Inspection Speed Research Articles

Performances of remote digital processing in automated non-destructive evaluation

Recent advances in transducers, electronic techniques, computer architectures and digital signal processing provide substantial improvements in automated non-destructive evaluation (ANDE) techniques. ANDE must fulfil very strict specifications, especially related to reliability (i.e., the quality of images, repetitiveness of acquisitions, etc.) and to the inspection speed. Multi-channel operation and digital signal processing provide a good solution to overcome strict NDE demands. In this paper, an approach based on a multi-channel remote system (RSENDAS 1) is presented, in which both signal conditioning and digital processing are done locally. The remote system performs complex algorithms for noise suppression, resolution improvement and smart data reduction at a sustained global processing rate of 10M samples per second. The central computer operates in parallel to attend to other jobs such as movement control, image display, etc. Several features of this solution related to noise immunity, system flexibility, system parallelism and other advantages are analysed in the paper.

An evaluation of economic and performance feedback in an inspection task with explicit economic consequences

Feedback has been consistently shown to improve inspection speed and accuracy, provided it is given in a timely and appropriate manner. Traditionally, this feedback has been with regard to speed and accuracy performance. Nevertheless, when there are explicit economic consequences associated with the inspection process, economic feedback is also an alternative. This study compares performance feedback with economic feedback in a realistic simulation of an aircraft inspection task. Subjects provided with performance feedback showed overall improvement in speed and accuracy. Subjects provided with economic feedback not only demonstrated similar improvement, but also approached economically optimal trade-offs between speed and accuracy. Relevance to industry The results of this study have direct implications on developing training strategies for improving industrial inspection performance. In particular, in the context of an aircraft inspection task, it was demonstrated that an appropriate training program could both improve accuracy (and therefore safety) and reduce the costs associated with inspection. These results are generalizable to other inspection applications in a variety of industries, such as electronics, textiles, manufacturing, and agriculture.

Transmit-receive eddy current probes for circumferential cracks in heat exchanger tubes : Obrutsky, L.S.; Cecco, V.S.; Sullivan, S.P.; Humphrey, D. Materials Evaluation, Vol. 54, No. 1, pp. 93–98 (Jan. 1996)

Conventional eddy current bobbin probes are known to be ineffectual in detecting circumferential cracks in tubing. Multi-pancake and/or rotating pancake probes are required to detect circumferential cracks. It has recently been demonstrated in CANDU steam generator tubes, with deformation, ferromagnetic deposits, and copper deposits, that multi-channel probes with transmit-receive (T/R) coils are superior to those using surface impedance coils. Unlike rotating probes, the design of a new probe denoted as C3 permits fast, single-scan inspection of a full-length tube at inspection speeds comparable to conventional bobbin probes. Since 1992, the probe has been used routinely for steam generator in-service inspection at 4 CANDU plants. Defective tubes have been plugged and the units returned to service, and they continue to operate without leaks. This paper describes the basic features of T/R surface probes. Two-dimensional voltage diagrams showing computer-derived probe response to frequency, lift-off, carbon steel supports, magnetite deposits and copper deposits are presented and compared with corresponding signals from impedance coils. Theoretical and experimental results show that T/R probes are able to detect defects in the presence of variable lift-off (due to tube-wall deformation) with ten times the signal-to-noise ratio as that exhibited by comparable pancake-type impedance probes. In addition, T/R more » probes are less sensitive to magnetite deposits, and possess good phase discrimination to internal defects. « less

Application of x-ray instrumentation in medicine: discrimination of neoplasms in radiographs by digital image processing

The application of digital image processing to mass screening for breast cancer is demonstrated in the context of adaptive spatial filtering. Experimental results for feature extraction by texture analysis and grey level thresholding reveal there is significant potential for computer-aided mass screening as an adjunct to radiological examination. Digital image processing can improve the visibility of image features, such as neoplasms or tumours, providing also speed and automation of inspection, whilst saving the radiologist time which is otherwise spent on checking clear samples (which represent the majority of screening cases).Zusammenfassung. Die Anwendung ziffernmäßiger Bildbearbeitung zur Massenabblendung für Brustkrebs wird im Zusammenhang mit räumlichem Filtern demonstriert. Experimentresultate für Besonderheiten-Extraktion im Falle von Gewebeanalyse und Grau-Niveaubegrenzung enthüllen, daß hier eine bedeutende Möglichkeit für computer-unterstützte Massenabblendung als ein Zusatz zur radiologischen Untersuchung besteht. Ziffernmäßige Bildbearbeitung kann die Sichtbarkeit von Bildmerkmalen, wie Neoplasmen und Tumoren, verbessern, außerdem Geschwindigkeit und Automation der Inspektion bietend, während dem Radiologen Zeit ersparend, die andernfalls zur Überprüfung klarer Muster verwendet würde (was für die Mehrzahl der Abschirmfälle zutrifft).

