neighborhood. The authors have been principally concerned in the study of this mathematical problem. In this paper we have attempted to bring together in summary form the most significant results arising from the studies on several different projects.

An exploratory investigation was conducted to evaluate the applicability of state-of-the-art eddy current nondestructive evaluation techniques to the characterization of applied and residual stresses in structural steels. Eddy current response versus stress measurements were developed for ASTM Type A533B and A471 steels under tensile, bending and residual stress loading conditions. A “shrink fit” specimen was used to establish applicability to residual stresses. Results show that an eddy current approach can be used to provide an accurate quantitative measure of surface stresses. The technique can also be used to map surface stress contours. Details of the procedure are described along with the test results and proposed applications. Recommendations for further work needed to optimize and expand the technique are included.

Acoustoelasticity is a promising method for the in situ analysis of both applied and residual stresses. The object of this investigation is to establish a technique for scanned shear wave measurements so as to determine the individual components of an inhomogeneous stress state and their directions. A computer-controlled scanning system with a dry contact rubber backed transducer has been developed which provides complete automation of scanning and data reduction.The theory of acoustoelasticity for anisotropic material has been developed using perturbation techniques. The experimental results on rolled aluminum plates confirm that, to a reasonable approximation, the effects of material anisotropy and stresses can be uncoupled and the needed stress information thus derived.

The theory for stress determination using acoustoelasticity is most frequently based on the evaluation of the motion of an infinitesimal plane wave propagating through an isotropic, elastic body which is subjected to a homogeneous deformation. The assumption of isotropy in this analysis allows the characterization of the acoustoelastic response to be carried out in terms of two second-order and three third-order elastic constants. Unfortunately, most structural materials do not behave isotropically, but instead have some degree of texture caused by the crystals aligning themselves in certain preferred orientations during the forming process. This paper examines the effect of texture on the acoustoelastic response of polycrystalline materials. In particular, the five second-order and nine third-order elastic constants of bodies exhibiting transverse isotropy are computed in terms of the elastic constants and orientation of the constituent crystals. Two methods of evaluating the constants are presented, the first being a Voigt type procedure in which the elastic stiffnesses are averaged for the chosen crystal orientation distribution, and the second being a Reuss type procedure in which the compliances are averaged. Acoustoelastic constants of the various waves in aluminum and copper are presented for the entire range of ideal textures in which all grains have the same direction cosines between the symmetry axis and the crystallographic axes.

The presence of a state of residual stress in a material can impair its structural quality by adversely affecting its elastic limit, yield point, etc.1 Most common nondestructive measurements of residual stress use x-ray techniques.2 However, these techniques determine only the surface residual stresses, while in many practical cases knowledge of the bulk residual stresses is desired. Ultrasonic methods3,4 appear most natural for measuring bulk residual stress but are used infrequently, in part because of difficulty in adequately measuring small effects and in part because of the absence of theoretical results treating the inhomogeneous nature of residual stress fields.

A new approach for using acoustic measurements to evaluate residual stresses in the presence of unknown material property variations is presented. It is shown that measurements using shear waves propagating along the normal to the surface of a plate do not provide sufficient information to separate the influences of stress and material property variations. To overcome this fundamental limitation, an alternative theory is developed that governs the propagation of shear waves polarized horizontally with respect to the surface of a plate (SH-waves), but propagating at oblique angles with respect to the surface normal. The question of separating the effects of residual stress and material properties on acoustic velocity is addressed in detail. In addition, a practical experimental procedure is developed that permits the evalution of the in-plane components of the principal stresses in a plate exhibiting an unknown inhomogeneous initial anistropy caused by material texture or microstructure. The procedure is then verified experimentally using an aluminum specimen with a known residual stress state, but unknown initial anisotropy.

The continuum theory of elastic wave propagation in deformed, anisotropic solids is reviewed with emphasis on those features which might be used to distinguish between stress induced changes in ultrasonic velocity and changes due to material anisotropy, such as would be produced by preferred grain orientation in a polycrystalline metal As noted by previous authors, one such feature is the difference in velocity of two shear waves, whose directions of propagation and polarization have been interchanged. In particular, when these directions fall along the symmetry axes of a rolled plate (assuming orthorhombic symmetry) and these are also the directions of principal stress, then the theory predicts that ρ(V₁₂2−V₂₁2) = T₁−T₂ where ρ is the density, V_ij is the velocity of a shear wave propagating along the i-axis and polarized along the j-axis, and T_i is a principal stress component. In addition to being independent of the degree of texture, this relationship has the advantage that no microstructurally dependent acoustoelastic coefficient is involved. The applicability of this prediction of continuum theory to heterogeneous engineering materials such as metal polycrystals is discussed using previously reported stress dependencies of ultrasonic velocities, and new experiments to answer some remaining questions are described. A possible configuration for using the effect to measure the value of a uniform stress in a plate of unknown texture is proposed.

