System and method for detecting microtextured regions in a subject

Abstract

To provide an inspection system and method for detecting an MTR present in a subsurface volume of a subject.SOLUTION: In this approach, an acoustic transducer and, optionally, a near surface sensor are used to introduce interrogating energy into the subject. Signal data representative of the test energy is analyzed to detect the MTR. In some approaches, a shift in the frequency distribution of the signal data is determined. In other approaches, a distribution of values for a given characteristic of the signal data, such as amplitude or frequency, is computed, and a quantified description of that distribution is computed. A response score and / or an intensity map for the subject may be generated based on the analysis of the signal data. An MTR score indicative of the MTR within the subject can be correlated to a response score and / or an intensity map. The subject may then be treated based on the response score and / or intensity map and their correlation with the MTR score.SELECTED DRAWING: Figure 1A

Description

[Technical Field]

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS This application claims the benefit of priority to Indian Provisional Application No. 202411047057, filed June 19, 2024, U.S. Provisional Application No. 63 / 665,988, filed June 28, 2024, Indian Provisional Application No. 202411076142, filed October 8, 2024, and U.S. Provisional Application No. 63 / 745,900, filed January 16, 2025, all of which are incorporated herein by reference.

[0002] These teachings relate generally to systems and methods for inspecting components, and in particular for inspecting components to determine one or more topographical characteristics. [Background technology]

[0003] Machine components in various industries often undergo periodic inspections to identify defects or anomalies present in the components that may adversely affect the performance of the components. The inspections may be used to evaluate the condition or quality of the components, to determine whether the components are suitable for continued service, and / or to develop maintenance or replacement schedules for the components.

[0004] In the aerospace industry, various components are formed from metal alloys, such as titanium alloys. Certain metal alloys can contain microstructures, such as micro-texture regions (MTRs), that can degrade the performance of components formed from the alloy. Therefore, it would be useful to have inspection devices and methods for evaluating the microstructural properties of such materials. Summary of the Invention [Means for solving the problem]

[0005] Various needs are met, at least in part, through the provision of systems and methods for detecting microstructural characteristics that are described in the following detailed description, particularly when studied in conjunction with the drawings. A full and useful disclosure of the illustrative embodiments of the invention, including the best mode, directed to one skilled in the art, is set forth in this specification, which makes reference to the accompanying drawings. [Brief explanation of the drawings]

[0006] [Figure 1A] 1 is a schematic diagram of an acoustic inspection system, according to various embodiments. [Figure 1B] 1 is a flow diagram of a method of operating an acoustic inspection system, according to some embodiments. [Figure 1C] 1 is a flow diagram of a method of operating an acoustic inspection system, according to some embodiments. [Figure 1D] 1 is a flow diagram of a method of operating an acoustic inspection system, according to some embodiments. [Figure 2] FIG. 1 is a schematic diagram of a sensor arrangement, according to some embodiments. [Figure 3] FIG. 1 is a schematic diagram of a sensor arrangement, according to some embodiments. [Figure 4A] FIG. 1 is a schematic diagram of a sensor arrangement, according to some embodiments. [Figure 4B] FIG. 1 is a schematic diagram of a sensor arrangement, according to some embodiments. [Figure 5] FIG. 1 is a schematic diagram of a sensor arrangement, according to some embodiments. [Figure 6A] FIG. 1 is a schematic diagram of a sensor arrangement, according to some embodiments. [Figure 6B] FIG. 1 is a schematic diagram of a sensor arrangement, according to some embodiments. [Figure 7A] FIG. 1 is a schematic diagram of a sensor arrangement, according to some embodiments. [Figure 7B] FIG. 1 is a schematic diagram of a sensor arrangement, according to some embodiments. [Figure 8]FIG. 1 is a schematic diagram of a sensor arrangement, according to some embodiments. [Figure 9A] FIG. 1 is a schematic side view of a sensor arrangement, according to some embodiments. [Figure 9B] 1 is a schematic cross-sectional side view of a sensor arrangement, according to some embodiments. [Figure 9C] FIG. 9C is a schematic bottom end view of the sensor arrangement of FIG. 9B. [Figure 9D] 1 is a schematic cross-sectional side view of a sensor arrangement, according to some embodiments. [Figure 9E] FIG. 9E is a schematic bottom end view of the sensor arrangement of FIG. 9D. [Figure 9F] FIG. 1 is a schematic side view of a sensor arrangement, according to some embodiments. [Figure 9G] FIG. 9F is a schematic bottom end view of the sensor arrangement of FIG. 9F. [Figure 9H] FIG. 9F is a schematic top view of the subject of FIG. 9F. [Figure 9I] 1 is a schematic side view of a sensor arrangement according to some embodiments. FIG. [Figure 9J] FIG. 9J is a schematic bottom end view of the sensor arrangement of FIG. 9I. [Figure 9K] 1 is a schematic cross-sectional side view of a sensor arrangement, according to some embodiments. [Figure 9L] 1 is a schematic side view of a sensor arrangement according to some embodiments. FIG. [Figure 9M] 1 is a schematic cross-sectional side view of a sensor arrangement, according to some embodiments. [Figure 9N] FIG. 9C is a schematic bottom end view of the sensor arrangement of FIG. 9M. [Figure 9O] 1 is a schematic cross-sectional side view of a sensor arrangement, according to some embodiments. [Figure 9P] FIG. 9B is a schematic bottom end view of the sensor arrangement of FIG. 9O. [Figure 9Q] 1 is a schematic cross-sectional side view of a sensor arrangement, according to some embodiments. [Figure 9R] FIG. 9Q is a schematic bottom end view of the sensor arrangement of FIG. 9Q. [Figure 9S] 1 is a schematic side view of a sensor arrangement according to some embodiments. FIG. [Figure 9T] FIG. 9C is a schematic bottom end view of the sensor arrangement of FIG. 9S. [Figure 10] FIG. 1 is a schematic diagram of a sensor arrangement, according to some embodiments. [Figure 11A] 1 is a flow diagram of a method for analyzing signal data, according to some embodiments. [Figure 11B] FIG. 11B includes exemplary data from the analysis described in FIG. 11A. [Figure 11C] FIG. 11B includes exemplary data from the analysis described in FIG. 11A. [Figure 12] 1 is a flow diagram of a method for analyzing signal data, according to some embodiments. [Figure 13] 1 is a flow diagram of a method for analyzing signal data, according to some embodiments. [Figure 14] 1 is a flow diagram of a method for analyzing signal data, according to some embodiments. [Figure 15] 1 is a flow diagram of a method for analyzing signal data, according to some embodiments. [Figure 16A] 1 is a flow diagram of a method for analyzing signal data, according to some embodiments. [Figure 16B] FIG. 16B includes exemplary data from the analysis described in FIG. 16A. [Figure 16C] FIG. 16B shows the formula used in the method of FIG. 16A. [Figure 17] 1 is a flow diagram of a method for analyzing signal data, according to some embodiments. [Figure 18] 1 is a flow diagram of a method for analyzing signal data, according to some embodiments. [Figure 19] 1 is a flow diagram of a method for analyzing signal data, according to some embodiments. [Figure 20A]1 is a flow diagram of a method for analyzing signal data, according to some embodiments. [Figure 20B] FIG. 1 illustrates a formula used by some embodiments. [Figure 21] 1 is a flow diagram of a method for analyzing signal data, according to some embodiments. [Figure 22] 1 is a flow diagram of a method for analyzing signal data, according to some embodiments. [Figure 23] 1 is a flow diagram of a method for analyzing signal data, according to some embodiments. [Figure 24] 1 is a flow diagram of a method for analyzing signal data, according to some embodiments. [Figure 25] 1 is a flow diagram of a method for analyzing signal data, according to some embodiments. [Figure 26A] 1 is a flow diagram of a method for analyzing signal data, according to some embodiments. [Figure 26B] FIG. 26B includes exemplary data from the analysis described in FIG. 26A. [Figure 26C] FIG. 26B includes exemplary data from the analysis described in FIG. 26A. [Figure 27A] 1 is a flow diagram of a method for analyzing signal data, according to some embodiments. [Figure 27B] 1 is a schematic diagram of a subject and a macroslice. [Figure 27C] FIG. 1 shows a sensor inspecting a macro-slice. [Figure 27D] FIG. 1 shows an intensity map for a macro slice. [Figure 27E] FIG. 10 shows a raw intensity map for a macro slice after calibration. [Figure 27F] FIG. 10 shows an intensity map for a macro slice after calibration. [Figure 28]1 is a graph of a signal envelope for exemplary waveform data in accordance with some embodiments. [Figure 29] FIG. 1 illustrates equations used in signal data analysis, according to some embodiments. DETAILED DESCRIPTION OF THE INVENTION

[0007] Elements in the figures are illustrated for simplicity and clarity and are not necessarily drawn to scale. For example, the dimensions and / or relative positions of some of the elements in the figures may be exaggerated relative to other elements to help improve understanding of various embodiments of the teachings herein. Also, common but well-understood elements that are useful or necessary in commercially feasible embodiments are often not depicted to facilitate understanding of these various embodiments of the teachings herein. While some operations and / or steps may be described or depicted in a particular order of appearance, those skilled in the art will understand that such details regarding order are not actually required.

[0008] Conventional inspection systems and methods used to determine microstructural characteristics in samples utilize destructive testing, such as electron backscatter diffraction (EBSD). Typically, specimens to be inspected via EBSD testing are also highly prepared. For example, preparing a specimen for EBSD testing may involve polishing the specimen to a mirror finish. The acoustic inspection devices and methods described herein involve a non-destructive approach to detecting microstructural characteristics of specimens formed from materials having a crystalline structure. Non-destructive testing may evaluate components while avoiding damage to the components during the inspection process. Furthermore, the inspection systems and methods described herein may be performed on simple surfaces that are not highly prepared and do not need to be polished to a mirror finish. Advantageously, the inspection systems and methods herein may be used to evaluate peened and machined surfaces.

[0009] The inspection systems and methods described herein can be used to determine one or more properties of an object without destroying the object or requiring tedious or time-consuming surface preparation. Material properties that can be identified or detected by these inspection systems and methods include microstructural properties related to the object's microstructure. For example, the inspection systems and methods can be used to identify microstructural properties related to external discontinuities such as cracks, voids, inclusions, and material anomalies (regions with different densities or moduli of elasticity). The microstructural properties detected by these inspection systems and methods can also be intrinsic properties (e.g., non-discrete type properties).

[0010] Advantageously, the described inspection systems and methods can identify or detect microstructural characteristics indicative of microtextured regions (MTRs) present within a volume or near-surface volume of a specimen. Examples of MTR characteristics that can be detected or identified by the inspection systems described herein include, but are not limited to, grain (crystal) size, grain orientation (crystalline orientation), grain shape, grain volume, discontinuous microstructure, grain phase, grain chemical composition, presence of MTRs, MTR size, quantiles of an MTR distribution determined via an experimental distribution, Weibull distribution, etc., such as B50 MTR size (50th percentile of MTR size or area distribution), B70 MTR size, B95 MTR size, intensity of MTRs, orientation of MTRs, combination of adjacent MTRs with different orientations, MTR fraction, macrotexture, dislocation content, residual elastic compressive or tensile stress, or the like. The material property can be a property of the object exhibiting the MTR, or any property of the MTR, such as size, shape, intensity, density, frequency, orientation, orientation variation, adjacent region properties, volume fraction, etc.

[0011] As used herein, B50, B70, and B95 refer to the 50th, 70th, and 95th percentiles, respectively, of the distribution of an MTR property, such as MTR size, MTR area, etc.

[0012] In addition to standard statistical measures derived from MTR characteristics, such as size data, including mean, median, etc., each set of raw MTR distribution data can also be fitted to multiple statistical distributions. It is contemplated that statistical measures from such distributions can be related to signal data derived from one or more of the transducer arrangements and / or analysis methods described herein.

[0013] As used herein, MTR may refer to agglomerations of grains with similar crystallographic orientation in titanium and nickel alloys. As used herein, macrotexture may refer to MTR or individual grains with preferred crystallographic orientation relative to the sample axes. It should be noted that changes in material properties can affect the performance of an alloy. For example, as MTR increases in size and / or intensity in titanium alloys, room temperature hold fatigue properties decrease. Similarly, as macrotexture increases in titanium alloys (as measured by the intensity of the c-axis alignment of the crystals in the sample), the 0.2% yield strength and ultimate tensile strength increase. In nickel alloys, fatigue properties deteriorate as grain size increases.

[0014] As used herein, an MTR score is derived from one or more MTR characteristics or statistics of those MTR characteristics. For example, the MTR score can be the average MTR size, the B50 MTR size, the B70 MTR size, the B95 MTR size, the MTR fraction, the MTR intensity, or a combination of one or more MTR characteristics, such as the MTR size multiplied by the MTR fraction, the MTR size divided by the MTR intensity, the MTR orientation gradient, the MTR area multiplied by the MTR fraction, or the MTR area divided by the MTR intensity. The MTR characteristics of a specimen can be obtained from EBSD, which is considered an authoritative method. An exemplary code used to calculate MTR properties is DREAM.3D, which includes MTR property definitions that correlate to room temperature hold fatigue performance, as described in detail in A. Pilchack et al., "DOT / FAA / TC-23 / 40, US Dep't of Transp., F, Cold Dwell Fatigue of Titanium Alloys: History, Current State, and Aviation Industry Perspective," September 2024. In the present case, the DREAM.3D code definitions traditionally used to correlate MTR scores with room temperature hold fatigue performance can be adjusted to provide a good correspondence between such MTR scores and ultrasonic or electromagnetic signals (or parameters derived therefrom).

[0015] It should be understood that the term "MTR score" may be referred to interchangeably with one or more alternative terms or phrases that reflect a ground truth measurement of the MTR or MTR characteristic. For example, the phrase "indicating that an MTR is present" may indicate that an MTR score is calculated to indicate that an MTR is present in a portion of a subject, such as within a subsurface volume. In another example, "predicting the level of subsurface MTR" may indicate that an MTR score is calculated. Additionally, the term "MTR characteristic" may be used interchangeably with "MTR score."

[0016] Particles and particle clusters (MTRs) in metals and alloys scatter and attenuate acoustic or electromagnetic energy. When an acoustic or electromagnetic signal strikes a particle or MTR, the particle or MTR may reflect the signal or act as a point source that vibrates and generates a signal with a frequency characteristic of the MTR. The resulting scattered and reflected signal, referred to herein as scattering or backscattering, conveys information about the scatterer. This scattering / reflection process reduces the amplitude / energy of the transmitted signal, which can provide information about the nature and size of the MTR. This effect can be visible when comparing time, frequency, or time-frequency spectra from various locations and time windows of signal data.

[0017] Given a known relationship between acoustic or electromagnetic signal characteristics and MTR characteristics, the approaches described herein use acoustic or electromagnetic response scores generated from one or more testing methods to guide subject disposition (pass, fail, grade). These testing methods can obtain one or more intensity maps or response scores for a subject using different combinations of transducer arrangements, sensor arrangements, and analysis methods described herein. The acoustic or electromagnetic response scores are related to the MTR score. In addition, the analysis methods described herein can be used to classify subjects based on their response scores or subregions of the subject based on individual subregion response scores.