CT Multiscan: Using Small Area Detectors to Image Large Components

The small size and dynamic range of the best two-dimensional X-ray detectors are impediments to the use of three-dimensional X-ray computed tomography (3D-XRCT) for 100 percent inspection of large ceramic components. The most common industrial 3D-XRCT systems use a “rotate-only” geometry in which the X-ray source and the area detector remain stationary while the component placed between them is rotated through 360 deg. This configuration offers the highest inspection speed and the best utilization of X-ray dose, but requires that the component be small enough to fit within the X-ray/detector “cone.” Also, if the object is very dense, the ratio of an unattenuated X-ray signal to that through the longest path in the component may exceed the dynamic range of the detector. To some extent, both of these disadvantages can be overcome by using “Multiscan CT,” i.e., scanning small overlapping regions of a large component separately while maximizing the X-ray dose to each. The over lapping scans can then be combined seamlessly into a single scan with optimal contrast.

Interactive three dimensional inspection of patient setup in radiation therapy using digital portal images and computed tomography data

Presently, the majority of clinical tools to quantify deviations in patient setup during external beam radiotherapy is based on two-dimensional (2D) analysis of portal images. The purpose of this study is to develop a tool for the inspection of the patient setup in three dimensions (3D) and to validate its clinical advantage over methods based on 2D analysis in the presence of out-of-plane rotations. We developed an interactive procedure to quantify the setup deviation of the patient in 3D. The procedure is based on fast computation of digitally reconstructed radiographs (DRRs) in two beam directions and comparison of these DRRs with corresponding portal images. The potential of the tool is demonstrated on three selected cases of prostate and parotid gland treatment where conventional 2D analysis produced inconsistent results. The measurements from 3D analysis are compared with those obtained from the 2D analysis. Despite application of an immobilization cast, two investigated parotid gland setups showed rotational deviations in 3D up to 3 degrees. Two-dimensional analysis of these deviations produced inconsistent results. Analysis of the selected prostate setup in 3D showed a rotational deviation of 7 degrees around the left-right axis, possibly causing displacement of the seminal vesicles toward the borders of the conformal boost fields. Using 2D analysis, this out-of-plane rotation was misinterpreted as a translation resulting in the failure to trigger the decision protocol to correct the setup after the first fraction. Using the 3D patient setup analysis procedure, an accuracy of the order of 1 mm and 1 degree (SD) could be obtained. The computation time of the interactive DRRs is of the order of 1 s on a 60 MHz PC. The complete interactive 3D analysis requires about 10 min. Quantification of the patient setup in 3D provides essential additional information in cases where conventional 2D analysis is inconsistent, e.g., in the presence of out-of-plane rotations or geometrical degeneracies. The speed and accuracy of the interactive 3D patient setup inspection are acceptable for use in offline clinical studies and analysis of problem cases.

Optical inspection system for the inner surface of a pipe using detection of circular images projected by a laser source

An optical inspection system for the inner wall of a pipe has been developed. This system uses the projection of a circular pattern formed by a conically reflecting mirror on the inner wall, and detection of circular images from the illuminated circumferential surface. Irregularities of the images due to corrosion or accumulation of deposits on the inner surface are inspected by this system. Inspection of the shape of the inner wall has been impossible by current methods. The experimental system reported in this paper has the following practical specifications: resolution of detection on ±0.1 mm, and inspection speed of 30 mm/s at an interval of 1 mm along the pipe. The system is now being used for practical inspection.

THE OPTIMUM SPEED OF VISUAL INSPECTION USING A RANDOM SEARCH STRATEGY

Abstract Choosing a standard speed for visual inspection is difficult using standard work measurement techniques. This paper uses a different approach, taking a model of human functioning in inspection and deriving an optimum speed of working so as to balance the cost of time against the cost of errors. Random search is used as the model and both self-paced and externally-paced conditions are derived for multiple fault types. Optimization using a Golden Section search is used to provide optimum inspection speed.