The behavior of the temperature dependence of longitudinal ultrasonic velocity in type A533B steel in the presence of external as well as residual stresses has been investigated. In all measurements, the ultrasonic velocity in the vicinity of room temperature is found to vary linearly with temperature, and the slope of the linear relationship increases or decreases according to whether the stress is applied in tension or in compression respectively. The results also indicate that the temperature dependence of the velocity is a linear function of applied stress, and the slope of this linear relationship is the same for all specimens tested, and for both tensile and compressive stresses. These results are then used to obtain a relationship between the temperature dependence and residual stress in type A533B steel.

A correlation between the change in ultrasonic wave forms and applied strain in aluminum (6061-T6) has been obtained at high strain levels. Sophisticated signal processing techniques have indicated a complex interaction of the frequency components of a high frequency ultrasonic pulse as it passes through an aluminum tensile specimen. Strain induced microstructural changes in the aluminum attenuate the acoustic energy. One of the attenuation mechanisms is the formation of deformation induced cavities at precipitates and inclusions which scatter the ultrasonic energy. Measuring the signal attenuation at the appropriate frequencies determines the degree of deformation induced damage.

In a previous conference we presented a comparison of several different acoustic techniques to estimate residual stresses in complex situations.1 Of the several methods, the use of the attenuation of broad band pulses appeared to be better than the usual method of inferring strain from changes in the propagation velocity. The measurement of strain through changes in the velocity is effected through the equations 1$$\Delta V/V{\text{ }} = {\text{ }}f({\varepsilon _{ij}})$$2$$\Delta t/t = \Delta d/d - \Delta V/V$$ where ε_ij = strain tensor, t = time for the wave to traverse the specimen, d = specimen thickness, V = wave velocity.

It is known that the energy-release rates associated with translation, rotation, and self-similar expansion of cavities or cracks in solids are expressed by path-independent integrals J, L and M, respectively. These integrals are of interest to NDE in that they can be used to characterize nondestructively defects in solids.It is shown that these integrals for a crack may be evaluated by first considering an ellipse and then performing a limiting process. This obviates dealing with singularities at crack tips and holds promise for a more efficient numerical method in complicated cases, since modeling of singularities is always associated with difficulties and uncertainties.

An ultrasonic technique has been developed that enables measurement of grain size in uranium without metallographic preparation or destructive analysis. Pitch-catch ultrasonic analysis using transverse (shear) waves was conducted in the determination of grain size in wrought uranium parts.

In heat-treatable aluminum alloys it has long been accepted that decreased values of strength were accompanied by increases in electrical conductivity (C). In quality or processing control and trouble-shooting situations this has been useful for finding anomalies in or among aluminum alloy maill products. But the regression was always found as a wide scatterband where conductivity could not give a narrow range of possible strengths.It was discovered for several alloys and quantified for 2219, that the scatterband formed by data from several lots and sources actually could be divided into groups with different histories. When specimens produced by created combinations of quenching-time and aging-time had their Hardness (H) vs Conductivity plotted on a H vs C format a fan-like dispersion of coordinated points was seen. Drawing locuses thru like times divided this fan into age-time and quench-time grids. Any particular C-H coordinate in this envelope then was seen as identifying the thermal history of the piece with that of C-H value. It was also found that progress in one direction on this format marked out the increase in the 2219 hardening precipitates θ″ and θ′. Progress in the other direction marked out the increase in the softening precipitate θ. So that even the particular metallurgical status could be found from the C-H coordinate of the specimen.This work taught that the large C-H variations seen in accumulations of data most often represented variations in the material itself, not in measurement systems. The work also taught that variation in production material was tracible mainly to variation in quench quench times. Should such variations be reduced the standard deviation of strength would be reduced and higher design strengths could be assigned to the alloy. In practical situations increases in design strengths (which conversely means reductions in assembly weight) are seen at 12%. The NDE measurements can serve this end by identifying and certifying grades of material before pieces are put into service. This strategy involves avoiding that apparently sound material which will fail early in its service life.