[0018] As used herein, a response score is generated from an analysis of acoustic, electromagnetic, or other signals generated when inspecting a metal piece, which correlates with an MTR score. A response score may be generated from any of the analysis methods described herein for analyzing signal data. Some phrases may indicate that a response score is generated or may be generated. In one example, a response score may be generated from a spatial intensity map. In another example, a response score may be generated from a linear MTR profile, or a spatial MTR profile may be used to generate a response score. In some examples, a response score characterizes a subject based on macro-slice quantification signal data and subject signal data. The response score corresponds to the signal data alone. The MTR score represents the "ground truth" of the subject. By correlating a sufficient number of "response scores" with "MTR scores," a determination can be made as to which "response scores" indicate undesirable "MTR scores" and the subject can be treated accordingly.

[0019] After a classification or other determination is made, various actions are taken. In some embodiments, if the classification is favorable (e.g., an acoustic or electromagnetic response score correlated to an MTR score below a threshold), the specimen is moved from the inspection area to the next station in the manufacturing process (e.g., manually moved or electronically controlled machinery such as a robot to physically move the specimen forward in the manufacturing process for its intended use). In other embodiments, if the classification is unfavorable (e.g., an acoustic or electromagnetic response score correlated to an MTR score above a threshold), the affected material may be cut or otherwise removed from the specimen (manually or by electronically controlled bandsaws or similar instruments), or the entire specimen may be scrapped (e.g., moved to a designated scrap area manually or by electronically controlled machinery such as a robot) or downgraded to a lower-quality manufacturing use. In still other embodiments, the determined acoustic or electromagnetic response score correlated to the MTR score may be used in future manufacturing processes. For example, parameters of a manufacturing process or machine producing a new part may be adjusted, set, and / or changed based on the acoustic or electromagnetic response score to reduce the MTR score formed during production of the new part. Adjusting, setting, or changing a parameter may involve, for example, changing a parameter value stored in electronic memory associated with the manufacturing process or machine. In other aspects, it may involve manually adjusting an actuator at the machine or an actuator controlling the process. In some examples, the adjusted, set, or changed parameter causes the manufacturing process or machine to operate differently than before the parameter was adjusted, set, or changed. In still other aspects, the parameter may be updated periodically or continuously.

[0020] The inspection systems and methods described herein may be used on engine parts and components, such as gas turbine engines. Exemplary components include, but are not limited to, components made from titanium (Ti) or nickel (Ni) alloys, such as fan disks / blades, compressor disks / blades, etc. In addition, the systems and methods described herein may be used to inspect titanium or nickel billets or forgings.

[0021] The terms and expressions used herein have the ordinary technical meaning as ascribed to such terms and expressions by one of ordinary skill in the art as set forth above, unless a different specific meaning is otherwise stated herein. The word "or," when used herein, shall be construed as having a disjunctive rather than a conjunctive meaning, unless otherwise specified herein. "Coupled," "fixed," "attached," and similar phrases refer to both direct coupling, fixing, or attachment, as well as indirect coupling, fixing, or attachment via one or more intermediate components or features, unless otherwise specified herein.

[0022] In the English original, the singular forms with the articles "a," "an," and "the" include the plural forms unless the context clearly dictates otherwise.

[0023] As used throughout this specification and claims, the term "approximate" is applied to modify any quantitative expression that can be acceptably varied without resulting in a change in the basic function to which it pertains. Thus, a value modified by one or more words such as "approximately," "about," and "substantially" should not be limited to the exact value specified. In at least some instances, the term "approximate" may correspond to the precision of an instrument for measuring a value or the precision of a method or machine for constructing or manufacturing a component and / or system. For example, the term "approximate" may refer to being within a 10% margin.

[0024] 1A, there is illustrated an inspection system 100. The inspection system 100 may be used to inspect an object under test 102 to determine at least one MTR characteristic present within a volume of the object under test 102.

[0025] The specimen 102 can be any milled product (e.g., billet, bar, plate, sheet, slab, etc.), intermediate machined product, or component formed from the milled product. The specimen 102 can also be a forged or fabricated part. The specimen 102 can be alpha-beta processed, beta annealed, or an as-cast structure in the form of a casting or additively built part that is cooled through the beta transus. It is contemplated that the specimen 102 can be formed from other material processing methods that include cooling through the beta transus, such as welding.

[0026] The specimen 102 can be a billet, such as a titanium or nickel billet. In some examples, the specimen 102 is an end slice of a billet, a macro-slice of a forged part (e.g., a fan disk), the surface of a forged part, or any other suitable sample taken from a billet, forging, or part before or after any step of thermomechanical processing of a metal alloy. In some approaches, the specimen 102 is a part or component of a gas turbine engine (e.g., a part or component before assembly or a part or component disassembled from a gas turbine engine). As used herein, a macro-slice refers to a thin-section sample extracted from a larger sample. The specimen 102 can be characterized non-destructively and destructively for the purpose of characterizing the larger sample.

[0027] The inspection system 100 comprises an inspection device 104 including one or more volume acoustic transducers 106, one or more near-surface sensors 134, and a controller 108. The system 100 comprises a transducer positioning system 110 and, optionally, one or more databases 112.

[0028] The inspection device 104 is configured to transmit and / or receive inspection energy that propagates within the volume of the subject 102. The inspection device 104 can use various types of inspection energy, including acoustic energy (e.g., acoustic waves), electromagnetic energy (e.g., induced current), or a combination thereof. In some examples, the inspection energy is ultrasound. The inspection energy is used to interrogate the subject 102 to detect or identify material properties, such as MTR or other particle properties, present within the volume of the subject 102. The MTR or particle properties may affect the manner in which the inspection energy propagates within the subject 102. In this manner, properties of acoustic waves scattered, reflected, and / or transmitted through the subject 102 by the acoustic transducer 106 of the inspection device 104 may be analyzed by the controller 108 to identify or detect MTR properties associated with the subject 102.

[0029] The inspection device 104 may be configured to transmit and / or receive acoustic waves or induced currents along or near the surface of the subject 102. In some examples, the acoustic waves are Rayleigh surface waves. In some examples, the induced currents are eddy currents generated by a coil. The acoustic waves and / or induced currents are used to interrogate the subject 102 to detect or identify MTR characteristics present at or near the surface of the subject 102. The MTR characteristics may affect the manner in which the acoustic waves or induced currents propagate within the subject 102. In this manner, characteristics of the acoustic waves or induced currents sensed by the near-surface sensor 134 may be analyzed by the controller 108 to identify or detect MTR characteristics associated with the subject 102.

[0030] The near-surface sensor 134 may include a variety of sensor arrangement configurations. One such embodiment is described in detail with reference to Figures 9A-9T. In some embodiments, the near-surface sensor 134 may include a separate acoustic or electromagnetic transmitter 136 and acoustic or electromagnetic receiver 138. The near-surface sensor 134 may include one or more of an ultrasonic sensor, an eddy current sensor, an optical polarization sensor, or a combination thereof.

[0031] Acoustic transducer 106 may include a variety of transducer arrangements. Particular embodiments of transducer arrangements that may be used in acoustic transducer 106 are described in detail with reference to Figures 2-7B.

[0032] In some embodiments, the acoustic transducer 106 may be a transducer array including multiple transducer elements acting as transmitters and receivers. In some configurations, a single transducer array may be used to generate and receive acoustic waves propagating through the volume of the subject 102. As such, the transducer array may be configured to generate and receive acoustic waves propagating through the volume of the subject 102. The elements may operate in a transmit mode, in which the transducers operate to activate piezoelectric elements to generate acoustic waves. The elements may also operate in a receive (or standby) mode, in which the transducers wait to receive acoustic waves. The elements may be activated or operated between the transmit mode and the receive mode. FIG. 2 shows an exemplary transducer arrangement using a transducer array including multiple transducer elements acting as transmitters and receivers. It is also contemplated that such a transducer array may be implemented as the acoustic transducers in FIGS. 3 and 5.

[0033] In some embodiments, the acoustic transducer 106 may include a separate acoustic transmitter 114 and acoustic receiver 116. Figures 4A-4B, 6A-6B, and 7A-7B show exemplary transducer arrangements using separate acoustic transmitters and receivers.

[0034] In some configurations, the acoustic transmitter 114 may be a single-element acoustic transmitter configured to propagate acoustic waves through a volume of the subject 102. In other configurations, the acoustic transmitter 114 may be a transmit acoustic transmitter array including multiple acoustic transmitters (e.g., elements) configured to generate acoustic waves that propagate through the volume of the subject 102. The transmit acoustic transmitter array may be a linear array or a matrix array. In some configurations, the acoustic receiver 116 and the acoustic transmitter 114 are a single device, for example, a single sensor element configured to pass acoustic waves through the volume of the subject 102 and receive the acoustic waves from the subject 102. In other configurations, the acoustic receiver 116 may be a receive acoustic transmitter array including multiple acoustic receivers (e.g., elements) configured to receive or detect acoustic waves. The receive acoustic transmitter array may be a linear array or a matrix array.

[0035] The acoustic transducer 106 and / or near-surface sensor 134 may be configured to inspect or interrogate multiple inspection zones 164 within the subject 102. FIG. 6 illustrates an inspection zone 664 that may be used as an inspection zone 164 according to one exemplary, non-limiting embodiment. It is contemplated that the subject 102 may be divided into any number of inspection zones 164 to facilitate testing the subject 102 for MTR. The inspection zones 164 may have any suitable shape or configuration relative to a larger area or volume of the subject 102. In some embodiments, the inspection zone 164 is a chord (see, for example, first inspection zone 664A and second inspection zone 664B in FIG. 6). In some embodiments, the inspection zone 164 is a region extending across the subject from one side to the other; for example, the inspection zone 164 may extend axially, longitudinally, radially, or in other directions. In some embodiments, the examination zones 164 may be arranged at a variety of different depths within the subject 102 .

[0036] The acoustic transducer 106 may be configured to generate shear waves, longitudinal waves, mixed-mode acoustic waves, or a combination thereof. In some examples, when the subject 102 is round, a transducer arrangement using shear waves may not reach deep into the subject 102, so a transducer arrangement using shear waves may be used to inspect shallower examination zones 164 or regions of the subject, while a transducer arrangement using longitudinal waves may be used to inspect both shallow and deeper examination zones 164 or regions of the subject 102.

[0037] The controller 108 is in operative communication with the acoustic transducer 106. The controller 108 is configured to operate the acoustic transducer 106, e.g., to control the transmission and reception of acoustic waves. By some approaches, the controller 108 is configured to control the timing and / or frequency of the acoustic waves transmitted or generated by the acoustic transducer 106. The controller 108 also receives signal data representative of the acoustic waves generated by the acoustic transducer 106 and scattered, reflected, and / or transmitted through the subject 102. For example, in some embodiments, the controller 108 receives signal data 130, which is described further below. In some embodiments, the controller 108 receives electrical measurement signals indicative of the acoustic waves received by the acoustic transducer 106.

[0038] In some embodiments, when the acoustic transducer 106 includes a transducer array, the controller 108 is configured to activate or fire elements of the transducer array (e.g., individual acoustic transducers) simultaneously or at specified times to adjust the incident wavefront. For example, the controller 108 can activate the elements of the transducer array using timing delays to control when each element of the transducer array is activated or pulsed. The controller 108 can control when each element of the transducer array is fired to steer the wavefront and / or generate coherent or focused waves through the interaction of the acoustic waves generated by each element. In some approaches, the controller 108 can operate the transducer array to form a focused acoustic wave having a focal point at a predetermined or controlled focal depth.

[0039] In some embodiments, the controller 108 may be configured to excite the elements of the transducer array with any waveform shape, such as a negative square wave, a bipolar square wave, a chirp excitation, a tone burst excitation, or a coded excitation. The controller 108 may function as a computing device that performs the functions and methods described herein. For example, the controller 108 may perform one or more of the methods described in FIGS. 1B-1D. Furthermore, the controller 108 may be configured to perform one or more of the signal analysis techniques described herein. In some embodiments, the controller 108 is configured to receive signal data from the near-surface sensor 134 and / or the acoustic transducer 106.

[0040] The controller 108 may be capable of performing one or more signal transformation techniques to transform the signal data 130. Such signal transformation techniques may include one or more of a Hilbert transform, a fast Fourier transform, wavelet analysis, convolution, or deconvolution. The controller 108 may be configured to perform a Hilbert transform, a fast Fourier transform, or the like on the signal data 130 to isolate amplitude and phase characteristics of the signal data 130. The controller 108 may also be configured to rescale the signal data to a standardized magnitude or time framework for direct comparison with other signal data 130.

[0041] The controller 108 may include one or more processors 124, input / output (I / O) devices 126, and memory devices 118. The processor 124 may include any suitable processing device, such as a microprocessor, microcontroller, integrated circuit, logic device, or other suitable processing device. The processor 124 may be used to perform or assist in performing steps of the processes, methods, functions, and techniques described herein, and to control various communications, decisions, programs, content, lists, services, interfaces, logging, reports, etc. Additionally, the one or more processors 124 may access memory devices 118, which may store instructions 122, code, and the like, that are executed by the processor 124 to implement the intended functions.

[0042] Memory device 118 typically includes one or more processor-readable and / or computer-readable media accessed by at least processor 124 and may include volatile and / or non-volatile media, such as RAM, ROM, EEPROM, flash memory, and / or other memory technologies. Additionally, while memory device 118 is shown as being internal to controller 108, memory device 118 may be internal memory, external memory, or a combination of internal and external memory. Similarly, some or all of memory device 118 may be internal memory, external memory, or a combination of internal and external memory of processor 124. Memory device 118 may be virtually any associated memory, such as, but not limited to, one or more of a solid-state storage device or drive, a hard drive, a universal serial bus (USB) stick or drive, a flash memory, a secure digital (SD) card, other memory card, and other such memory, or a combination of two or more such memories, and some or all of the memory may be distributed at multiple locations on a computer network.

[0043] The memory device 118 may store data 120, such as code, software, executable files, scripts, data, content, lists, programming, programs, log or historical data, engine information, component information, and the like. While FIG. 1A illustrates various components coupled via a bus, it is understood that in practice, various components may be directly coupled to the controller 108 and / or one or more other components. The memory device 118 stores sets of operational code or instructions 122 that are executed by the controller 108 and / or processor 124 to implement the functionality of the inspection system 100 or portions thereof. In some embodiments, the memory device 118 may also store some or all of the data 120 that may be needed to inspect the subject 102.

[0044] I / O device(s) 126 may be any relevant port or combination of ports, such as, but not limited to, USB, Ethernet, or other such ports. I / O device(s) 126 may be configured to allow wired and / or wireless communication coupling to external components. For example, I / O device(s) 126 may provide wired and / or wireless communication (e.g., Wi-Fi, Bluetooth, cellular, RF, and / or other such wireless communication) and, in some cases, may include any suitable wired and / or wireless interface devices, circuits, and / or connection devices, such as, but not limited to, one or more transmitters, receivers, transceivers, or combinations of two or more such devices.