Improved eddy-current inspection for steam generator tubing : Proceedings of the 17th Water Reactor Safety Information Meeting, Rockville, Maryland, (United States), 23–25 Oct. 1989. pp. 285–299, The US Nuclear Regulatory Commission, NUREG/CP-0105, Vol. 3, (Mar. 1990)

A major limitation of eddy-current inspection of steam generator tubing is that small flaw signals can be masked by the effects of benign variables, such as tube supports. To identify the critical flaw properties accurately and reliably in the presence of signals caused by these other property variations, we must have enough information to distinguish the flaw signals from the extraneous ones. Therefore, we developed instrumentation to measure both the amplitude and the phase of the eddy-current signal at several frequencies, as well as computer equipment to process the data quickly and reliably. This need to detect small flaws in the presence perturbing property variations has also required the development of more sensitive and more complicated probes, such as pancake and reflection probes. These smaller coils can detect much smaller flaws and are less sensitive to artifacts outside the tube, such as tube supports, magnetite, or copper. By being pressed against the tube wall, they also avoid liftoff effects. To increase the inspection speed an array of these small coils has been constructed and tested. Finally, new and more complicated tube standard were constructed to include the range of property variations. 19 figs.

An improvement of Compton scatter imaging with wide aperture detectors a Monte Carlo study

A novel NDT (nondestructive testing) imaging technique which is based on the measurement of Compton-scattered photons and uses wide-aperture detectors has been developed. With this technique, the inspection speed can be improved because the scanning activities can be reduced by one degree of freedom. Using wide-aperture detectors serves to simplify the design of measurement systems, and the data collection time at each step in the scanning process can be reduced due to the removal of the usual efficiency-reducing focusing collimator. To study the imaging process of the novel NDT system, a number of Monte Carlo calculations have been performed. The Monte Carlo code has been validated by comparison to the widely accepted general-purpose MCNP (Monte Carlo neutron-photon) code and has been found to run over ten times faster than the MCNP code for this particular research. As a result, the NDT system has been designed to deal, in particular, with the pipe wall thinning problems widely observed in nuclear power plants and some fossil power plants. In this application, the contour of a pipe inner surface can be reconstructed in a very straightforward manner, and the scanning activities can be reduced from three dimensions to two dimensions. >

3-D OBJECT INSPECTION FROM MULTIPLE 2-D CAMERA VIEWS

A new approach to 3-D object inspection using 2-D camera views is proposed. With input objects put on a turntable for inspection, a top view of the object is first taken by a top camera right above the table. The object shape boundary points are extracted, and the centroid as well as the principal axis of the shape are computed for top-view registration with the model shape, using the generalized Hough transform. The distance-weighted correlation is then used as the similarity measure for shape matching. Input objects with low similarity values are rejected. Non-rejected objects are inspected further from lateral views, using a lateral camera for image taking and the turntable for object rotation. Experimental results with a high inspection rate and a fast inspection speed show the feasibility of the proposed approach.

Efficient data acquisition in automated ultrasonic inspection

The large volumes of data resulting from automated ultrasonic inspections illustrate the difficulties of efficient data acquisition and analysis. Consequently, all the available inspection systems impose a compromise between the quantity and the quality of data that they collect and the maximum possible speed of inspection. The paper introduces the problems of ultrasonic data acquisition and assesses several techniques so far employed in inspection systems. A further technique is then described in which only the minimum data required by ultrasonics specialists for defect detection, sizing and characterisation are gathered in a single high-speed scan. This is a technique which is presently being investigated so that the plotting and analysis may be carried out rapidly by discarding data which would otherwise obstruct the interpretation process. The technique requires intelligent discrimination of the features contained in the baseband envelope of the ultrasonic signal. A high-speed signal preprocessor which will implement this technique is under development and is known as the digital ultrasonic signal conditioner (DUSC); its design and applications are described

Speed-accuracy trade-off and time of day

Data were collected at six times of day from a total of 48 subjects working at a self-paced version of a visual, signal identification task. Speed of inspection and confidence of report both tended to increase over the day, while signal detectability ( d′) declined. The response criterion (ln β) remained invariant with time of day. To determine whether variations in speed were being offset by those in accuracy, the efficiency indices d′/ T and d′/√ T were derived. Analyses showed that efficiency was stable over test times: 8:00, 11:00, 17:00, 20:00, indicating the operation of a speed-accuracy trade-off; but efficiency dropped significantly at 14:00 (possibly a post-lunch dip) and again at 23:00, at the end of the day. The shift in the trading relation between speed and accuracy was in the direction of faster but less careful performance as the day progressed; the possibility that this is due to accumulated fatigue rather than to time of day per se, is discussed.

Variable inspection time as an indicator of cognitive reading behaviors

Previous attempts to relate inspection speed to text relevance have produced ambiguous results. Two experiments were conducted to demonstrate that the inconsistencies previously observed were probably due to the experimental manipulations employed to establish relevant and irrelevant text material. While specific goals or questions were shown to produce variable inspection speeds, a general purpose in reading did not. In both cases, text information related to the reading goals was more likely to be retained. Results are explained in terms of the storage demands imposed by each category of reading goal.