There are a limited number of nondestructive evaluation techniques available for field inspection of large composite structures and practically no viable techniques for in-service inspection. With this in mind, an innovative Damage Assessment System is proposed which is based on a concept of using an optical fiber mesh, implanted into the body of a fiber reinforced composite structure. Such a mesh would become an integral part of the structure during the course of its fabrication. The selection of the mesh fibers would be predicated on their strain to failure characteristics and strain compatibility with the base, composite reinforcing fibers. This optical system will be capable of locating damage, assessing severity and monitoring damage growth. A successful implementation of the total Damage Assessment System would involve the interaction of the optical fiber mesh with an adequately designed interrogative electronic package. This paper focuses on the former aspect of the total system. It will address some recent experimental work showing the practicality of the concept in assessing various modes of failure due to impact of composite plates, optical fiber selection, location and spacing of fibers, as well as the utility of the system for damage assessment in large, complex composite structures.

It is well known that nonmetallic inclusions can adversely affect the metallurgical properties of engineering alloys. For critical components such as aircraft engine gears and bearings it is important then to quantitatively assess the severity of the inclusion content in the alloy material before performing costly manufacturing operations. This poster paper will describe the operation and the initial results of a computer controlled steel cleanliness inspection system.The severity of the inclusion content is determined by this system through a statistical analysis of the internally reflected ultrasonic indications from the alloy material as a transducer is scanned in a raster fashion. These indications are sorted by a computer with respect to signal amplitude and location in the material. The initial results from a sampling of known clean and dirty steel billets show that the amplitude distribution function for the ultrasonic indications from clean material is nearly a normal distribution while the same distribution function for material with high nonmetallic inclusion content is highly skewed. Work has just started which will determine how well the ultrasonic cleanliness data from bar stock material will correlate with the mechanical performance of helicopter gears manufactured from this stock. The initial results will be presented.

Acoustical holography is rapidly approaching commercial status with applications to nondestructive evaluation, underwater imaging, and underground pipe location being pursued. In this paper we review the techniques involved, show some experimental results, and describe the latest commercial system.

A number of alloys exhibit a phenomenon which has come to be known as ‘shape memory’. Without exception these alloys undergo a rather special type of martensitic transformation which results from one or another of the elastic constants1–7 weakening with changing temperature to the point that at the transformation temperature the transforming phase becomes both mechanically and thermodynamically unstable. ‘Shape memory’ is exhibited in the following way. A specimen at a temperature above the martensitic transformation is shaped to some desired form. The specimen is then cooled below its transformation temperature and deformed. Subsequent heating of the specimen through the transformation temperature causes reversion to the originally fabricated shape.

X-ray analysis has traditionally been wavelength dispersive, manpower intensive, and an analytical laboratory tool employing massive and costly x-ray generators and diffraction spectrometers. Following the development of the lithium drifted silicon and germanium [Si(Li) and Ge(Li)] semiconductor radiation detectors in the mid 1960’s, and more recently the high purity germanium detector [HPGe], the analytical technique of energy dispersive x-ray fluorescence analysis (XRFA)1 matured during the 1970’s.

Advances in elastic wave scattering and inversion techniques have shown that advances in transducer technology are needed in order to fully exploit them. This is particularly true in the case of flaw sizing algorithms in which it has been demonstrated that a need exists for transducers with specified bandwidth characteristics. As a part of an effort to develop a composite, multiviewing ultrasonic transducer for flaw characterization, a computational tool has been developed which provides a convenient way to select driver pulse shapes and transducer characteristics which optimize this property. The purpose of this paper is to discuss this computational tool.

In order to provide a capability for performing advanced signal processing on ultrasonic and acoustic emission signals at speeds that are sufficient for practical applications, a high speed Digital Ultrasonic Instrument (DUI) has been developed. The DUI performs its processing entirely digitally and therefore can do the phase- and frequency-sensitive processing which is necessary in many advanced NDE techniques. Its speed of computations is sufficient to handle pulse repetition frequencies (PRFs) of several hundred Hertz. Three applications of the DUI are described, one each in the areas of flaw detection, flaw characterization and acoustic emission source characterization. The first application is improved near surface flaw detection by the use of subtraction of front surface echoes. The second application is a real-time operator-interactive method for correcting a flaw signal to remove system response and interface signals and thereby prepare the flaw signal for flaw characterization techniques such as the Born Inversion. The third application is the automatic identification of sources of acoustic emission in a fastener-hole geometry.