[0045] The user interface 128 may be used to control one or more components of the inspection system 100. The inspection system 100 and its various components may be operated and controlled via the user interface 128. The user interface 128 may be used to display user input and / or output. For example, the user interface 128 may include any known input device, such as one or more buttons, knobs, selectors, switches, keys, touch input surfaces, voice inputs, displays, etc. Additionally, the user interface 128 may include one or more output display devices, such as lights, visual indicators, display screens, etc., to communicate information to the user, such as, but not limited to, communication information, status information, order information, shipping information, notifications, errors, conditions, and / or other such information. Similarly, the user interface 128 in some embodiments may include an audio system capable of receiving voice commands or requests verbally issued by the user and / or outputting audio content, alerts, and the like.

[0046] The transducer positioning system 110 is in operative communication with the inspection device 104. More particularly, the transducer positioning system 110 is in operative communication with the acoustic transducer 106 and the controller 108. The transducer positioning system 110 is configured to adjust the position of the acoustic transducer 106 with respect to the subject 102. Furthermore, the controller 108 may send signals instructing the transducer positioning system 110 to adjust the position of the acoustic transducer 106 or the near-surface sensor 134.

[0047] The transducer positioning system 110 may include a movable platform, a turntable, a robotic arm, or any other device capable of holding and / or moving the subject 102. The transducer positioning system 110 may also include a device operable to move the acoustic transducer 106 or near-surface sensor 134, such as a robotic arm or a mast that moves via a drive system. Such a mast may be a straight XY scanner or may have a gimbal and / or swivel angle control device. The transducer positioning system 110 may also be two separate controllable devices that operate in conjunction with each other. In some approaches, the transducer positioning system 110 may also be configured to adjust the angle of incidence of one or more of the acoustic transducer 106 or near-surface sensor 134.

[0048] In some embodiments, the controller 108 may cause the transducer positioning system 110 to move the acoustic transducer 106 or near-surface sensor 134 adjacent to multiple zones within the subject 102. In this manner, the transducer positioning system 110 positions the acoustic transducer 106 or near-surface sensor 134 such that the acoustic transducer 106 or near-surface sensor 134 can interrogate multiple zones within the subject 102.

[0049] In some embodiments, controller 108 is in communication with a manufacturing system 111, for example, via network 140. Manufacturing system 111 may be configured to perform a manufacturing process or a portion thereof, such as a manufacturing process used to fabricate subject 102 or a component similar to the subject. Manufacturing system 111 includes one or more machines 113 for fabricating, processing, handling, and / or manipulating subject 102 or a component similar to (e.g., of the same type as) subject 102. Machine 113 may be operably coupled to controller 108.

[0050] The machine 113 may be a device used to move the specimens 102 from an inspection area associated with the manufacturing process to another portion (e.g., a station) of the manufacturing system 111. For example, the machine 113 may be a robot, a conveyor belt, or the like. In some examples, the machine 113 is a device used to separate or handle the specimens 102. The specimens 102 may be part of a batch of components produced through a manufacturing process, and the machine 113 may be used to sample or physically move the specimens 102 for testing via the inspection system 100. The machine 113 may also be a device that can be used to move or separate a batch of components with which the specimens 102 are associated. In this manner, the machine 113 may be used to physically pick or separate the batch of components. The machine 113 may also be used to physically pick or separate the specimens 102 from the batch of components.

[0051] The machine 113 may also be a device for processing the subject 102, for example, for removing a portion of the subject 102 in which an MTR is detected (e.g., an affected portion of the subject 102). For example, the machine 113 may be a band saw or similar tool that is electronically controlled by the controller 108.

[0052] The machine 113 may be an actuator or other device used to control the manufacturing system 111 or a portion thereof. The machine 113 may be operably coupled to the controller 108 such that the controller 108 can adjust one or more parameters of the manufacturing system 111.

[0053] The database 112 may be in communication with the inspection device 104 and / or the transducer positioning system 110. The database 112 may include any data collected or used by the inspection system 100. In some examples, the database 112 includes data collected, generated, and / or otherwise received by the acoustic transducer 106, the near-surface sensor 134, the transducer positioning system 110, and / or the controller 108. In one example, the database 112 includes signal data 130, such as acoustic waves generated and / or received by the acoustic transducer 106. In another example, the database 112 includes data, such as position data 132 recorded by the transducer positioning system 110. The database 112 need not be a single database, but may include one or more databases. In some examples, the database 112 includes signal data 130 and position data 132.

[0054] The signal data 130 may be referred to herein as waveforms or signal data 130 and may include any data related to characteristics of a sensor signal representing a recorded signal, such as an acoustic wave, an electromagnetic wave, a sensor voltage, or a sensor impedance. The waveform may be from a rectified or unrectified signal. The waveform may be from an impulse excitation, a single excitation frequency, multiple frequencies, or a gradually varying frequency spectrum, such as a chirp. The signal data 130 may include any characteristic of an analog or digitized signal. In some examples, the characteristics of the signal data 130 include at least one of a transit time, a signal amplitude, a voltage, an impedance, or the like. In some examples, the signal data 130 may be used to generate one or more intensity maps or histograms generated by the approaches described herein.

[0055] The position data 132 includes data indicating the location on the subject 102 where the data was acquired. In some examples, the position data 132 indicates the location on the subject 102 where the signal data 130 was acquired.

[0056] In some examples, the database 112 is separate from the controller 108 and / or the inspection device 104. In other examples, the database 112 is part of the controller 108 and / or the inspection device 104. In some embodiments, the controller 108 performs analysis of the signal data 130 without storing the signal data 130. For example, the processor of the controller 108 may have access only to the signal data 130 in RAM and can perform analysis of the data in real time.

[0057] In some embodiments, one or more components of the inspection system 100, including the acoustic transducer 106, the controller 108, the transducer positioning system 110, the near-surface sensor 134, and the database 112, communicate information to and from each other over a network 140. The network 140 may be any suitable communication network, such as, for example, a LAN, a WAN, the Internet, cellular, Wi-Fi, and other such communication networks, or a combination of two or more of such networks. The network 140 may also include wired communication.

[0058] In some embodiments, during operation, the controller 108 operates the acoustic transducer 106 to transmit acoustic waves that propagate within a volume of the subject 102 or a portion thereof. The controller 108 then receives data indicative of the acoustic waves from either the acoustic transducer 106 or the database 112. The controller 108 then detects or determines an MTR characteristic of the subject 102 based on the signal data.

[0059] Referring to FIG. 1B, a method 170 of operating the controller 108 (FIG. 1A) is shown. In block 172, the controller 108 causes the acoustic transducer 106 (FIG. 1A) to propagate acoustic waves into a volume of the subject 102 (FIG. 1A). In method 170, the acoustic transducer 106 comprises a transducer array that propagates acoustic waves into the subject 102. Transmitted, reflected, and / or scattered waves from the acoustic transducer 106 are received by the acoustic transducer 106 and / or the acoustic receiver 116 (FIG. 1A), which then output or generate a signal that is received by the controller 108. In block 174, the controller 108 then acquires signal data. In block 176, based on the signal data, the controller 108 determines at least one MTR score indicative of MTRs present within the subsurface volume of the subject 102 based on the signal data.

[0060] 1C, a method 180 for operating the controller 108 (FIG. 1A) is shown. In block 182, the controller 108 causes the inspection device 104 (e.g., the acoustic transducer 106 (FIG. 1A) and / or the near-surface sensor 134) (FIG. 1A) to transmit inspection energy (e.g., acoustic waves) to the object 102 (FIG. 1A) at two or more locations in time, space, or frequency. The controller may cause the inspection device 104 to propagate the inspection energy into the incident surface of the object 102 (FIG. 1A) and throughout the volume of the object 102. In block 184, the controller 108 acquires signal data (e.g., acoustic signals) indicative of acoustic waves scattered or reflected from the inspection device 104 (e.g., the acoustic transducer 106) toward the incident surface of the object 102 at two or more locations in time, space, or frequency. In block 186, the controller 108 analyzes the signal data indicative of the acoustic waves using at least one of the methods described with reference to FIGS.

[0061] In some embodiments, the controller 108 is configured to determine or calculate a quantified description of two or more acoustic signals relative to one another. In some examples, the quantified description includes at least one of wavelet analysis, statistical properties of the signal data, information entropy determination for the subject, joint probability matrix or co-occurrence matrix, phase coherence between features of waveforms representing acoustic waves, comparison of similarity between elements of a transducer array, comparison between waveforms acquired through full matrix capture, recognition and cataloging of regions within the subject, and / or patterns of signal envelopes representing acoustic waves. In some examples, the quantified description includes at least one of a Hilbert transform, a continuous wavelet transform, a Fourier transform, or a short-time Fourier transform.

[0062] In block 188, the controller 108 determines at least one response score indicative of MTR characteristics present within the subsurface volume of the subject 102 based on the analysis results.

[0063] Referring now to FIG. 1D , a method 190 for operating the controller 108 ( FIG. 1A ) is shown. In block 192, the controller 108 determines or generates one or more response scores and / or one or more intensity maps for the subject 102 ( FIG. 1A ). The controller 108 determines or generates the response scores and / or intensity maps using or based on signal data. The signal data may be acquired via one or more of the approaches or sensor configurations described herein. FIG. 24 details an exemplary method for determining a response score for the subject 102 based on signal data. The signal data may be acquired using one or more of the transducer arrangement configurations shown in FIGS. 2-10 . The response scores and / or intensity maps may be acquired or generated using one or more of the analysis methods described herein, including those shown and described with reference to FIGS. 11-27 .

[0064] In block 194, the controller 108 generates or determines one or more MTR scores based on the response scores and / or intensity map. For example, the response scores and / or intensity map may be correlated to an MTR score that represents the ground truth MTR content or level of the specimen 102. As used herein, ground truth MTR content or level refers to an MTR level from direct observation that reflects the actual MTR content or level, such as from measurements via EBSD. The MTR score is indicative of or predictive of the level of subsurface MTR present in the specimen 102.

[0065] In block 196, the controller 108 treats the subject 102 based on the one or more response scores and / or intensity maps and the correlation with the one or more response scores. For example, the controller 108 may decide to accept or reject the subject based on the MTR score. In some approaches, the controller 108 may not determine the MTR score (e.g., block 194 may be omitted) but may decide to accept or reject the subject based on the response score and / or intensity map. The controller 108 may also assign a material grade to the subject 102 based on the signal, the response score and / or intensity map, or the MTR score. For example, the controller 108 may assign a material grade that serves as an industry specification for the inspection of a particular material, such as titanium (Ti). It is contemplated that materials may be graded into multiple classes. These classes may reflect different design requirements.

[0066] In some embodiments, the accept / reject decision and / or material grade for the subject 102 is output on the user interface 128 (FIG. 1A) of the controller 108. For example, the accept / reject decision and / or material grade may be output on a screen of the user interface 128 to display the results to an operator. In some examples, the controller 108 may assign an accept or reject identifier to the subject 102 based on at least one MTR score and / or based on the accept / reject decision.

[0067] In other embodiments, the controller 108 adjusts, operates, or controls the manufacturing system 111 (FIG. 1A) to take action based on the material grade treatment for the subject 102. In some examples, the controller 108 controls, operates, or adjusts one or more machines 113 of the manufacturing system 111. It is contemplated that the controller 108 may also cause the manufacturing system 111 to take action based on the response score and / or intensity map or based on the MTR score. sensor

[0068] 2-10, various embodiments of sensor arrangements are illustrated that may be used in the inspection device 104 of the inspection system 100 of FIG. 1A. Any one or combination of sensor arrangements, including the options shown in FIGS. 2-10, may be selected to be consistent with the inspection systems and methods described herein. Volume Sensor Configuration

[0069] 2-7B include sensor arrangements for volumetric sensors including acoustic transducers. While FIGS. 2-7B illustrate transducer arrangements for cylindrical objects, it is contemplated that objects to be inspected using the inspection systems and transducer arrangements described herein may have any shape and / or geometry. FIGS. 2, 3, and 5 illustrate transducer arrangements that use a pulse-echo transducer configuration to interrogate an object using scattered or reflected acoustic waves. As used herein, a pulse-echo transducer configuration refers to a transducer configuration in which one or more transducers or transducer arrays are used to emit acoustic waves (e.g., pulses) and receive or detect acoustic waves (e.g., echoes) scattered or reflected from the object being inspected.

[0070] 4A, 4B, 6A, 6B, 7A, and 7B illustrate exemplary transducer arrangements that use a pitch-catch transducer configuration to interrogate a subject using transmitted acoustic waves. As used herein, a pitch-catch transducer configuration refers to a transducer configuration in which two transducers or transducer arrays are used, one acting as a transmitter (e.g., pitch) and the other acting as a receiver (e.g., catch). In a pitch-catch transducer configuration, the transmitter emits acoustic waves that propagate through at least a portion of the subject and are then detected by the receiver.

[0071] FIG. 2 illustrates a sensor arrangement 200 according to a first embodiment. The sensor arrangement 200 may be implemented as the acoustic transducer 106 shown and described with reference to FIG. 1A. The sensor arrangement 200 uses a transducer array 266 in a pulse-echo configuration to interrogate a subsurface volume of a subject 202. In some examples, the subject 202 is the subject 102 described with reference to FIG. 1A. The sensor arrangement 200 utilizes acoustic waves scattered and / or reflected from the subject 202 to detect MTR characteristics within the subsurface volume of the subject 202. The sensor arrangement 200 utilizes the interplay between MTR and reflective properties such as reflectivity and scattering.

[0072] The transducer array 266 includes a plurality of acoustic transducers 206. Each of the plurality of acoustic transducers 206 may act as an acoustic transmitter and an acoustic receiver. The transducer array 266 may be implemented as the acoustic transducer 106 of FIG. 1A. As such, the transducer array 266 may be in operative communication with the controller 108 (FIG. 1A). The transducer array 266 is configured to transmit and receive acoustic waves 260 propagating through a volume of the subject 202.

[0073] In some examples, the transducer array 266 is a linear array in which the acoustic transducers 206 are arranged in a line. In other examples, the transducer array 266 is a matrix array in which the acoustic transducers 206 have a two-dimensional arrangement of transducer elements.

[0074] 2, the subject 202 has a geometry that defines a longitudinal direction X and a transverse direction Y that extends perpendicular to the longitudinal direction X. The transducer array 266 is oriented along the longitudinal direction X of the subject 202, although it is contemplated that the transducer array 266 may be positioned in other orientations depending on the geometry of the subject 202. For example, in some embodiments, the transducer array 266 may be oriented along the transverse direction Y.

[0075] In operation, the controller 108 can be configured to cause the transducer array 266 to propagate acoustic waves into a volume of the subject 202. The controller 108 then acquires signal data of the acoustic waves from the transducer array 266. The controller 108 then determines at least one MTR characteristic present within the subsurface volume of the subject 202 based on the signal data.

[0076] The controller 108 may be configured to implement one or more of the signal analysis techniques described herein to analyze the signal data received from the transducer array 266 .

[0077] FIG. 3 illustrates a sensor arrangement 300 according to a second embodiment. The sensor arrangement 300 may be implemented as the acoustic transducer 106 shown and described with reference to FIG. 1A. The sensor arrangement 300 uses the transducer 306 in a pulse-echo configuration to interrogate a subsurface volume of a subject 302. In some examples, the subject 302 is the subject 102 described with reference to FIG. 1A. The sensor arrangement 300 utilizes acoustic waves scattered and / or reflected from the subject 302 to detect MTR characteristics within the subsurface volume of the subject 302. The sensor arrangement 300 utilizes the interplay between MTR and reflective properties such as reflectivity and scattering.

