Ultrasonic inspection of welds and joints. The service of the Research and Production Center ECHOPLUS in Moscow. Advantages and equipment for narrow control. Scheme and results of the work.
In this article you will learn what non-destructive ultrasonic testing is and what advantages it has. Types of non-destructive testing. Methods and algorithms of work.
In this article you will learn what mechanized ultrasound control is and when it is applied. Equipment for mechanized ultrasonic inspection. Areas of application and algorithms of work.
To increase the speed of recording echo signals and increase the speed of image recovery of reflectors, it is proposed to use a thinned switching matrix (SMC). To obtain a switching matrix that allows to obtain images minimally different from the image obtained by the full switching matrix (FMC), it is proposed to use a genetic algorithm. Two variants of optimization of the switching matrix are considered: element-wise thinning and column thinning. Numerical and model experiments have shown that a thinned switching matrix determined using a genetic algorithm, filled by 25%, allows the formation of images that differ from the image obtained by FMC, with an error of about 3%. Working with the switching matrix by columns allows you to increase the speed of recording echo signals by 4 times. The speed of image recovery increases by the same number of times.
An effective method of replacing zonal focusing with an antenna array is proposed, which is traditionally used in automated ultrasonic inspection of welded joints with narrow cutting to detect defects at the fusion boundary. This method, based on the use of multi-circuit digital focusing antenna technology (CFA), allows you to obtain and analyze high-quality images of reflectors. The proposed method, compared with zonal focusing made using phased array technology (HEADLIGHTS), is less sensitive to the accuracy of positioning the antenna array relative to the seam axis and to changes in the thickness of the control object, makes it possible to estimate the height of defects not by amplitude, but by the size of the glare reflectors.
Currently, in order to increase the speed of preparation of the ultrasound control protocol and reduce the influence of the human factor, reflector recognition (classification) systems based on artificial neural networks are being actively developed. For their more efficient operation, the images of the reflectors must be processed in order to increase the signal-to-noise ratio of the image and its segmentation (clustering). One of the segmentation methods consists in image processing with an adaptive anisotropic diffuse filter, which is used for processing optical images. In model experiments, the effectiveness of using this texture filter for segmentation of images of reflectors reconstructed from echo signals measured using antenna arrays has been demonstrated.
Ultrasonic flaw detection has developed methods for recording and analyzing echo signals to determine the type of reflector and its dimensions. The method of digital focusing with an antenna (CFA) allows you to restore the image of the entire discontinuity boundary using echo signals reflected from the bottom of the control object, taking into account the transformation of the wave type. However, this approach is not always applicable in practice, since the shape of the bottom of the object of control may be unknown. Using the features of the behavior of the reflection coefficient for different types of waves, it is possible to draw a conclusion about the type of reflector from images only on a direct beam. Numerical and model experiments have confirmed the operability of the proposed approach.
Image restoration of reflectors by digital focusing antenna (CFA), along with such advantages as high resolution over the entire area of image restoration of reflectors, the ability to obtain images taking into account the reflection and transformation of the wave type from the boundaries of the object of control, has several disadvantages: a large volume of measured echo signals, a long image recovery time and insufficiently high energy of ultrasonic waves, entered into the object of control. The Plane Wave Imaging (PWI) method allows you to combine the advantages of phased array antenna technology (FAR) and CFA technology. In PWI mode, when a plane wave is emitted, all elements of the antenna array (AR) work (as in the HEADLIGHT mode), which allows you to increase the energy entered into the control object, and echo signals are recorded by all elements of the AR (as in the CFA mode). The images of the reflectors are restored by the combination SAFT method. To obtain an image, the number of emitted plane waves can be used less than the number of antenna array elements, which reduces the volume of measured echo signals. The transfer of calculations to the area of spatial sectors makes it possible to increase the speed of restoration of the presentation of reflectors. Model experiments have shown the positive and negative sides of obtaining images of reflectors by the PWI method in comparison with the CFA method both for the case of using a prism and without a prism.