We present a simple analytical method for predicting the eddy current signal (ΔZ) produced by a surface flaw of known dimensions, when interrogated by a probe with spatially varying magnetic field. The model is easily parameterized, and we use it to construct inversion schemes which can extract overall flaw dimensions from multiposition, multifrequency measurements. Our method is a type of Born approximation, in which we assume that the probe’s magnetic field at the mouth of the flaw can be used as a boundary condition on the electromagnetic field solutions inside the flaw. To simplify the calculation we have chosen a “rectangular” 3-dimensional flaw geometry for our model. We describe experimental measurements made with a new broadband probe on a variety of flaws. This probe operates in a frequency range of 200 kHz to 20 MHz and was designed to make the multifrequency measurements necessary for inversion purposes. Since inversion requires knowledge of the probe’s magnetic field shape, we describe experimental methods which determine the interrogating field geometry for any eddy current probe.

This paper describes progress in eddy-current methods to identify and size defects. An eddy-current imaging method is used to generate data for analysis of small defects. Characterization of the defect is derived from this information and the validity of the derivation is determined by study of artificial and natural defects. Initial results have a theoretical foundation but more advanced analysis is needed.

The important steps in eddy current testing for defects in test objects are: detection, classification and sizing. This paper concentrates especially on the last two items.

In this report we describe an approach to the reconstruction of flaws, not merely their detection. This will give us the ability to obtain much more information about the nature of the flaw. By “flaw” we mean virtually any departure of the medium from a standard condition, which is known a priori, such as may be produced not only by a crack but also by conductivity in homogeneities produced by stresses, magnetite build-up, etc. Our approach is very much in the spirit of contemporary work in inverse methods in electromagnetics [1–3] and electromagnetic-geophysical prospecting [4–11].

This paper describes the continuation of the work1,2 on synthetic aperture focusing techniques (SAFT) and ultrasonic imaging as well as new work on the measurement of the amplitude and phase as a function of frequency and aperture position.

Acoustic images offer good spatial resolution in two dimensions, but do not give information on the third spatial dimension. In this paper we will describe a new method of measuring flaw depth profiles which makes use of the corner reflections from a surface breaking flaw. A calibration experiment is described, showing that at the present time this method of flaw depth determination can be calibrated to an accuracy of about 20%.

The imaging of interior planes in a solid object involves difficulties with spherical aberration, with the elimination of the front surface echo, and with the need in some cases to use signal processing techniques in order to enhance the signal-to-noise ratio. This paper shows that suitably designed spherical lenses can be used to minimize the spherical aberration. The elimination of the front surface echo requires careful time-gating and very short, broadband pulses. It is shown by using extended chirp pulses, the signal-to-noise ratio for subsurface objects can be very greatly improved. A number of examples of the use of these techniques for NDE are presented.

Ultrasonic imaging is of use in a number of important areas, including nondestructive testing and medicine. The field has grown considerably in the past ten years and remains an active and growing area of research. For nondestructive evaluation of materials (NDE), the aim is to provide means for obtaining estimates of the size, shape and orientation of flaws in sufficiently a quantitative manner so that failure of mechanical structural parts can be predicted.1,2 As an imaging technique, ultrasonic imaging is appropriate whenever the medium considered is opaque to other sources of radiation, such as optical radiation. It is versatile and convenient and may be used in both the transmission and reflection, active and passive modes. The technique is also versatile in another important sense. Use of (nearly) spatial coherent radiation sources and of linear detectors results in the availability of a wide variety of digital signal processing procedures for enhancing signals, improving lateral resolution, range gating and so on.3

We have developed a technique for imaging variation in the surface characteristics of a sample by measuring the local perturbation of the Rayleigh wave velocity. A 50 MHz acoustic microscope operated out-of-focus is excited with a very short tone burst so that the on-axis longitudinal and off-axis Rayleigh reflection pulses are temporally separated. The relative phase between these two signals is measured using a synchronous detection scheme. This technique has a potential sensitivity of 10 ppm. We have demonstrated experimentally that we can detect a 240 Å thick film of indium deposited on glass which corresponds to a velocity perturbation of 0.18%.