[0078] The transducer 306 acts as an acoustic transmitter and an acoustic receiver. In some embodiments, such as the embodiment depicted in FIG. 3, the transducer 306 is a single-element transducer that acts as both an acoustic transmitter and an acoustic receiver. In some embodiments, the transducer 306 is a transducer array including multiple elements that act as both an acoustic transmitter and an acoustic receiver. The transducer 306 may be implemented as the acoustic transducer 106 of FIG. 1A. As such, the transducer 306 may be in operative communication with the controller 108 (FIG. 1A). The transducer 306 is configured to transmit and receive acoustic waves propagating through the volume of the subject 302. In some configurations, the transducer 306 is positioned at a zero-degree angle of incidence, propagating longitudinal waves through the volume of the subject 302.

[0079] In operation, the controller 108 may be configured to cause the transducer 306 to propagate acoustic waves 360 into a subsurface volume of the subject 302. The controller 108 then acquires signal data from the transducer 306. The controller 108 may analyze the signal data and, based on the analysis, determine at least one MTR characteristic present within the subsurface volume of the subject 302.

[0080] 4A-4B illustrate a sensor arrangement 400 according to a third embodiment. The sensor arrangement 400 may be implemented as the acoustic transducer 106 shown and described with reference to FIG. 1A. The sensor arrangement 400 uses a pitch-catch configuration of an acoustic transmitter 414 and an acoustic receiver array 416 to interrogate a subsurface volume of a subject 402. In some examples, the subject 402 is the subject 102 described with reference to FIG. 1A. The sensor arrangement 400 utilizes acoustic waves 460 scattered and / or reflected from the subject 402 to detect MTR characteristics within the subsurface volume of the subject 402. The sensor arrangement 400 utilizes the interplay between MTR and reflective properties such as reflectivity and scattering.

[0081] The acoustic transmitter 414 and the acoustic receiver array 416 may be implemented as the acoustic transducer 106 of Figure 1A. As such, the acoustic transmitter 414 and the acoustic receiver array 416 may be in operative communication with the controller 108 (Figure 1A).

[0082] The acoustic transmitter 414 is configured to propagate acoustic waves 460 into a subsurface volume of the subject 402. In some embodiments, the acoustic transmitter 414 includes a single acoustic transmitter element. In other embodiments, the acoustic transmitter 414 is a transducer array including multiple acoustic transmitter elements.

[0083] The acoustic receiver array 416 is configured to receive the acoustic waves 460 transmitted by the acoustic transmitter 414. The acoustic receiver array 416 is disposed on the same side of the subject 402 as the acoustic transmitter 414. In this manner, the acoustic receiver array 416 can receive the acoustic waves 460 scattered or reflected from the subject 402. In some embodiments, the acoustic receiver array 416 can be an acoustic receiver including a single acoustic receiver element.

[0084] 4A and 4B, the subject 402 has a geometry that defines a longitudinal direction X and a transverse direction Y that extends perpendicular to the longitudinal direction X. The acoustic transmitter 414 and the acoustic receiver array 416 are aligned on the same side of the subject 402 and oriented along the longitudinal direction X, although it is contemplated that the acoustic transmitter 414 and the acoustic receiver array 416 may be positioned in other orientations depending on the geometry of the subject 402. For example, in some embodiments, the acoustic transmitter 414 and the acoustic receiver array 416 may be oriented along the transverse direction Y.

[0085] In operation, the controller 108 may be configured to cause the acoustic transmitter 414 to propagate acoustic waves 460 into a volume of the subject 402. The acoustic transmitter 414 may cause the acoustic waves 460 to propagate through the volume of the subject 402 and reach the acoustic receiver array 416 via at least one of scattering or refraction. The controller 108 acquires signal data from the acoustic receiver array 416. The controller 108 then determines at least one MTR characteristic present within the subsurface volume of the subject 402 based on the signal data.

[0086] 4B, individual MTRs 462 present within the volume of the subject 402 backscatter acoustic waves 460 to the surface of the subject 402. The scattered or reflected acoustic waves 460 are received or detected by the acoustic receiver array 416.

[0087] FIG. 5 illustrates a sensor arrangement 500 according to a fourth embodiment. The sensor arrangement 500 may be implemented as the acoustic transducer 106 shown and described with reference to FIG. 1A. The sensor arrangement 500 uses a transducer 506 in a pulse-echo configuration to interrogate a subsurface volume of a subject 502 at various angles of incidence 570. In some examples, the subject 502 is the subject 102 described with reference to FIG. 1A. The sensor arrangement 500 utilizes acoustic waves scattered and / or reflected from the subject 502 to detect MTR characteristics within the subsurface volume of the subject 502. The sensor arrangement 500 utilizes the interplay between MTR and reflective properties such as reflectivity and scattering.

[0088] The transducer 506 acts as an acoustic transmitter and an acoustic receiver. In some embodiments, such as the embodiment depicted in FIG. 5, the transducer 506 is a single-element transducer that acts as both an acoustic transmitter and an acoustic receiver. In some embodiments, the transducer 506 is a transducer array including multiple elements that act as both an acoustic transmitter and an acoustic receiver. The transducer 506 may be implemented as the acoustic transducer 106 of FIG. 1A. As such, the transducer 506 may be in operative communication with the controller 108 (FIG. 1A). The transducer 506 is configured to transmit and receive acoustic waves propagating through a volume of the subject 502. In some approaches, the transducer 506 is configured to transmit at least one of longitudinal waves, shear waves, or mixed-mode acoustic waves.

[0089] The transducer 506 is configured to transmit acoustic waves at an angle of incidence 570. The angle of incidence 570 is an angle defined with respect to a normal 568 to the surface of incidence of the object 502. The angle of incidence 570 is the angle between the incident acoustic wave 560 and the normal 568 to the surface of incidence. As used herein, the surface of incidence refers to the surface through which the acoustic wave passes as it enters the object 502. In some approaches, the transducer 506 may transmit acoustic waves at angles of incidence from greater than 0 degrees to about 40 degrees.

[0090] In operation, the controller 108 may be configured to cause the transducer 506 to propagate acoustic waves into a subsurface volume of the object 502. The controller 108 then acquires signal data from the transducer 506. The controller 108 may analyze the signal data and, based on the analysis, determine at least one MTR characteristic present within the subsurface volume of the object 502.

[0091] 6A-6B illustrate a sensor arrangement 600 according to another embodiment. The sensor arrangement 600 may be implemented as the acoustic transducer 106 shown and described with reference to FIG. 1A. The sensor arrangement 600 uses a pitch-catch configuration of an acoustic transmitter 614 and an acoustic receiver array 616 to interrogate a subsurface volume of a subject 602. The sensor arrangement 600 utilizes acoustic waves 660 scattered from and / or propagated through the subject 602 to detect MTR characteristics within the subsurface volume of the subject 602. The sensor arrangement 600 utilizes the interplay between MTR and transmission characteristics, such as variations in attenuation and scattering, in addition to variations in velocity and refraction.

[0092] 6A-6B show end views of a subject 602. While the subject 602 is depicted as having a circular cross-sectional area, it is contemplated that the sensor arrangement 600 may be implemented with a subject 602 having any cross-sectional shape when viewed from an end view. In some examples, the subject 602 is the subject 102 described with reference to FIG. 1A. The subject 602 has a first side 602-1 (e.g., covering a first half of the subject 602) and a second side 602-2 (e.g., covering a second half of the subject 602). The first side 602-1 is spaced apart from and faces the second side 602-2 of the subject 602.

[0093] The acoustic transmitter 614 is disposed on a first side 602-1 of the subject 602. The acoustic transmitter 614 is configured to propagate acoustic waves 660 into a subsurface volume of the subject 602. In some embodiments, the acoustic transmitter 614 includes a single acoustic transmitter element. In other embodiments, the acoustic transmitter 614 is a transducer array including multiple acoustic transmitter elements.

[0094] The acoustic receiver array 616 is disposed opposite the acoustic transmitter 614. In particular, the acoustic receiver array 616 is disposed on the second side 602-2 of the object 602. The acoustic receiver array 616 is configured to receive the acoustic waves 660 transmitted by the acoustic transmitter 614. In this manner, the acoustic receiver array 616 can receive the acoustic waves 660 that have propagated through the object 602. In some embodiments, the acoustic receiver array 616 is an acoustic receiver that includes a single acoustic receiver element.

[0095] The acoustic transmitter 614 and the acoustic receiver array 616 may be implemented as the acoustic transducer 106 of Figure 1A. As such, the acoustic transmitter 614 and the acoustic receiver array 616 may be in operative communication with the controller 108 (Figure 1A).

[0096] 6A-6B, the subject 602 has a geometry that defines a longitudinal direction X (not shown) and a transverse direction Y that extends perpendicular to the longitudinal direction X. The acoustic transmitter 614 transmits acoustic waves along the transverse direction Y, although it is contemplated that the acoustic transmitter 614 and acoustic receiver array 616 may be positioned, depending on the geometry of the subject 602, to propagate waves in other directions into the subject 602. For example, in some embodiments, the acoustic transmitter 614 and acoustic receiver array 616 may be positioned to transmit acoustic waves along the longitudinal direction X.

[0097] In operation, the controller 108 may be configured to cause the acoustic transmitter 614 to propagate acoustic waves 660 through a volume of the subject 602. The acoustic transmitter 614 may cause the acoustic waves 660 to propagate through the volume of the subject 602 via transmission in at least one of a longitudinal direction X (not shown) and a transverse direction Y from the acoustic transmitter 614 to the acoustic receiver array 616. The controller 108 obtains signal data for the acoustic waves 660 from the acoustic receiver array 616. The controller 108 then determines at least one response score for the subject 602 based on the signal data for the acoustic waves 660.

[0098] In some embodiments, the controller 108 may cause the acoustic transmitter 614 to propagate the acoustic waves 660 into multiple zones within the subject 602. For example, the controller 108 may be in operative communication with a transducer positioning system, such as the transducer positioning system 110 described with reference to FIG. 1A. The transducer positioning system 110 may move one or more of the acoustic transmitter 614 or the acoustic receiver array 616 to position the acoustic transmitter 614 and / or the acoustic receiver array 616 adjacent to one or more zones of the sample.

[0099] 6A-6B illustrate variations in the positioning of an acoustic transmitter 614 and an acoustic receiver array 616 relative to a subject 602 to interrogate different zones of the subject 602. FIG.

[0100] In Figure 6A, the acoustic transmitter 614 and acoustic receiver array 616 are positioned at normal incidence corresponding to a first depth 662A such that the acoustic wave 660 traverses a first inspection zone 664A of the object 602. Thus, the acoustic transmitter 614 can interrogate multiple different inspection zones 664, including, but not limited to, a first inspection zone 664A and a second inspection zone 664B. Because the object 602 is circular in shape, the first inspection zone 664A is a first chord of the object 602. In Figure 6B, the acoustic transmitter 614 and acoustic receiver array 616 are positioned at a second angle of incidence corresponding to a second depth 662B such that the acoustic wave 660 traverses a second inspection zone 664B of the object 602. Because the object 602 is circular in shape, the second inspection zone 664B is a chord. Selection of the angle of incidence 668B can place the second inspection zone 664B at any desired depth or acoustic path. Selection of the angle of incidence 668B can also be used to select an acoustic wave mode, such as a longitudinal wave or a shear wave.

[0101] In some approaches, the controller 108 is configured to analyze the signal data received from the acoustic receiver array 616 using statistical characteristics of the signal data. Amplitude is one example of a characteristic of the signal data that is analyzed. Due to the pitch-catch configuration of the acoustic transmitter 614 and the acoustic receiver array 616, the amplitude is the transmitted amplitude.

[0102] 7A-7B illustrate a sensor arrangement 700 according to a sixth embodiment. The sensor arrangement 700 may be implemented as the acoustic transducer 106 shown and described with reference to FIG. 1A. The sensor arrangement 700 uses a pitch-catch configuration of an acoustic transmitter 714 and an acoustic receiver array 716 to interrogate a subsurface volume of a subject 702. In some examples, the subject 702 is the subject 102 described with reference to FIG. 1A. The sensor arrangement 700 utilizes acoustic waves 760 scattered and / or reflected from the subject 702 to detect MTR characteristics within the subsurface volume of the subject 702. The sensor arrangement 700 utilizes the interplay between MTR and transmission characteristics such as variations in velocity and refraction.

[0103] The acoustic transmitter 714 and the acoustic receiver array 716 may be implemented as the acoustic transducer 106 of Figure 1A. As such, the acoustic transmitter 714 and the acoustic receiver array 716 may be in operative communication with the controller 108 (Figure 1A).

[0104] The acoustic transmitter 714 is configured to propagate acoustic waves 760 into a subsurface volume of the subject 702. In some embodiments, the acoustic transmitter 714 includes a single acoustic transmitter element. In other embodiments, the acoustic transmitter 714 is a transducer array including multiple acoustic transmitter elements.

[0105] The acoustic receiver array 716 is configured to receive acoustic waves 760 transmitted by the acoustic transmitter 714. The acoustic receiver array 716 is disposed substantially perpendicular to the acoustic transmitter 714. In this manner, the acoustic receiver array 716 can receive acoustic waves 760 that have propagated through the subject 702. In some embodiments, the acoustic receiver array 716 can be an acoustic receiver including a single acoustic receiver element.

[0106] 7A and 7B, the subject 702 has a geometry that defines a longitudinal direction X and a transverse direction Y that extends perpendicular to the longitudinal direction X. Although the acoustic transmitter 714 transmits acoustic waves along the longitudinal direction X, it is contemplated that the acoustic transmitter 714 and the acoustic receiver array 716 may be positioned, depending on the geometry of the subject 702, to propagate waves in other directions into the subject 702. For example, in some embodiments, the acoustic transmitter 714 and the acoustic receiver array 716 may be positioned to transmit acoustic waves along the transverse direction Y.

[0107] In operation, the controller 108 can be configured to cause the acoustic transmitter 714 to propagate acoustic waves 760 into a volume of the subject 702. The acoustic transmitter 714 can cause the acoustic waves 760 to propagate through the volume of the subject 702 and via transmission to the acoustic receiver array 716. The controller 108 obtains signal data from the acoustic receiver array 716. The controller 108 then determines at least one MTR characteristic present within the subsurface volume of the subject 702 based on the signal data.

[0108] 7B , individual MTRs 762 present within the volume of the object 702 refract some of the acoustic waves 760. For example, the acoustic waves 760 enter the object 702 in a primary beam 764, propagate through the object 702, and the MTRs 762 cause secondary beams 766 to be refracted away from the primary beam 764. The refracted acoustic waves in the secondary beams 766 are received or detected by the acoustic receiver array 716. The acoustic receiver array 716 may receive the acoustic waves in the secondary beams 766 and / or the primary beams 764.

