Abstract: A computerized method of image processing using processing circuitry to apply a previously trained machine learning model in a system for non-destructive testing (NDT) of a material is described. The method can include acquiring acoustic imaging data of the material, the acoustic imaging data acquired at least in part using an acoustic imaging modality, generating an acoustic imaging data set corresponding to an acoustic propagation mode, applying the previously trained machine learning model to the acoustic imaging data set, and generating an image of the material depicting a probability of a flaw per pixel or voxel based on the application of the previously trained machine learning model.

Abstract: Acoustic data, such as a full matrix capture (FMC) matrix, can be reconstructed by applying a previously trained decoder machine learning model to one or more encoded acoustic images, such as the TFM image(s), to generate reconstructed acoustic data. A processor can use the reconstructed acoustic data, such as an FMC matrix, to recreate new encoded acoustic images, such as TFM image(s), using different generation parameters (e.g., acoustic velocity, part thickness, acoustic mode, etc.).

Abstract: A full matrix capture (FMC) acoustic acquisition technique can be modified in a manner to compensate for one or more of variation from a nominal distance between an acoustic transducer array and the structure being scanned, or a variation from a nominal angle (e.g., inclination or declination) of the array surface with respect to the structure being scanned. Such modification can include determining incremental delay factors to be added or removed from nominal delay values in A-Scan data based on offset values corresponding to the transmit and receive transducer elements (or apertures) used for acquiring a particular A-Scan instance. The compensated A-Scan data in the FMC acquisition can then be used TFM imaging, for example.

Abstract: Acoustic evaluation of a target can be performed using an array of electro-acoustic transducers. For example. a technique for such evaluation can include generating acoustic transmission events using different transmitting apertures, the apertures defined by corresponding zones along the array, the zones including multiple electro-acoustic transducer elements. In response to the respective acoustic transmission events. respective acoustic echo signals are received. Representations of the respective received acoustic echo signals are coherently summed. The coherently summing includes applying determined nominal element delay factors to the respective representations to approximate a virtual probe normal to a nominal shape of a surface of a structure being inspected. A pixel or voxel value is corresponding to a specified spatial location within the structure being inspected is generated using the coherently summed representations.

Abstract: Flaw detection information, acquired by one or more NDT modalities, can be applied to a trained machine learning model to automatically associate a detected flaw with a cluster of similar flaws. Then, a flaw identification output can be generated based on the association, such as indicted the flaw and an associated probability. In this manner, the described techniques can automatically classify and identify each detected flaw and, in some cases, no flaw conditions. These techniques can shorten the time between part rejection and flaw diagnosis and allow for automated process control.

Abstract: Examples of the present subject matter provide techniques to accurately size flaws using acoustic inspection without expending significant computing resources and time. Examples described herein include techniques to convert amplitude-based inspection images, such as TFM or phased array ultrasonic testing (PAUT) images, into sizing images based on Acoustic Influence Maps (AIMs).

Abstract: Acoustic inspection productivity can be enhanced using techniques to perform compression of acquired acoustic data, such as data corresponding to elementary A-scan or other time-series representations of received acoustic echo data. In various approaches described herein, time-series data can be decimated for efficient storage or transmission. A representation of the time-series data can be reconstructed, such as by using a Fourier transform-based up-sampling technique or a convolutional interpolation filter, as illustrative examples. The techniques described herein can be used for a variety of different acoustic measurement techniques that involve acquisition of time-series data (e.g., A-Scan data). Such techniques include Full Matrix Capture (FMC) applications, plane wave imaging (PWI), or PAUT, as illustrative examples.

Abstract: Systems and methods are disclosed for conducting an ultrasonic-based inspection. The systems and methods perform operations comprising: receiving a plurality of scan plan parameters associated with generating an image of at least one flaw within a specimen based on acoustic echo data obtained using full matrix capture (FMC); applying the plurality of scan plan parameters to an acoustic model, the acoustic model configured to determine a two-way pressure response of a plurality of inspection modes based on specular reflection and diffraction phenomena; generating, by the acoustic model based on the plurality of scan plan parameters, an acoustic region of influence (AROI) comprising an acoustic amplitude sensitivity map for a first inspection mode amongst the plurality of inspection modes; and generating, for display, a first image comprising the AROI associated with the first inspection mode for capturing or inspecting the image of the at least one flaw.

Abstract: Using various techniques, a position of a probe assembly of a non-destructive inspection system, such as a phase array ultrasonic testing (PAUT) system, can be determined using the acoustic capability of the probe assembly and an inertial measurement unit (IMU) sensor, e.g., including a gyroscope and an accelerometer, without relying on a complex encoding mechanism. The IMU sensor can provide an estimate of a current location of the probe assembly, which can be confirmed by the probe assembly, using an acoustic signal. In this manner, the data acquired from the IMU sensor and the probe assembly can be used in a complementary manner.

Abstract: Acoustic evaluation of a target can be performed using an array of electro-acoustic transducers. For example, a technique for such evaluation can include generating pulses for transmission by respective ones of a plurality of electro-acoustic transducers in a transducer array to contemporaneously establish respective acoustic beams corresponding to at least two different acoustic beam steering directions for an acquisition, the pulses comprising at least a first sequence having pulses of defining a profile having a first polarity, the first sequence corresponding to a first beam steering direction (e.g., angle or spatial beam direction), and a second sequence having pulses defining a profile having a second polarity opposite the first polarity, the second sequence corresponding to a second beam steering direction. In response to transmission of the pulses, respective acoustic echo signals can be received and aggregated to form an image of a region of interest on or within the target.