Reliable and repeatable flaw detection and accurate flaw characterization are critical to reliable evaluation of component serviceability. Digital filtering and signal processing techniques can be used in real time to reliably and repeatedly detect flaws even at low signal-to-noise ratios. The same techniques provide near optimum flaw resolution and dimensional precision using standard unfocused transducers. This paper describes these techniques and demonstrates their effectiveness with sample data. The data was processed and collected by an ultra-high-speed general purpose programmable signal processor which is used to detect, locate, and generate images of the component with flaw locations.

In this paper we describe a contacting shear wave array which we have used to characterize flaws in aluminum samples. By using shear waves instead of longitudinal waves, this array has doubled the resolution of our real-time synthetic aperture imaging system. The images we have obtained are significantly better than those obtained with longitudinal or surface wave arrays.

A computer based acoustic model has been developed for the NDE of multilayered structures [1]. The model is applicable for normal or off-normal incident excitation, with receiver in pulse-echo, pitch-catch or in array mode. The model can simulate the observed signal for arbitrary frequency response of the transmitting and receiving transducer. In addition, the model considers attenuation and mode conversion effects in each layer in predicting the ultrasonic response.While earlier works [2,3,4] considered normal incidence and non attenuative media, the current model’s capabilities have been expanded to include off-normal incident angles and the attendant mode conversions created with this inspection configuration. Another feature included in this model is provision for attenuation within each layer. It can be modeled as constant over all frequencies or as a frequency dependent quantity, such as a constant “0” model. Common materials such as aluminum, stainless steel, rubber and biological tissues exhibit this property.The paper will focus on model development and discuss results obtained for a bronze-rubber multilayer structure.

The interpretation of ultrasonic signals in the inspection for flaws in bonded regions of multilayered specimens is difficult because of signal energy loss due to material attenuation, undesired reverberations within certain layers, and overlapping responses from different interfaces because of finite transducer bandwidth. The flaws are usually air gaps, lack of adhesion, and porosity within the bonding agent.While ultrasonic NDE signal interpretation can always be improved with appropriate instrumentation — broader bandwidth transducers for increased resolution, for example — signal processing allows for further enhancement using a digital computer. Material attenuation can be compensated for by using digital filters that preferentially allow the higher frequency components in the ultrasonic signal, similar to “preemphasis” filters used in communications. Deconvolution of signal response broadens the effective bandwidth of the transducer and can be used to minimize dominant reverberations within a layer. The use of the cepstrum — a relatively new signal processing method — allows for the separation of overlapped responses which are visually difficult to separate.The use of these signal processing methods is demonstrated for inspecting bronze-rubber structures with a small layer of epoxy in between.

In the fabrication of components with thermosetting polymers, an essential step is the proper optimization of the cure cycle. In the early phases of the cure the polymer must have the appropriate flow properties to assure the proper wetting, spreading, and forming. It must then harden without excessive build up of residual stresses and flaws or the loss of adhesion at any interfaces that are present. To complicate the situation further, many applications such as composites, adhesives, paints, and protective coatings involve thin films whose cure behavior is significantly different than that for bulk samples. To help address this problem, an ultrasonic shear wave propagation technique has been developed. It is a laboratory device that measures the dynamic shear properties of a thin film in a way which is both nondestructive and nonperturbing to any chemical reactions that may be occurring. The applicability of this test method was demonstrated with cure studies on two model systems: one based on tung oil the other based on an epoxy resin.

Optical interferometric techniques using single mode optical fiber waveguide embedded in composites and other layered materials have been used to measure one- and two-dimensional stress distributions and acoustic emission caused by applied point source loads. By interferometrically comparing the phases of coherent optical signals propagated through an embedded sample fiber and a bypass reference fiber, a signal proportional to the instantaneous strain integrated along the embedded length of the sample fiber has been detected. System calibration has been obtained by applying a one-dimensional dc strain field to a cantilever beam containing the fiber. Using this calibrated system, an array of fibers attached to a 15cm × 15cm x 0.3cm plate simply supported at the corners and subjected to point loading on the surface has been used to quantitatively determine the two-dimensional dc stress field in the plate. Finally, the calibrated ac response of the interferometer to acoustic emission events in a composite panel has been demonstrated. Potential applications are discussed.