[0109] Transducer placement configurations, including the options shown in Figures 2-7B, can be selected to match existing inspection configurations used for other purposes, such as inspecting for fusion-related anomalies, cracks, or porosity in multiple types of materials. For example, the approach shown in Figure 2 can be used to detect porosity or hard alpha inclusions in a titanium billet, or the approach shown in Figure 5 can be used to detect porosity in a nickel billet. Such approaches can be used in conjunction with the analytical methods listed herein to accomplish multiple purposes with a single scan. This approach allows for simultaneous MTR inspections without additional testing. This approach also allows for the avoidance of transducer or instrument reconfiguration between inspections.

[0110] Figure 8 illustrates an interaction volume 810 that is interrogated by inspecting energy transmitted or generated by a sensor arrangement that includes an inspection device 804 for interrogating a subject 802. The inspection device 804 may be implemented as the inspection device 104 of Figure 1A. In Figure 8, the inspection device 804 is placed on or above the subject 802.

[0111] The inspection device 804 includes one or more excitation sources 806 and one or more signal receivers 808. In some configurations, a single sensor may serve as both the excitation source 806 and the signal receiver 808. The excitation source 806 may be any device operable to introduce inspection energy into the subject 802 or a portion thereof. In some examples, the excitation source 806 is a transducer including a piezoelectric element configured to generate acoustic waves. In other examples, the excitation source 806 includes a laser configured to generate acoustic waves within the subject. In some examples, the excitation source 806 includes one or more coils configured to generate induced currents, such as eddy currents.

[0112] In some embodiments, the excitation source 806 includes a transmit sensor array of one or more active elements (e.g., acoustic transducers, coils) configured to transmit, generate, or otherwise introduce test energy into the subject 802. The transmit sensor array can be a linear array or a matrix array. In some configurations, the transmit sensor array is an acoustic transducer array. The acoustic transducer array can include multiple acoustic transducers as active elements acting as acoustic transmitters and receivers. When configured in this manner, the acoustic transducer array transmits and receives acoustic waves as the test energy. In other embodiments, the transmit sensor array is an eddy current sensor array.

[0113] The excitation source 806 generates one or more of acoustic waves, laser pulses, electromagnetic waves, and / or induced currents within the interaction volume 810 of the object under test 802, and the signal receiver 808 returns corresponding signal data to the inspection system 100 (FIG. 1A). The excitation for the acoustic waves may be in the frequency range of 1 MHz to 100 MHz using pulses, chirps, tone bursts, or coded excitation. The excitation for the induced current may be in the frequency range of 100 KHz to 50 MHz using single frequencies, multiple frequencies, or frequency sweeps, such as pulses, chirps, tone bursts, or coded excitation.

[0114] In some embodiments, the signal receiver 808 includes a receiving sensor array of one or more elements configured to receive test energy from the object under test 802. The test energy may be scattered, reflected, or otherwise transmitted or transferred from the object under test 802 to the receiving sensor array. In some configurations, the receiving sensor array is an acoustic transducer array. In other embodiments, the receiving sensor array is an eddy current transducer array.

[0115] The interaction volume 810 refers to the volume of material that generates a discernible change in the sensor response. While both surface and volumetric inspection of an object have an interaction volume 810, it should be understood that the interaction volume 810 in a surface inspection approach is much smaller and adjacent to the surface of the object being inspected. For example, volumetric inspection of an object may reach a greater penetration depth using the object's volume than surface inspection methods. In an exemplary Rayleigh wave inspection, the interaction volume 810 may generally reach a depth of less than two wavelengths within the object 802. In an example of eddy current inspection, the interaction volume is on the order of two times the skin depth. The skin depth may represent the effective penetration depth (e.g., effective inspection depth) of eddy currents within a material.

[0116] The inspection device 804 may actively or passively steer or preferentially orient the interaction volume 810. In one example of passive orientation, the excitation source 806 is shaped to generate an anisotropic interaction volume 810 due to the geometry of the excitation source 806. In one example of active orientation, the excitation source is mounted on a rotating fixture that rotates relative to the orientation of the subject 802, resulting in a rotation of the interaction volume 810. In another example of active orientation, multiple excitation sources 806, such as an eddy current array or a transducer phased array, are excited by the inspection system 100 to adjust, steer, or otherwise shape the interaction volume 810 based on electrical input to the inspection device 804.

[0117] Although FIG. 8 illustrates a sensor arrangement for a flat specimen, it is contemplated that the specimen to be inspected using the inspection systems and sensor arrangements described herein may have any shape and / or geometry. Near-Surface Sensor Configuration

[0118] 9A-9T, examples of sensor arrangements for the inspection device 104 of FIG. 1A are shown. In particular, FIGS. 9A-9T show exemplary sensor arrangements for near-surface sensors. The inspection device 104 of FIG. 1A may be comprised of any combination of one or more of the devices shown in FIGS. 9A-9T configured with the inspection system 100 to work separately or in combination. It is further contemplated that any one or combination of the inspection systems and transducer arrangements described in U.S. Non-Provisional Application No. 18 / 678,918, the entirety of which is incorporated herein by reference, may be used.

[0119] Referring particularly to FIG. 9A , in one example, the inspection device may include a phased array pitch-catch configuration configured to generate Rayleigh surface waves. In such a configuration, multiple transducers are arranged in a specific pattern, including one or more acoustic transmitters 904 and one or more acoustic receivers 906. The transducers may be single elements or arrays of multiple elements, such as a linear array. The acoustic transmitter 904 is positioned at a first location relative to the subject 902 through a coupling medium and configured at a position and angle sufficient to generate and transmit acoustic surface waves, such as Rayleigh surface waves, across the surface or near-surface volume of the subject 902. The acoustic receiver 906 is positioned at a second location relative to the subject 902 and configured to simultaneously receive the acoustic surface waves. The acoustic transmitter 904 may be separated from the acoustic receiver 906 by any suitable distance.

[0120] 9B and 9C, in another example, an inspection device may include a single-element transducer 910 that generates a spherically focused wave on the object under test. FIG. 9C is a bottom end view of the single-element transducer 910 of FIG. 9B. The transducer 910 includes a concave surface 912 and an active driven element 908. In some approaches, a lens may be positioned between the driven element and the concave surface 912. In other approaches, the driven element 908 itself may be placed on the concave surface 912. The radius of curvature of the concave surface 912 may be selected to achieve a particular wave incidence angle 982 that results in surface acoustic waves (Rayleigh waves) propagating across the surface of the object under test. In some approaches, the acoustic wavefront incidence angle 982 ranges from greater than 0 degrees to approximately 40 degrees. The acoustic waves are then received by an acoustic receiver portion of the transducer 910.

[0121] In some examples, concave surface 912 may further include one or more acoustic transmitters spaced apart from one or more acoustic receivers. In some examples, the acoustic transmitter and acoustic receiver are separate elements. In other examples, the acoustic transmitter and acoustic receiver are a single piezoelectric element, and acoustic waves emitted by one portion of the element may be received on another portion of concave surface 912.

[0122] 9D, 9E, 9F, and 9G, the inspection device 104 may include a single element transducer that generates preferentially oriented Rayleigh surface waves.

[0123] In Figures 9D and 9E, a transducer 914 includes an element 916 that is shaped to generate preferentially oriented Rayleigh waves. Figure 9E is a bottom end view of the transducer 914 of Figure 9D. The shaped element 916 may include a single-element strip, which may be curved in one or more directions.

[0124] 9F, 9G, and 9H, an inspection device 950 includes a transducer 952 including a spherical element with a mask 965 that shapes the Rayleigh waves. The mask 965 may include slits that define openings to provide paths through which the acoustic waves propagate. The mask 965 may be constructed of foam or other material that provides sufficient sound insulation. The shape of the mask 965 may be tailored to optimize inspection performance for a particular feature size. The mask 965 may be adjustable, removable, or permanently attached.

[0125] Although the transducer 952 is cylindrical, in other embodiments, the transducer 952 may have other suitable shapes. The transducer 952 has a concave surface 956. The transducer 952 is mounted adjacent to the surface of the subject 960. The concave surface 956 of the transducer 952 is disposed at an end of the transducer 952. The concave surface 956 is spaced a distance 966 from the surface of the subject 960. The concave surface 956 may be a hemispherical or non-hemispherical surface surrounded by a periphery 959. In one example, the non-hemispherical surface forming the concave surface 956 may be a curved, concave rectangular strip. Such a strip may or may not have curved corners. The periphery 959 of the concave surface 956 includes an acoustic transmitter portion 957 and an acoustic receiver portion 958. The acoustic transmitter portion 957 and the acoustic receiver portion 958 form a transmission gap 962 therebetween. The acoustic transmitter portion 957 faces the acoustic receiver portion 958. In this embodiment, the acoustic transmitter portion 957 and the acoustic receiver portion 958 are part of a single piezoelectric device.

[0126] Concave surface 956 is the active element of transducer 952 and is a single piezoelectric element that operates as both an acoustic transmitter and an acoustic receiver. The piezoelectric element can operate in a transmission mode, in which the transducer operates to activate the piezoelectric element (e.g., concave surface 956) to generate acoustic waves. The piezoelectric element can also operate in a reception (or listen) mode, in which the transducer waits for acoustic waves to reflect off the surface of the object 960 and bounce back onto the piezoelectric element (e.g., concave surface 956). When electrical energy is applied to concave surface 956, it expands and contracts, causing vibration of concave surface 956. The vibration of concave surface 956 results in acoustic waves propagating from the acoustic transmitter portion 957 of transducer 952 to the object 960. The thickness of concave surface 956 (e.g., the active element) can determine the frequency of the acoustic waves. The acoustic waves propagate across the surface of the object 960. The acoustic waves are reflected back off the concave surface 956 and return to the acoustic receiver portion 958 of the transducer 952 .

[0127] Because the concave surface 956 is a piezoelectric material, when an acoustic wave reaches the concave surface 956, the motion is converted back into an electrical signal. When the acoustic wave is ultrasonic, the electrical signal may be referred to as an ultrasonic waveform. The transducer 952 may emit the acoustic waves at a particular pulse repetition rate. The pulses (e.g., acoustic wave transmissions) may be spaced so that the acoustic wave has enough time between pulses to reach the subject 960 and return to the transducer 952 before the next pulse is generated. It is also contemplated that the acoustic transmitter portion 957 may act as an acoustic receiver and the acoustic receiver portion 958 may act as an acoustic transmitter, such that the acoustic wave may propagate across the subject 960 in opposing or two directions between the acoustic transmitter portion 957 and the acoustic receiver portion 958.

[0128] The acoustic transmitter portion 957 generates or transmits acoustic waves. As shown, the acoustic waves generated by the acoustic transmitter portion 957 of the transducer 952 strike the surface of the subject 960 at an angle of incidence 964. The angle of incidence 964 is such that the acoustic waves propagate across the surface of the subject 960 across the propagation gap 962. In some approaches, the angle of incidence is within a range of about 20 degrees to about 40 degrees. In other approaches, the angle of incidence is within a range of about 25 degrees to about 35 degrees. In yet other approaches, the angle of incidence is within a range of about 28 degrees to about 32 degrees, and in some embodiments, the angle of incidence is about 30 degrees. The radius of curvature of the concave surface 956 can be selected to achieve a particular angle of incidence that focuses the acoustic energy and results in surface acoustic waves (Rayleigh waves) propagating across the surface of the subject 960. The radius of curvature of the concave surface 956 corresponds to the angle of incidence, and therefore it is contemplated that the aforementioned ranges for the angle of incidence also apply to the radius of curvature. For example, the radius of curvature may be from about 20 degrees to about 40 degrees.

[0129] The propagation gap 962 is the distance that the acoustic wave propagates across the surface of the object 960. The acoustic wave propagates across the propagation gap 962 and is received by the acoustic receiver portion 958 of the transducer 952. The propagation time of the acoustic wave across the propagation gap 962 is affected by properties such as the microstructure of the object 960. For example, the size and orientation of the crystalline grain structure can affect the propagation time of the acoustic wave across the propagation gap 962.

[0130] The inspection device 950 further includes a coupling medium (not shown). In some approaches, the space surrounding the propagation gap 962, the object under test 960, and the concave surface 956 is filled with the coupling medium. In some approaches, the entire inspection device 950 is immersed in the coupling medium. In some approaches, a first volume 961A, which immerses the entire concave surface 956, is filled with the coupling medium. The first volume 961A includes the volume below line Z in FIG. 9F. In other approaches, a second volume 961B, which encompasses the space between the acoustic transmitter portion 957 and the acoustic receiver portion 958 through which the acoustic wave may propagate between the transducer 952 and the surface of the object under test 960, is filled with the coupling medium. The second volume 961B includes the volume of the space between the acoustic transmitter portion 957 and the acoustic receiver portion 958 and the volume between the acoustic transmitter portion 957 and the acoustic receiver portion 958 and the object under test 960.

[0131] The bottom end of transducer 952 includes a concave surface 956. A periphery 959 of concave surface 956 is an annular edge to which a mask 965 is coupled. A portion of periphery 959 of concave surface 956 is blocked by mask 965. Mask 965 includes slits 967 ( FIG. 9G ) that form a path from acoustic transmitter portion 957 to acoustic receiver portion 958. Slit 967 may help to guide or direct acoustic waves from acoustic transmitter portion 957 to acoustic receiver portion 958. In some approaches, mask 965 may be omitted by not including a portion of the piezoelectric active element that is covered by mask 965.

[0132] 9H is a top view of a portion of the upper surface of the object 960 of FIG. 9F. FIG. 9H illustrates various acoustic path angles for acoustic waves transmitted and received by the inspection device 950 when the mask 965 is not present. As shown, the inspection device 950 can perform scans at a number of different acoustic path angles. As illustrated, the inspection device 950 performs four scans, including a scan at a 0-degree acoustic path angle, a scan at a 45-degree acoustic path angle, a scan at a 90-degree acoustic path angle, and a scan at a 135-degree acoustic path angle. The direction between the acoustic transmitter portion 957 and the acoustic receiver portion 958 is used to determine or set the acoustic path angle. It should be understood that the particular acoustic path angles illustrated are not limiting, and the inspection device 950 can perform scans at any combination of angles and any number of different angles.

[0133] 9I and 9J illustrate an inspection sensor 970 having an alternative configuration to that shown in FIGS. 9F, 9G, and 9H. FIG. 9I presents a cross-sectional side view of the inspection device. The inspection sensor 970 is substantially similar to the inspection device 950 shown in FIGS. 9F, 9G, and 9H, except that the acoustic transmitter portion 976 and the acoustic receiver portion 978 on the concave surface 974 of the transducer 927 are separate piezoelectric elements. In contrast to FIGS. 9F, 9G, and 9H, the transducer 972 includes the acoustic transmitter portion 976 and the acoustic receiver portion 978 within a single piezoelectric element. The acoustic transmitter portion 976 may include one or more acoustic transmitters. Similarly, the acoustic receiver portion 978 may include one or more acoustic receivers.