Aging and Pretraining in Industrial Inspection

Learning-to-learn skills were hypothesized as reasons for poorer learning performance by older workers. To test this on a complex, simulated industrial inspection task, 84 subjects in three age groups were trained for the task. Half received restraining on organization of memory material and on size discrimination using tasks not directly related to the inspection task. Inspection speed decreased with age while errors increased. Pretraining reduced size discrimination errors and decision errors, having a larger effect than the age decrement and a consistent effect across age groups. There was no interaction between pretraining and the type (active or passive) of task training scheme used to train for the inspection task. It is concluded that intervention by pretraining and by the use of active training can improve the employability of older workers by removing the perceived barrier of their trainability.

Performance Effects of Variables in Dynamic Visual Inspection

Abstract Three task variables of dynamic visual inspection were examined in laboratory experiments: (V) conveyor-belt velocities, (T) target exposure time and (P) the viewing position of the inspector. Measurements were obtained on both error types in inspection of equally-spaced targets for 36 experimental condition. Task variables V and T provided significant main effects on inspection accuracy and variables T and P had an interacting effect. The probability of correctly identifying a target was found to be describable as l-2.58e−a where a = 10.587T- 0.0.75/V-0.012/[T+ P2] and the coefficient of determination, R2 = 0.88. Inspector's eye motions were recorded and gross eye-motion patterns were identified but none appeared to be directly associated with inspection accuracy. Shifts in the inspection error criterion were observed and described for changes in these task variables. Both inspection speed and accuracy are combined in a cost model. This study shows that different conveyor speeds have better economics with different error costs but that shorter exposure times were never economically better.

Studies of Visual Inspection

This paper describes the results of four experiments in a series aimed at understanding and improving visual inspection in general and of small integrated circuits (i.e. 'chips') in particular. Stimuli consisted of chips that, although electrically sound, contained visual anomalies. The first experiment found that the modal duration of eye fixations of trained inspectors was about 200 msec. The most accurate inspectors made the fewest eye fixations and were the fastest. The second experiment evaluated the performance of inspectors at one of the many sequential stages of chip inspection and found that 23% of the chips containing anomalies were accepted, whereas only 2% of the chips without anomalies were rejected. When the same chip was judged more than once by an individual inspector the consistency of her judgment was very high whereas the consistency between inspectors was somewhat less. The third experiment showed that variation by a factor of six in inspection speed led to variation of less than a factor of two in inspection accuracy. The fourth experiment showed that inspection via a ground glass screen is only a little worse than the usual method of looking through a binocular microscope. This was true even though the inspectors had no previous experience with the screen.

The effect of speed of working on industrial inspection accuracy

One of the major determinants of inspection accuracy is speed of working. From a knowledge of the speed-accuracy trade off in inspection, it is possible to predict the optimum economic speed of inspection and hence to determine manning levels for inspection. This paper describes seven studies and shows that similar results were obtained irrespective of the type of product examined. Theoretical models of inspection are described which partially account for these effects. A method for the determination of optimum speed in economic terms is given.

Automatic rotary probe system for the ultrasonic inspection of steel tube

Manual systems for tube inspection are unable to keep pace with high-speed tube manufacturing processes, and only provide an imprecise test. A range of high speed rotating ultrasonic probe systems and static probe assemblies has been developed for automatic quality control and acceptance testing of tubes. These systems afford automatic inspection of tubes in the diameter range 0.5–44in(12.7mm–12.118m), seamless or welded tubes, at inspection speeds up to 200ft/ min(60.96m/min). The systems incorporate several novel features in the ultrasonic and electronic design. One of these systems is described in detail.

Ultrasonic Imaging Systems for Nondestructive Testing

Methods for detecting ultrasonic images are briefly reviewed, and discussed in light of their potential application to nondestructive test problems. In a typical nondestructive test system involving a water immersion tank, the ultrasonic image detector should respond to an ultrasonic intensity of 10−2 W/cm2 or less. This requirement eliminates from consideration number of detection methods involving photographic, chemical, or thermal mechanisms. Of the remaining image detection systems, only a few have been applied to nondestructive test problems. These include the surface relief technique, the Pohlman cell, and scanning methods involving piezoelectric transducers. The latter method has been most used for image applications, with mechanical movement of the transducer(s) in a C scan method. Electron scan of a single piezoelectric detector to provide a television presentation is coming into wider use. A description of one such system is given. For a thin, flat inspection object, the system offers good inspection speed, resolution, and sensitivity to inhomogeneities. Thick, irregularly shaped inspection objects present problems because lens systems are involved. Lenses for imaging are not well developed. Also, the use of a lens tends to slow the inspection of a thick object because one must examine the object by focusing at different depths. Pulsed television systems and holographic techniques offer some promise in that area. A new phase-sensitive color television ultrasonic image system offers hope for improved detection of subtle discontinuities.