Abstract: Techniques for compensating the sensitivity variations induced by lift-off variations for an eddy current array probe. The techniques use the eddy current array probe coils in two separate ways to produce a first set of detection channels and a second set of lift-off measurement channels without the need to add coils 5 dedicated to the lift-off measurement operation.

Abstract: A material profile can be determined by assuming an elliptical or circular arc geometry and by using acoustic noise generated by diffuse internal reflection and/or specular reflection on the internal (ID) or external (OD) interface of the material under inspection, such as a pipe or curved plate. A non-destructive testing (NDT) technique can acquire acoustic data of the material using an ultrasonic signal. The acquired acoustic data can be filtered such that the acoustic noise generated by diffuse internal reflection and/or specular reflection is separated from any flaws in the material. Positions of the acoustic noise can be determined and then a regression technique can be applied to the positions, which can generate an equation of a circle, for example, such as to provide a radius and thickness of the pipe or curved plate.

Abstract: Transmit-Receive Longitudinal (TRL) probes can be used for the inspection of noisy material, such as austenitic materials. By using various techniques, an inspection area is not constrained by a wedge design of an ultrasonic probe and the benefits of using a linear probe array (rather than a matrix) are maintained. Volumetric or TFM-like imaging on austenitic materials using a linear transmit array and a linear receive array that are out of plane with one another (a TRL configuration) and not in the main imaging place can simplify the inspection and analysis of such materials. For each scan position, an ultrasound probe can acquire acoustic imaging data. Then, a processor can then combine acquisitions from adjacent scan positions to create an imaging result using synthetic aperture focusing technique (SAFT) principles to recreate a focalization in a passive axis of the probe.

Abstract: An eddy current probe assembly for analyzing an article is provided. The eddy current probe assembly includes a housing, an eddy current probe disposed on or within the housing, and a barrier layer. The housing has a coolant passage that extends into the housing and forms a loop within the housing. The coolant passage allows for coolant flow through the housing. The eddy current probe thermally couples with a first portion of the coolant passage at a first surface of the eddy current probe. The barrier layer is on a second surface of the eddy current probe opposite the first surface of the eddy current probe.

Abstract: Generally, in a non-destructive test application, a compressed representation of an acoustic echo signal acquired by a non-destructive test (NDT) probe assembly can be received, such as via a network. The compressed representation can include data indicative of changes in phase values of the acoustic echo signal. Using the compressed representation, a time-domain representation of an instantaneous phase signal can be constructed from the compressed representation. The constructed instantaneous phase signal can be used in constructing at least one of an uncompressed acoustic echo signal representation or an image. As an illustration, amplitude values of sampled acoustic echo signals can be suppressed in the compressed representation, reducing data volume associated with transmitting a representation of the acquired acoustic echo signal.

Abstract: An ultrasound probe is capable of detecting flaws in an object in a non-destructive manner. The probe includes a row-column addressed (RCA) array with a plurality of row and column electrodes. The probe can perform volumetric inspection of an object using the RCA array in different transmission and reception configurations.

Abstract: An acoustic inspection system can be used to generate a surface profile of a component under inspection, and then can be used to perform the inspection on the component. The acoustic inspection system can obtain acoustic imaging data, e.g., FMC data, of the component. Then, the acoustic inspection system can apply a previously trained machine learning model to an encoded acoustic image, such as a TFM image, to generate a representation of the profile of one or more surfaces of the component. In this manner, no additional equipment is needed, which is more convenient and efficient than implementations that utilize additional components that are external to the acoustic inspection system.

Abstract: A computerized method of image processing using processing circuitry to apply a previously trained machine learning model in a system for non-destructive testing (NDT) of a material is described. The method can include acquiring acoustic imaging data of the material, the acoustic imaging data acquired at least in part using an acoustic imaging modality, generating an acoustic imaging data set corresponding to an acoustic propagation mode, applying the previously trained machine learning model to the acoustic imaging data set, and generating an image of the material depicting a probability of a flaw per pixel or voxel based on the application of the previously trained machine learning model.

Abstract: Systems and methods are disclosed for conducting an ultrasonic-based inspection. The systems and methods perform operations comprising: receiving a plurality of scan plan parameters associated with generating an image of at least one flaw within a specimen based on acoustic echo data obtained using full matrix capture (FMC); applying the plurality of scan plan parameters to an acoustic model, the acoustic model configured to determine a two-way pressure response of a plurality of inspection modes based on specular reflection and diffraction phenomena; generating, by the acoustic model based on the plurality of scan plan parameters, an acoustic region of influence (AROI) comprising an acoustic amplitude sensitivity map for a first inspection mode amongst the plurality of inspection modes; and generating, for display, a first image comprising the AROI associated with the first inspection mode for capturing or inspecting the image of the at least one flaw.

Abstract: An eddy current (EC) detection system comprises an EC probe including a plurality of sensors to provide corresponding EC response signals; and processing circuitry to evaluate speed of the EC probe based on a measurement of similarity of the EC response signals; determine whether the speed of the EC probe is too fast or two slow based on quality of the measurement; and generate a command to adjust speed of the EC probe during further EC inspection.