[0134] The inspection sensor 970 further includes a sensor 979 to indicate the position of the transducer 972 relative to the object 980. The sensor 979 may sense, detect, or otherwise determine a distance 984 between the concave surface 974 and the surface of the object 980. The sensor 979 may be a piezoelectric element that acts as an acoustic transmitter and an acoustic receiver. Thus, the sensor 979 may transmit acoustic waves to the object 980 and receive waves reflected from the object 980. The travel time of the acoustic waves to propagate to the object 980 and back to the sensor 979, as well as the speed of sound through a coupling medium such as water, may be used to determine the distance 984. In some embodiments, the sensor 979 may also be used to verify that the object 980 is level with the transducer 972. For example, the sensor 979 may be used to measure the distance 984 between the transducer 972 and the object 980. The level of the object 980 may be adjusted until the distance 984 is substantially the same at different points across the surface of the object 980 or is otherwise substantially the same at multiple points along the object 980. To achieve this, the level of the object 980 may be adjusted until the propagation time is substantially the same at different points along the surface of the object 980. It is contemplated that the surface of the object 980 is substantially flat.

[0135] The distance 984 can be used to determine a propagation distance over which the acoustic wave propagates between the acoustic transmitter portion 976 and the acoustic receiver portion 978. The propagation distance is exemplified by the sum of a first distance 977A, a propagation gap 986, and a second distance 977C. The first distance 977A extends between the acoustic transmitter portion 976 and the surface of the object 980. The second distance 977C extends between the surface of the object 980 and the acoustic receiver portion 978. In some approaches, the propagation distance can be determined via one or more geometric formulas when the distance 984 is known. The propagation gap 986 can also be determined via one or more geometric formulas when the distance 984 is known. The propagation gap 986, along with the propagation time, is used to determine the speed of the surface acoustic wave propagating across the surface of the object 980.

[0136] Figure 9J is a bottom end view of the transducer 972 of Figure 9I. The bottom end of the transducer 972 includes a concave surface 974. The concave surface 974 does not include a mask. Figure 9J illustrates an acoustic transmitter portion 976, which in this embodiment includes an array of acoustic transmitters, and an acoustic receiver portion 978, which includes an array of acoustic receivers. The acoustic transmitter portion 976 and the acoustic receiver portion 978 are positioned on opposite sides of the concave surface 974. It is also contemplated that the acoustic transmitter portion 976 may act as an acoustic receiver and the acoustic receiver portion 978 may act as an acoustic transmitter, such that acoustic waves may propagate across the subject 980 simultaneously in opposite directions.

[0137] 9K, in another example, an inspection device may include a transducer 924 that generates high-angle shear waves to perform near-surface inspection. The transducer 924 generates the high-angle shear waves by propagating sound into the object under test at an angle of incidence close to the second critical angle. This angle of incidence causes the shear waves inside the sample to refract at an angle nearly parallel to the surface of the sample. This angle of incidence may be generated by orienting the central ray of the transducer 924 at a preferred angle of incidence, as in a pulse-echo method, or it may be generated by orienting the central ray normal to the surface, as in a pitch-catch method, with the lens curvature of the transducer 924 creating peripheral rays at the preferred angle of incidence.

[0138] 9L, the inspection device may include one or more lasers 940 that can generate acoustic waves on or near the surface of the specimen. The one or more lasers 940 may emit a pulsed laser beam 941 to generate acoustic waves on or near the surface of the specimen.

[0139] 9M-9T show examples of eddy current sensors that can be used to generate induced currents near a specimen surface, which can be used for the near-surface inspection method presented herein. Near-surface inspection can be accomplished by a near-surface inspection device. The near-surface inspection device can include any combination of sensors described herein, such as a combination of an absolute coil and two opposing D-shaped differential coils. All eddy current sensors shown herein can be reinforced with a magnetic core and / or a magnetic shield. In some examples, the magnetic core can be made of ferrite. All eddy current sensors described herein can also be combined with an electric field sensor or a magnetic field sensor, such as a magnetoresistive sensor. All eddy current sensors described herein can be controlled and multiplexed via any pattern of driver and receiver coils. Furthermore, the eddy current sensors described herein can be a single coil acting as both an excitation source and a detector (e.g., a receiver), or can include separate coils for excitation and detection.

[0140] 9M and 9N, the eddy current sensor may include an absolute coil 925. The absolute coil 925 may include or be a single coil that generates a circumferentially symmetric induced current field.

[0141] 9O and 9P, the eddy current sensor can include a differential coil. In a differential coil, the measured response is based on the change in signal between a first coil 928 and a second coil 930. In some embodiments, the shape of the first coil 928 and the second coil 930 can resemble the letter "D." In other embodiments, the first coil 928 and the second coil 930 can differ in terms of aspect ratio or other geometric parameters.

[0142] 9Q and 9R, an eddy current sensor can include a "racetrack" coil that generates an anisotropic induced current field within a subject. The "racetrack" coil 932 has an elongated oval shape to vary the spatial distribution of the induced current. The racetrack configuration can also be combined with a differential coil.

[0143] 9S and 9T, an eddy current sensor may include an eddy current array 934 that combines multiple printed circuits to form a customizable coil. The eddy current array 934 may include a flexible array. The eddy current array 934 includes multiple eddy current coils or wires 938 arranged on a printed or flexible circuit board 936. The multiple eddy current coils or wires 938 may be multiplexed to control the distribution of induced currents within the subject. In some configurations, the coils or wires 938 may directionally steer the induced currents within the subject.

[0144] Any one or combination of inspection system and transducer arrangement configurations, including the options shown in Figures 9A-9T, may be selected to match the inspection configuration used in conjunction with the inspection methods or signal data analysis methods described herein, including the methods shown and described with reference to Figures 11-27.

[0145] Any one or combination of inspection system and transducer arrangement configurations, including the options shown in Figures 9A-9T, may be selected to match the inspection configuration used in the methods of Figures 1B-1D and described below. Near-surface and volumetric sensor configurations

[0146] 10 shows sensor arrangements using near-surface sensor configurations and volume acoustic transducer configurations to compare near-surface and bulk measurements. Near-surface measurements can be performed using Rayleigh surface waves generated by an ultrasonic transducer or using induced currents generated by a coil. Bulk measurements can be performed using one or more of the transducer arrangements described herein.

[0147] The sensor arrangement may include one or more volumetric ultrasound transducers 1010 arranged at different angular positions relative to the subject. The sensor arrangement may further include one or more surface wave ultrasound transducers or eddy current sensors 1012 arranged at different angular positions relative to the subject 1014. Chirp Excitation

[0148] It is contemplated that any of the above-described sensor arrangements of Figures 2-10 may use a time-varying excitation signal, which may be used to extend the frequency content of the signal data. In one example, because the ultrasonic response of acoustic waves to particle features is frequency-dependent, inspecting an object under test over a wider frequency range may provide improved inspection capabilities. Advantageously, a time-varying excitation may provide a wider frequency range than the square-wave excitation commonly used in industry today, and the varying frequency content may enable frequency-based analysis methods to identify one or more MTR characteristics of the object under test. Also advantageously, because the ultrasonic response to features is based on the wavelength of the wave, a time-varying excitation may be sensitive to a wider range of MTR characteristics than a conventional square-wave excitation.

[0149] Additionally, it is contemplated that a longer time-varying excitation may enable examination of larger subjects and / or deeper regions within the subject.

[0150] In one approach, the controller 108 of FIG. 1A generates a time-varying excitation signal and applies it to the inspection device 104 (FIG. 1A). The time-varying excitation signal may be a chirp, a tone burst, a Gaussian pulse, or other frequency-modulated signal. In some embodiments, a linear sweep chirp is used, starting from a low frequency f0 and sweeping to a higher frequency f1. In one example, the low frequency may be in the range of 0.1 MHz to 2 MHz, and the high frequency may be in the range of 10 MHz to 20 MHz. In some examples, the time-varying signal may be a logarithmic chirp.

[0151] In other embodiments, coded excitation may be used, in which amplitude and / or frequency modulated excitation is applied in a particular pattern, such as a Golay code, a Barker code, or a convolution of a Barker code and a Golay code. Advantageously, modulating the excitation signal allows the controller to search the signal data for the excitation pattern, which may allow for improved signal-to-noise ratio (SNR) when examining a subject.

[0152] As described above, and consistent with chirp excitation, in some embodiments, the inspection device 104 may include a single transducer. In other embodiments, the inspection device 104 may include multiple transducer elements connected in one or more phased arrays. When multiple elements are used, the controller 108 may vary the time-varying excitation signal for each element individually.

[0153] In some embodiments, the excitation signal may have a constant voltage, while in other embodiments, the excitation may be amplitude modulated, such as Hanning modulated.

[0154] The transducer responds to the time-varying excitation by emitting ultrasound waves having a frequency spectrum similar to that of the excitation signal, hi some embodiments, the characteristics of the excitation signal are selected to optimize the characteristics of the emitted pulses.

[0155] The frequency-varying acoustic waves interact with the subject, and the controller 108 records the waveform response.

[0156] In some embodiments, the time-frequency response of the back surface can be used as an intensity metric representing one or more MTR characteristics.

[0157] In other embodiments, the time-frequency response of the backscattered signal may be analyzed using any of the methods herein. Analysis method

[0158] 11-27, various signal analysis methods are illustrated that may be implemented by the controller 108 of FIG. 1A. The signal analysis methods are used to analyze signal data and may be implemented by any one of the sensor arrangements described above. It is contemplated that any of the signal analysis methods, or combinations thereof, may be used in conjunction with any of the sensor arrangements described with reference to FIGS. 2-10. Frequency Shift Method

[0159] FIG. 11A illustrates a method 1100 for analyzing signal data acquired using inspection system 100 (FIG. 1A) based on frequency response, according to some embodiments. Method 1100 is a frequency shift-based method for classifying subjects based on MTR severity and the presence of localized MTR. Method 1100 may be implemented by a controller, such as controller 108 as shown in FIG. 1A. Any one or combination of inspection system and sensor placement configurations, including the options shown in FIGS. 1-10, may be selected to match the inspection configuration used for method 1100.

[0160] Method 1100 utilizes the changes in frequency distribution or composition that can appear in an acoustic wave interacting with an MTR relative to the incident wave. Particles and particle clusters (MTRs) in metals and alloys scatter and attenuate ultrasonic energy. When ultrasonic waves strike a particle or MTR, the latter acts as a point source, causing vibration and the generation of waves detectable by the transducer. The resulting scattered and reflected waves, referred to herein as backscatter, convey information about the scatterer. The center frequency of the backscattered signal may match the center frequency of the incident wave (the center frequency of the transducer). However, due to the inertial properties of particles and MTRs, changes in the frequency distribution or composition (e.g., broadening and shifting of the frequency spectrum) in the backscattered signal may occur. Thus, depending on their size, MTRs may cause amplitude spikes at frequencies other than the center frequency of the transducer. Large MTRs downshift the scattered wave frequency to a greater extent than small MTRs. In other words, apart from the shift in center frequency, subpeaks may occur, thereby indicating the presence of MTRs of various sizes.

[0161] Another approach to assessing MTR severity may calculate the crystallographic orientation of a feature of interest based on the dominant frequency of the signal reflected from such feature. The intent of method 1100 is not to perform such characterization of a specific feature, but rather to enumerate features that exhibit a frequency shift away from an expected center frequency. Yet another approach uses one dominant frequency from a segmented fast Fourier transform to calculate the ultrasonic velocity, from which the crystallographic orientation of the reflecting feature is calculated. The approach in method 1100 selects one or more peaks whose amplitude exceeds a fixed percentage of the peak amplitude and uses that percentage of occurrence to classify the material as acceptable or unacceptable.

[0162] The method 1100 can be used to calculate a shift in the frequency distribution in the frequency characteristics of the signal data. The shift in the frequency distribution in the frequency characteristics can be determined between two or more configurations of the subject 102 (FIG. 1A). The change or shift in the frequency distribution can include any change or shift in the magnitude of a frequency spectrum derived from or associated with the signal data. In some embodiments, such a change or shift is quantified by a change in the peaks in the frequency spectrum.

[0163] Method 1100 can be used to analyze the specimen 102 referenced in FIG. 1A. In some examples, the specimen analyzed via method 1100 includes a billet, specifically a titanium or nickel billet. In some examples, the specimen can be an end slice of a billet, a macro-slice of a forged part (e.g., a fan disk), the surface of a forged part, or any other suitable sample taken from a billet, forging, or part before or after any step of thermomechanical processing of a metal alloy. In some approaches, the specimen can also be a part or component of a gas turbine engine (e.g., a part or component before assembly or a part or component disassembled from a gas turbine engine). Any one or combination of inspection system and transducer arrangement configurations, including the options shown in FIGS. 2-10, can be selected to match the inspection configuration used for method 1100.

[0164] At block 1102, signal data indicative of signal energy received from the subject 102 is received. The signal data may be received from the subject 102 at two or more locations on the subject 102. The signal data may represent acoustic waves propagating through a volume of the subject 102. As used herein, signal data may include backscattered signals of incident acoustic waves. The acoustic waves propagate between the front and rear walls of the subject 102. In some approaches, the signal data may be collected by one or more transducers in a pulse-echo or pitch-catch configuration and represent acoustic waves scattered or reflected from the subject 102. The controller 108 may be configured to collect and / or receive the signal data.

[0165] At block 1104, a subset of the signal data is selected. The subset of the signal data may be a window of an appropriate width. The window may include a spatial domain, a time domain, or a combination thereof. The window may further include a window of signal data corresponding to the incident wave and / or a window of signal data corresponding to the backscattered wave.

[0166] At block 1106, a short-time Fast Fourier Transform (FFT) is calculated on the signal data (e.g., a subset or windowed segment). The short-time FFT may be used to determine a power spectrum using a subset or window of the signal data. Various windows may be used for the short-time Fourier transform, such as a box window, a Gaussian window, or a discrete prolate spheroidal sequence window. Other suitable windows may also be used for the short-time Fourier transform. The FFT converts the spatial or time domain signal data into the frequency domain. The FFT also separates frequency characteristics from phase characteristics of the signal data. Characteristics of the FFT result are analyzed within a selected region to determine the power spectrum. The controller 108 may be configured to calculate the Fast Fourier Transform.

[0167] In block 1108, the FFT results are analyzed. The FFT results are analyzed to determine and identify amplitude spikes that are within a threshold of the peak frequency. In some embodiments, the threshold is equal to a fraction of the highest peak value detected. The amplitude spikes may appear at frequencies other than the center frequency of the transducer. The controller 108 may be configured to detect and identify the amplitude spikes using a peak detection algorithm. For example, FIG. 11B illustrates a power spectrum of an incident wave having an amplitude at the center frequency of the transducer, and FIG. 11C illustrates an example of a power spectrum with amplitude spikes at frequencies other than the center frequency of the transducer.

[0168] A shift in frequency distribution is then determined for the signal data between the two or more configurations based on the frequency characteristics of the signal energy. As used herein, frequency distribution refers to the spectrum of energy for a given frequency in a signal. The change or shift in frequency distribution may be measured by a change in the magnitude of the frequency spectrum. In some embodiments, the change may be quantified by a change in the peaks in the frequency spectrum. Blocks 1102 through 1116 provide one exemplary approach for determining a shift in frequency distribution between two or more configurations of the subject 102.