A computerized ultrasonic scanning bridge has been developed for the scanning and imaging of defects in structures. The raster scaning pattern can be implemented with any pair of the six available axes. The digitized ultrasonic signal can be imaged using a Peritek Graphic system. Details of the ultrasonic scanning bridge and imaging system will be reviewed.Examples of the evaluation of a graphite epoxy component will be reviewed. The scanning of the composite part requires the use of the two angulation axis for the raster scanning. The correlation of the ultrasonic inspection with failure pressure of the graphite epoxy component will be presented.

A new sensor is presented for detecting surface acoustic waves. The sensor constructed, using single mode fiber components, is small and rugged and has better sensitivity (.0003 Å) than has been reported for other SAW sensors. Optical reflection changes encountered while scanning surfaces can be divided out, making the probe applicable to rough samples of practical interest. Normal surface displacements near a defect (100 μm deep crack) have been measured in both long and short wavelength acoustic regimes.

A program is described which employs lasers for ultrasonic NDE. A high-power laser is used to generate a brief sound pulse in the test specimen. A second low-power laser then measures the response of the specimen to that sound pulse.The response of the specimen is measured by a “Laser Vibrometer.” This is a novel type of heterodyne interferometer which focuses a Helium-Neon laser beam onto the surface of the specimen and measures its displacement. Displacements as small as 2×10-12 meters on a 0.15 sec averaging time can be detected and also displacements of 1.5×l0-9 meters on a 10-MHz bandwidth. The Laser Vibrometer has a well defined frequency response and does not introduce distortion.The sound generating laser is either a pulsed carbon dioxide TEA laser or a YAG laser. The peak power exceeds 10 M watt. Two mechanisms for generating the sound are discussed. The thermoelastic mechanism relies on the thermal expansion of the surface, causing it to move. The reaction to this causes a pressure pulse in the specimen. Another mechanism allows a small amount of the surface to be ablated and the reaction to this causes a substantial pressure pulse in the specimen.Both laser beams can be scanned over the surface of the specimen by a microprocessor controlled mirror. The microprocessor generates a raster scan of arbitrary size, number of lines, step size and speed.Eventually this technique will allow the inspection of complex specimens without direct contact. This will eliminate the tedium and contact reliability problems associated with conventional piezo-ceramic NDE.

If a nondestructive evaluation system is designed to detect the presence or absence of a flaw in a material, typically one transducer may be sufficient. If, however, a characterization of the flaw is desired, then an array of transducers is in most cases required. Besides the capability of two and three dimensional imaging, array data has the advantages of increased resolution, improved signal-to-noise ratio after preprocessing and sharper focusing.In any NDE system, the acquisition of data is only one step towards the final objective of flaw characterization. The other step is that of processing the data in order to extract the desired information. In this paper, we consider one signal processing aspect of data obtained by a linear array of transducers. Each element on the array normally operates as transmitter and receiver simultaneously, and the data is collected by exciting one transducer at a time. The measured signals, after suitable time shifting for alignment, are summed in order to focus (or beamsteer) the array at a specific point. The resolution of this summing process depends on the side lobes of the array reject response, and this in turn depends on the number of elements and spacing between elements on the array.While summing is the simplest signal processing procedure to perform, it is however, as far as beamforming is concerned, not the most effective. The side lobe levels decrease as the number of elements N increases, and this has a lower bound of about -14 dB as N →∞. In this paper, we introduce an additional processing step with specially designed optimum filters before summing. The design methodology for these filters will be discussed in detail, and it will be shown that these filters have a superior frequency reject response which becomes more apparent if the array has a small number of elements.

The use of lasers to generate thermoelastic waves has received considerable attention since the publication of the theoretical work of White.1 The author derives equations for the propagation of longitudinal elastic waves produced by surface heating of a semi-infinite medium with harmonic thermal (laser) excitation. The importance of the boundary conditions in determining the magnitude of the resulting elastic waves is described. Two conditions are imposed and the resulting amplitudes derived: (1) Laser absorption at x = 0 with elastic amplitude u=0 at x = 0 for all times t ≥ 0 (perfectly clamped or constrained surface) and (2) for the stress σ = 0 at x = 0 for t ≥ 0 (free surface). Propagation of the elastic wave is in the x direction.

The piezoelectric transducer is the electromechanical conversion element in NDE. Existing transducers of this type are found to impose limits on the improvement in system performance and, at present, there is a lack of standards as required for calibration procedures. Physical “standards” based on equivalent circuits have been produced which model individual transducers, single transducers in pulse-echo mode and two-transducer configurations. The performance of these “standards” is shown to be in good agreement with that given by real transducers, and subject to less variability.