The TOFD method, widely used in ultrasonic flaw detection, allows to distinguish a crack from a volumetric reflector by the phase of the echo signals and to determine its height with high accuracy. However, the TOFD method without scanning with piezoelectric converters across the welded joint does not allow to determine the displacement of the reflector from the center of the seam, which is very important when evaluating the control results. The scanning devices used for this purpose have a complex design, their price is higher than that of one—dimensional sanitizing devices, and, most importantly, the control time increases significantly. If we use echo signals reflected from the bottom of the object of control, taking into account the change in the type of wave, then a combined image of the reflector can be obtained from a set of partial images recovered by the digital focusing antenna (CFA) method. If the echo signals measured in the combined mode for each piezoelectric transducer are used, it is possible to estimate the displacement of the reflector across the welded joint with an accuracy of ± 1.5 mm. Numerical and model experiments have confirmed the operability of the proposed approach.
When conducting ultrasonic testing, a situation may arise when the values of the recorded echo signals will be greater than the dynamic range of the receiving amplifier and the analog-to-digital converter of the flaw detector. This will cause the echo signals of pulses of large amplitude to undergo a cut-off operation (clipping) and reduce their amplitude, which may lead to an error when estimating the size of the reflector. A declipping method based on the Gershberg algorithm is proposed―Papulis, and its comparison with the declipping method using the least squares method is carried out. Numerical and model experiments have shown that the Gershberg method―Papulisa works more steadily than the least squares method for noisy echo signals and in the case of a rough step of their sampling.
Practically any welded joint cannot be considered as a homogeneous isotropic medium for control by ultrasonic (ultrasonic) waves. If the phase changes during the propagation of the ultrasonic wave are less than 180 degrees, then the medium can be considered as isotropic and homogeneous. Otherwise, the restoration of the image of the reflectors using simple algorithms will lead to the displacement of the reflector blocks from their true positions, and the shape of the glare will be distorted. Moreover, distortions can lead to the fact that instead of one block there will be two or more with a smaller amplitude. As a result, the glare amplitude of the large reflector may not reach the rejection level and the defect will be missed.
The creation of automated recognition systems for reflectors based on images obtained by ultrasonic antenna arrays is a very urgent task. Its solution will increase the speed of preparation of control protocols and increase their reliability by reducing the influence of the human factor.
When conducting ultrasonic testing using antenna arrays, interference pulses may be present in the measured echo signals, which, after the image of the reflectors is restored, may form false glare that complicates image analysis. Such undesirable pulses include pulses of reverberation interference that occur when the probing pulse is reflected from the prism boundaries, and/or pulses reflected from the structural reflector of the object of control. The simplest way to reduce the amplitude of such pulses, in case of their high stability from measurement to measurement, is to subtract a pattern with interference pulses from the measured echo signals. However, if the interference pulses change slightly during ultrasonic testing, slightly changing the arrival time and amplitude, then their suppression by subtracting the noise pattern will not be effective. To reduce the level of weakly varying interference, it is proposed to apply the decorrelation procedure.
The surface of the control objects may be uneven due to its design features. After installation, during operation or preparation for control, the initially flat surface of the object of control may lose this property. Many methods of image reconstruction of reflectors using ultrasonic antenna arrays proceed from the fact that the surface of the object of control is a straight line. Currently, in the practice of ultrasonic monitoring, the method of digital focusing of the antenna array (CFA) is widely used, which involves recording echo signals during radiation and reception by all pairs of the antenna array and restoring the image of reflectors from the measured echo signals by the combined SAFT (C-SAFT) method. If the antenna array scans along the X axis, then by adding together the CFA images obtained for each position of the antenna array, it is possible to obtain a final image with a lower noise level and a higher frontal resolution.
AUGUR-Analysis software has a built-in function that allows for the headlights or CFR images of defects to determine the phase difference of two glare, which allows you to specify the type of defect - volumetric or planar. This information is used in advanced data analysis techniques.