[0169] In block 1110, frequencies corresponding to amplitude spikes are selected and divided into predetermined frequency bins. In some examples, the frequency bins may include bins encompassing a 1 MHz window, such as (1-2 MHz), (2-3 MHz), (3-4 MHz), (4-5 MHz), (5-6 MHz), ... (Xn-1 to Xn). In some examples, the frequency bins include frequencies within a range of 1 MHz to 2 MHz upshift or downshift from the center frequency of the transmit transducer used to acquire the signal data. In some examples, the frequency bins are defined before acquiring the signal data. In some examples, the frequency bins are defined after acquiring the signal data, before dividing the frequencies corresponding to the amplitude spikes. The controller 108 may be used to implement this approach. The controller 108 may be configured to create frequency bins, determine frequencies corresponding to amplitude spikes, and match each respective frequency to a respective frequency bin.

[0170] The steps described in blocks 1102-1110 may be repeated for all scans within a cross section and / or for all cross sections of the subject 102. In other words, the selected window is moved (or slid) to the next set of signal data. The amount by which the window moves (e.g., hop size) may be selected based on a desired time or frequency resolution. The FFT results from each window are analyzed, and frequencies corresponding to amplitude spikes are selected and matched within the respective frequency bins.

[0171] In block 1112, after repeating the steps described in blocks 1102-1108, the number of points in each frequency bin is determined for multiple configurations or positions of the analyzed subject 102. In this manner, the number of MTRs (corresponding to amplitude spikes) can be identified and counted. The controller can be configured to count the total number of points in each frequency bin.

[0172] At block 1112, the number of points in each frequency bin for each cross section is sorted. In some examples, the number of points is sorted in ascending order. In some examples, the number of points is sorted in descending order. In some embodiments, a statistical distribution, such as a cumulative distribution function, may be generated from these points. The controller is configured to sort the number of points in each frequency bin for each cross section. In some examples, the controller may be configured to plot the number of points in each frequency bin in ascending order across the length of the subject 102.

[0173] In block 1114, a mathematical transform is applied to the values ​​from the different frequency bins (corresponding to the number of points). In some examples, the mathematical transform may include addition, subtraction, multiplication, or the like. The mathematical transform may be any suitable transform to maximize the difference between cross sections of a subject or between two or more subjects. The controller may be configured to apply the mathematical transform to the values ​​from each frequency bin.

[0174] In block 1116, the subject 102 may be classified based on the results of the mathematical transformation. The subject 102 may be classified, for example, by determining at least one response score for the subject 102. In some examples, a suitable statistic is calculated from the bin values, such as a median, quartile, etc., and this statistic is used to classify the subject 102. The controller 108 may be configured to classify the subject 102 based on the mathematical transformation. In this manner, the subject 102 may be classified based on high or low MTR content based on the number and / or density of amplitude spikes. In addition, the location of the MTR may be identified using position information related to the amplitude peaks.

[0175] In some embodiments, an intensity map of the subject 102 is generated based on the analysis of the signal data. The intensity map may also be used to characterize the subject 102 based on, for example, MTR content. The intensity map may be used to determine an MTR score for the subject 102. In some examples, the intensity map is generated or quantified using an Otsu threshold. In some examples, the intensity map is generated or quantified using a Gaussian mixture. In some examples, the intensity map is based on a continuous wavelet transform or a short-time Fourier transform over a specific time period. In some examples, the intensity map is generated based on upshifting or downshifting frequency components of the signal data.

[0176] In some embodiments, the intensity map is quantified to determine a shift in the frequency distribution of signal energy. In some examples, the intensity map is quantified at least in part based on enumeration of spots that exhibit changes in frequency distribution below or above a threshold. In some examples, the intensity map is quantified using a statistical measure of the size of regions with uniform frequency shifts. In some examples, the intensity map is quantified using a cumulative distribution function of the size of regions with uniform frequency shifts.

[0177] In some embodiments, the transmit sensor array is configured to steer the direction of the interrogation energy so that the upshift or downshift of the frequency distribution can be compared to other steered angles.

[0178] In some embodiments, one or more frequency peaks in the signal data are quantified excluding a primary frequency peak, which may be selected to have an amplitude above a fixed percentage of the magnitude of the primary frequency peak to quantify the frequency peak occurrence rate.

[0179] In some embodiments, a spectrogram of the frequency distribution of the signal energy is formed.

[0180] In some embodiments, the intensity map is compared to a reference intensity map and a difference or variance score between the intensity map and the reference intensity map is calculated. spectrogram

[0181] FIG. 12 illustrates a method 1200 for analyzing signal data using time-frequency analysis to generate an intensity map representing the interaction of test energy with the subject 102 .

[0182] As used herein, a spectrogram is understood to be a multidimensional data set in which one dimension is the frequency spectrum, a second axis is at least one of position and time, and the third dimension is a measure of signal intensity. A spectrogram may include multiple spatial dimensions corresponding to the dimensions of the object.

[0183] At block 1202, a spectrogram (e.g., a time-frequency spectrogram) is generated from the signal data. In some embodiments, the spectrogram is generated by a fast Fourier transform, a short-time Fourier transform, a continuous wavelet transform, a discrete wavelet transform, a continuous chirplet transform, a chirp-Z transform, a fractional Fourier transform, or a Gabor transform. It is contemplated that any signal method that computes the variation of a signal in both time and frequency may be used.

[0184] In some examples, a variety of wavelets may be used for the continuous wavelet transform, including both orthogonal and non-orthogonal wavelets, such as Morlet wavelets, Mexican hat wavelets, generalized Morlet wavelets, etc. It is contemplated that other forms of wavelets may be used.

[0185] In some embodiments, synchrosqueezed versions of the Fast Fourier Transform, Short-Time Fourier Transform, Continuous Wavelet Transform, Discrete Wavelet Transform, Continuous Chirplet Transform, Chirp-Z Transform, Fractional Fourier Transform, or Gabor Transform are used. Advantageously, synchrosqueezed transforms may provide finer resolution of details of the frequency behavior of the signal data.

[0186] At block 1204, a similarity metric is selected to quantify the difference between the spectrograms. In some embodiments, the similarity metric is a function that takes two spectrogram matrices as input and returns a scalar value that represents the difference between the spectrograms. In other embodiments, the matrices are generated by subtracting one spectrogram from another, and the similarity function may only accept a single matrix as input. It is contemplated that the optimal selection of a similarity metric may depend on the characteristics of the sensor used and the manufacturing conditions of the analyte.

[0187] In some examples, Pearson's correlation coefficient is used as the spectrogram similarity metric, while in other examples the selected similarity metric is one of mean squared error, peak signal to noise ratio, feature similarity index, or structural similarity index.

[0188] In another example, cross-correlation and / or cross-coherence of spectrograms may be used as the similarity metric. In yet another additional example, time-frequency ridges, which represent patterns in time-frequency space, are extracted from spectrograms and the similarity of the ridge curves is compared.

[0189] In some approaches, metrics are generated to measure the area of ​​regions of uniform character in the spectrogram.

[0190] In block 1206, the intensity I is calculated by applying a selected similarity metric to two or more of the multiple time-frequency spectrograms. Comparing the two spectrograms may result in an intensity at a single point. Generating an intensity map may involve comparing spectrograms at multiple points on the object 102.

[0191] Alternatively, in block 1208, the intensity I may be calculated from the similarity between the measured spectrogram and a reference spectrogram. In some embodiments, the reference spectrogram may be obtained from a calibration standard of a neutral material, such as titanium alloy, powder metal alloy, and / or quartz, processed to have the desired MTR condition. In another example, the reference spectrogram may be obtained from a physics-based simulation of an ideal material. In another example, the reference spectrogram may be an average signal response from the subject 102.

[0192] In some examples, the intensity map is generated based on a structural similarity function of the signal data. In some examples, the intensity map is generated based on cross-coherence of the signal data. In some examples, the intensity map is generated based on mean square error of the signal data. In some examples, the intensity map is generated based on phase shifts in the signal data. In some examples, the intensity map is generated based on Pearson's correlation coefficient of the signal data. In some examples, the intensity map is generated by calculating features in the signal data having phase coherence and determining statistics of the features having phase coherence.

[0193] At block 1210, this process is repeated to generate a spatial intensity map representative of one or more MTR scores within a region of the subject 102, or one or more response scores representative of the overall MTR status of the subject 102. An MTR score indicative of the MTR present in the subject 102 may then be calculated for the subject 102 based on the response scores and / or the spatial intensity map. Phase Coherence

[0194] 13 illustrates a method 1300 for analyzing signal data. Method 1300 involves determining or calculating coherent features (e.g., phase coherence) within the signal data. In some examples, for signal data acquired from a single-element transducer, the phase coherence within the signal data is based on spatiotemporal coherence of the signal data within the subject and / or within a waveform. In other examples, for signal data acquired using an array transducer (e.g., a phased array), the phase coherence within the signal data is based on coherence of signals from multiple positions, times, and / or elements. The acoustic waves analyzed via method 1300 may be scattered or reflected acoustic waves.

[0195] Signal data representative of the interaction of the inspection energy with the volume of the subject is collected or received at block 1302. The controller 108 may be configured to collect or receive the signal data from the inspection device 104, for example.

[0196] The phase of the signal data is calculated at one or more positions, times, frequencies, or excitation modes at block 1304. The phase may be calculated via one or more of the argument of the complex impedance or voltage of the inspection device, a Hilbert transform, and / or application of a sign(x) function to the signal data.

[0197] At block 1306, the phase and / or phase shift of the signal data at one or more locations or times is compared.

[0198] At block 1308, regions where the phase of one region of the signal data aligns with the phase of one or more other signals (e.g., regions of phase coherence) are identified. The one or more other signals may be spatially and / or temporally separated from the waveform. In some examples, the controller 108 is configured to identify regions where the phase of one region of the signal aligns with the phase of one or more other signals. In one example, coherence in the signal data is based on regions in the samples having correlated phases. In one example, phase correlation is calculated using a sign(x) function (e.g., grouping by "up peak" and "down peak"). Such regions may be applied to signal data collected using single element and / or array transducers. In another example, coherence in the signal data may be envelope-based, for example, based on a Hilbert transform. In yet another example, coherence in the signal data may be based on sign(x) for distinct elements of an array transducer.

[0199] In one example, the signal data is an acoustic waveform, and the signal phase is calculated via the Sign function.

[0200]

number

[0201] teeth,

[0202]

number

[0203] can be calculated via where:

[0204]

number

[0205] is the position

[0206]

number

[0207] , signal data at time t and frequency f. The multiple coherence vectors are

[0208]

number

[0209] , t, or f. In some examples, the response score may be generated by summing the coherence vectors. In other examples, the local changes in the coherence vectors may be represented as an image. It is also contemplated that other metrics representing the similarity between multiple coherence vectors may be used to generate the response score. In other examples, the coherence vectors may be applied as a mask and / or filter to the signal data for use in other analysis methods described herein or in U.S. Non-Provisional Application No. 18 / 678,918.

[0210] In one example, the signal data is of an acoustic waveform generated from a Rayleigh wave. The signal data may be processed via a Hilbert transform to obtain the signal envelope.

[0211] In another example, the signal data is the complex impedance of an eddy current coil, the phase of the complex signal is calculated at multiple locations or times, and the phase coherence and / or change in phase coherence is determined at different locations, times, or frequencies.

[0212] At block 1310, one or more response scores and / or intensity maps are generated based on the regions of phase coherence. An MTR score indicative of the MTR present in the subject 102 may then be calculated for the subject 102 based on the response scores and / or spatial intensity maps. Instantaneous Phase

[0213] FIG. 14 illustrates a method 1400 for analyzing signal data based on local discontinuities in the phase of the signal data.

[0214] In block 1402, signal data representing an interaction between an inspection energy and the object 102 (FIG. 1A) is received. The signal data may represent an interaction between an acoustic wave and the object or an interaction between an induced electric field and the object. The inspection energy may be an acoustic wave or an induced electric field. The controller 108 (FIG. 1A) may be configured to collect or receive the signal data, for example, via the inspection device 104 (FIG. 1A). The signal data is collected at multiple regions of the object 102.

[0215] At block 1404, the instantaneous phase of the signal data is calculated. In some embodiments, the instantaneous phase is calculated from the argument of a complex-valued Hilbert transform. In other embodiments, the instantaneous phase may be calculated from a Taeger operator. In other embodiments, the instantaneous phase may be calculated from zero crossings of the signal data. The instantaneous phase of the signal data may be calculated at one or more positions, times, frequencies, or excitation modes. In some embodiments, the instantaneous phase is calculated based on the argument of a complex-valued impedance or voltage of the inspection device 104. In some embodiments, the instantaneous phase is calculated by applying a sign(x) function to the signal data. In some examples, the phase variation of the signal data is calculated at a specific time point. In some examples, the amplitude variation of the signal data is calculated at a specific time point.

[0216] In one example, the selected signal data may be an acoustic waveform generated from a Rayleigh wave. The signal data may be processed via a Hilbert transform to obtain the instantaneous phase. The Hilbert transform separates the amplitude and phase characteristics of the received signal energy captured by the signal data. It is also contemplated that other methods of calculating the instantaneous frequency of a signal may be used as a proxy for the instantaneous phase. In some embodiments, the instantaneous eddy current phase is determined when the test energy is electromagnetic energy.

[0217] At block 1406, regions having local phase discontinuities are identified. In some examples, the discontinuities are identified by unwrapping the phase into a linearly increasing signal. In other examples, the discontinuities may be identified by comparing the phase at nearby locations in time or space.

[0218] Note that the method illustrated in FIG. 13 computes phase-aligned regions, but the method identifies phase discontinuities.

[0219] In some embodiments, the phase data may be filtered to retain only regions where the phase discontinuity exceeds a size or magnitude threshold.

[0220] In block 1408, this process is repeated to generate a spatial intensity map representative of one or more MTR characteristics within a region of the subject, or one or more response scores representative of the subject's overall MTR state. One or more response scores and / or intensity maps may be generated by comparing instantaneous phase and / or changes in instantaneous phase at different locations, times, or frequencies. The phase signal may be combined or contrasted with other characteristics of the signal data, such as the signal envelope, time-of-flight peaks, amplitude peaks, or other signal characteristics described herein. An MTR score indicative of the MTR present in the subject 102 may then be calculated for the subject 102 based on the response scores and / or spatial intensity map. Nakagami

[0221] FIG. 15 illustrates a method 1500 for analyzing signal data by examining the distribution of values ​​within a magnitude envelope.

[0222] At block 1502, signal data indicative of a signal or test energy received from the object under test 102 (FIG. 1A) is received. The signal data may be collected using and / or received from the test device 104 (FIG. 1A).

[0223] At block 1504, a signal envelope may be calculated for the signal data. In some embodiments, the envelope is calculated via a Hilbert transform. The Hilbert transform may be performed to separate the amplitude and phase characteristics of the signal data. In another embodiment, a moving root-mean-square (RMS) filter may be used to calculate the envelope. In another embodiment, a spline fit may be used. In another embodiment, an infinite impulse response (IIR) or finite impulse response (FIR) filter may be used to calculate the envelope. It is contemplated that other methods of calculating the envelope of a signal may be used.

[0224] At block 1506, a distribution of values ​​for a given characteristic of the signal data is determined or calculated. The distribution of values ​​for the given characteristic may be calculated using a signal envelope of the signal data. The given characteristic of the signal energy may include at least one of amplitude, phase, frequency, or shape associated with the signal data, for example, associated with the envelope of the signal data. In some approaches, a distribution of envelope magnitudes associated with the envelope is calculated.

[0225] At block 1508, a quantified description of the distribution of values ​​for the given characteristic is determined or calculated. In one example, a parametric distribution may be numerically fitted to the magnitude distribution to obtain the fitted value. In one embodiment, a Nakagami distribution having shape and scale parameters is fitted. For example, the distribution of amplitude values ​​associated with the signal data may be extracted and fitted to at least one of a Nakagami distribution, a generalized Nakagami distribution, or a Rayleigh distribution (e.g., a parametric distribution) having a specific parametric description of the shape of the distribution. In another embodiment, a descriptive characteristic function is fitted to the shape of the distribution of values ​​for the given characteristic of the signal energy. In some embodiments, the slope of the shape of the distribution is measured.

[0226] In some embodiments, the parametric distribution may include one or more of a Nakagami distribution, a Rayleigh distribution, a gamma distribution, a homodyne K (HK) distribution, a generalized Nakagami distribution, or a Rice distribution.

[0227] In some embodiments, amplitude values ​​are extracted from the signal data and one or more higher order statistics are calculated. The higher order statistics may include one or more of skewness, kurtosis, etc. In such an approach, the amplitude values ​​are not limited to a particular model of the distribution function.

[0228] In block 1510, one or more of the fitted values ​​are selected as metrics of the intensity I corresponding to the location on the subject 102.

[0229] At block 1512, the calculation of I is repeated at one or more additional locations on the subject 102, thereby generating a spatial intensity map representative of one or more MTR scores in a region of the subject, or one or more response scores representative of the subject's overall MTR status. An MTR score indicative of the MTR present in the subject 102 may then be calculated for the subject 102 based on the response scores and / or the spatial intensity map. Shape Fitting

[0230] 16A illustrates a method 1600 for analyzing one or more sets of signal data to generate an intensity map that represents one or more MTR characteristics within the subject. The focusing and defocusing of acoustic waves is dependent on the geometry and material. As signal data is collected over increasing time intervals, the shape of the received signal can represent one or more MTR characteristics within the subject.

[0231] In block 1602, signals may be collected from longitudinal, shear, Rayleigh, or mixed-mode waves and may be generated and received using any of the inspection methods shown and described with reference to Figures 1-7B. In some approaches, scattered acoustic waveform data may be generated by the inspection system 100 and inspection device 104 of Figure 1A.

[0232] At block 1604, the signal data is pre-processed through one or more filtering and / or normalization processes. Pre-processing may involve a variety of operations.

[0233] In one example, the signal data may be pre-processed by performing an envelope calculation using one or more of a Hilbert transform, a spline fit, an FIR or IIR filter, or a moving RMS filter.

[0234] In some embodiments, the signal data may be averaged over multiple nearby locations within the subject.

[0235] In block 1606, the shape of the signal data is characterized and / or parameterized. In one example, various signal processing parameters may be used, including, but not limited to, peak height, time to peak, peak width, peak prominence, or integrated area. The signal processing parameters may further include scaling of the data to a specified range, one or more frequency-based filters, or one or more statistical-based filters.

[0236] At block 1608, the shape of the signal data (e.g., the shape of the waveform data and signal profile characteristics) is determined. The shape of the signal data may be quantified by or fitted to a shape-describing function. The shape-describing function may be a parametric function. The parametric function may include at least one of a polynomial function, an exponential function, a parametric function having an exponential term and a polynomial term, a parametric function associated with a probability density function (PDF) or cumulative distribution function (CDF) of a statistical distribution, or a parametric function associated with a physics-based model. The parametric function is fitted to match the signal data based on a numerical analytical method, including least squares, nonlinear least squares, and principal component regression. It is contemplated that any numerical analytical method for fitting a parameterized function to data may be used. The function may be fitted to any region of interest within the signal data. Characteristic shape parameters may be fitted to the signal data. The characteristic shape parameters may further include at least one of the following: signal peak height, signal peak position, signal peak width, signal peak prominence, integrated area of ​​the signal envelope, or integrated area of ​​a parametric function, to name a few. Quantifying waveform data and / or signal profile characteristics provides one approach to calculating a distribution of values ​​for a given characteristic of the signal data. In such an approach, the shape of the distribution is the given characteristic.

[0237] In some embodiments, the parametric function used to quantify the signal data may be a polynomial, an exponential function, a Gaussian function, or a nonlinear combination thereof. It is contemplated that in some examples, the parametric function used to quantify the signal data may take the form of a statistical distribution, such as a Lévy distribution, a lognormal distribution, a Weibull distribution, or similar distributions. The parametric function may also be related to the probability density function (PDF) or cumulative distribution function (CDF) of the statistical distribution, or to a physics-based model. It is also contemplated that a numerical model based on physical principles of acoustic interaction with particles or MTR features may be fitted in this regard. One or both of the approaches described in blocks 1606 and 1608 may be used to quantify the waveform shape.

[0238] Referring briefly to FIG. 16B, in one example, the preprocessed waveform data is fitted to a modified exponential function. Equation 1 and Equation 2 are exemplary equations that may be used to fit the signal data. Graph 1620 shown in FIG. 16C presents a graph illustrating preprocessed signal data 1622 fitted to a modified exponential equation 1624. Modified exponential equation 1624 may include Equations 1 and 2 as shown in FIG. 16B. Information regarding the MTR level in the subject may be derived from the fitted values ​​of one or more function parameters (a, b, c, d, etc.).

[0239] In another example, the data is fitted by a series of splines. In yet another example, a linear function is fitted to a domain of the waveform data.

[0240] In yet another example, waveform data is preprocessed by averaging over subregions of samples and applying a median filter to each signal. The preprocessed signal data is then element-wise exponentiated. A spatial map of the subject can then be generated from element-wise differences of cumulative sums of the preprocessed signal data. The spatial map can be used as a representation of one or more MTR characteristics of the subject or to generate one or more scores. The scores may be indicative of MTR regions at various locations on the subject.

[0241] In another example, the shape of the signal data (eg, waveform data) may be determined by calculating the area under the signal curve of the signal data.

[0242] In another example, the shape of signal data (eg, waveform data) may be determined by calculating the element-by-element difference of a running sum of the signal data raised to a specified exponent.

[0243] 16A, in block 1610, quantified signal shapes, including one or both approaches described in blocks 1606 and 1608, are compared spatially across the subject. Quantified waveform shape data is compared across multiple locations on the subject, which may represent individual locations / points, locally averaged positions, profiles of fit parameters across the subject, etc.

[0244] In one example, the quantified shape of the signal data may be used as an intensity value I. Intensities may be calculated at multiple locations on the subject to generate an intensity map representing one or more MTR scores.

[0245] In another example, a response score for the subject may be obtained from the quantified shape of the average signal waveform over a region of the subject. An MTR score indicative of the MTR present in the subject 102 may then be calculated for the subject 102 based on the response score and / or the spatial intensity map.

[0246] In block 1612, the subject may be treated or graded based on the value and distribution of the spatially quantified signal shape at a single location or across the entire subject. Template Cross-Correlation

[0247] FIG. 17 illustrates a method 1700 for analyzing signal data by quantifying correlation with a bank of one or more signal templates.

[0248] At block 1702, one or more signal templates are selected. The signal templates are selected to enable identification of one or more MTR characteristics. In some embodiments, the signal template is a time series representing a nominal material. In other examples, the signal template is a signal characteristic of an undesired material, such as the ultrasound signal response of a single MTR.

[0249] In other embodiments, the signal template may be obtained from a calibration standard of a neutral material, such as titanium alloy, powder metal alloy, and / or quartz, processed to have the desired MTR condition. In another example, the signal template may be obtained from a physics-based simulation of an ideal material. In another example, the signal template may be an average signal response from the subject 102. In another example, the signal template may be derived from an excitation waveform.

[0250] In block 1704, a cross-correlation between the signal data and the template is calculated. The cross-correlation may refer to a measure of similarity between two signals. The cross-correlation is based on a comparison between the signal data and the signal template. Optionally, the template may be used as a matched filter by time reversal before the cross-correlation. In some examples, the received signal data is correlated with the signal template, and a difference or variation score is calculated. In some aspects, a measure of the areal extent of the uniform received energy associated with the signal data is measured. Any function that makes the cross-correlation an impulse function may qualify as a matched filter, such as in pulse compression mode.

[0251] At block 1706, a second cross-correlation between the signal data and a second template may be calculated. The cross-correlation may optionally be repeated for another signal template. For example, the cross-correlation of the signal data may be repeated for multiple signal templates in a bank of signal templates.

[0252] At block 1708, a spatial intensity map or one or more response scores for the subject are generated based on the cross-correlation, and in some approaches, based on the second cross-correlation. For example, the signal template is cross-correlated with signal data from multiple regions of the subject to generate a spatial intensity map representative of one or more MTR characteristics within the regions of the subject, or aggregated into one or more response scores representative of the MTR state of the subject. An MTR score indicative of the MTR present in the subject 102 may then be calculated for the subject 102 based on the response scores and / or the spatial intensity map.

[0253] It should be noted that, unlike other correlation-based approaches, using template cross-correlation, signal data is not autocorrelated with time-lagged versions of signal data from the same location on the object under test, a technique used in the prior art. Advantageously, the approach herein does not require shear waves, whereas the prior art specifies only shear wave inspection. This allows the methods herein to be applied during longitudinal, shear, and / or Rayleigh wave inspection of the object under test. In the inspection of a round billet, for example, longitudinal waves may inspect a larger portion of the object under test compared to shear waves. Additionally, a rotating object under test is not required with this approach, and the object under test may be scanned at any orientation. Furthermore, advantageously, when calibration sample signals are included in the bank of signal templates, this may allow the inspection system 100 to operate with improved robustness in isolating the response of the MTR characteristics of the object under test 102. Signal Decomposition

[0254] FIG. 18 illustrates a method 1800 for analyzing signal data by decomposing the signal data into multiple components. The signal response to an acoustic wave may be governed by the fundamental frequency and bandwidth of the excitation transducer. Decomposing the signal data into multiple components may enable the inspection system 100 to isolate and remove signal features that do not correspond to MTR characteristics. As used herein, components are understood to refer to separate vibration modes of a signal, having different frequency spectra, time lags, and magnitudes.

[0255] In block 1802, signal data is received, for example, from the inspection device 104. The signal data represents an interaction between the object under test and inspection energy. In some approaches, the excitation energy is an acoustic wave. The acoustic wave may be a longitudinal wave, a shear wave, or a Rayleigh wave.

[0256] In block 1804, the signal data may be preprocessed to improve the results of the decomposition. In some embodiments, the preprocessing may include filtering, envelope calculation via a Hilbert transform, or FFT-based downsampling. In some embodiments, the signal data may be preprocessed using one or more of a Hilbert transform, a spline fit, an FIR filter, an IIR filter, or a moving RMS filter.

[0257] At block 1806, the signal data is separated into multiple components. The signal data may be separated into multiple separate components using one or more of empirical mode decomposition, variational mode decomposition, Hilbert-Huang transform, or multi-resolution mode decomposition. In some embodiments, such methods are used to decompose the signal data into components. It is contemplated that any other suitable method may be used to decompose the signal data into components. At block 1808, one or more of the multiple separate components may be post-processed before further analysis. Post-processing may include filtering, such as a median filter or a Gaussian filter. Additional filtering may include high-pass, low-pass, and band-pass filters. Post-processing may also include signal envelope calculation, a Hilbert transform, a moving RMS filter, spline fitting, an IIR filter, or an FIR filter.

[0258] In block 1810, one or more components that represent one or more MTR characteristics are identified or selected. In some embodiments, the selection of relevant signal components may be guided by comparison to a calibration specimen, such as a titanium alloy or powder metal alloy that has been processed to have a desired MTR condition. Components that are present in the specimen but not in the calibration specimen may represent an undesirable MTR characteristic.

[0259] In some embodiments, one or more intensity maps or response scores may be generated for the subject 102 based on one or more components indicative of the MTR. The intensity maps or response scores may then be correlated with one or more MTR scores indicative of the ground truth MTR present in the subject 102.

[0260] It is contemplated that in some embodiments, components representing one or more MTR characteristics may be used as input signal data for any of the additional analytical methods detailed herein. Advantageously, selecting specific signal components may improve the performance of the selected additional analytical method. Rolling A-Scan Dissimilarity

[0261] FIG. 19 illustrates a method 1900 for analyzing signal data based on local variations in the signal data.

[0262] In block 1902, signal data is received, for example, from the inspection device 104. The signal data represents an interaction between the object under test and inspection energy. In some approaches, the excitation energy is an acoustic wave. The signal data may correspond to a longitudinal wave, a shear wave, or a Rayleigh wave.

[0263] At block 1904, a window size is selected. The window size may include a spatial domain associated with the signal data, a time domain associated with the signal data, or a combination thereof. The selection of the window size may be influenced by the shape of the transducer beam interaction volume. The selection of the window size may also be influenced by the size and orientation of the MTR characteristic of the subject. In some embodiments, the signal data may be restricted to a subsequence within the focal zone of the transducer.

[0264] In one example, the window may be 31 pixels horizontally, 5 pixels vertically, and 1 time sample deep.

[0265] In another example, the window may be 3 pixels circumferentially, 3 pixels axially, and 3 time samples in depth.

[0266] At block 1906, the local mean and variance of the signal data are calculated for multiple positions, and in some embodiments, for all positions in the window.

[0267] At block 1908, the variance and mean are summed over multiple windows to obtain an intensity value. In some embodiments, this intensity value is calculated using the formula

[0268]

number

[0269] It is calculated from where:

[0270]

number

[0271] is the variance of the signal data within the window,

[0272]

number

[0273] is the mean of the signal data within the window region.

[0274] In some embodiments, this formula may be modified by weighting or non-linear transformations such as coefficient weights, power transformations, or logarithmic transformations.

[0275] At block 1910, the intensity calculation is repeated at one or more locations on the subject to generate an intensity map representative of at least one MTR characteristic. An MTR score indicative of the MTR present in the subject 102 may then be calculated for the subject 102 based on the spatial intensity map. Signal / Pattern Entropy

[0276] 20A illustrates a method 2000 for analyzing signal data based on the signal entropy of the test signal data. As used herein, signal entropy refers to a metric that quantifies the likelihood of observing similar patterns in the signal data, depending on both the characteristics of the patterns, the number of times each pattern is observed, and the degree to which the patterns match each other.

[0277] In block 2002, the signal data is optionally coarsened to improve the pattern matching algorithm. Coarsening may refer to a process by which signal data is modified or represented as a coarser or less detailed version. Advantageously, coarsening the signal may allow the inspection system 100 to optimize the inspection for a given size of the MTR feature. In some embodiments, the signal data may be coarsened by selecting every Nth sample from the original signal. In other embodiments, the signal may be coarsened by applying a moving average filter or a median filter to average adjacent sample points together. In other embodiments, the signal may be coarsened by FFT-based downsampling. It is contemplated that the coarsening may be adapted based on the size and depth of the microstructural features within the object.

[0278] In block 2004, the magnitude of the signal data is optionally re-digitized into N bins. The bins may be evenly or non-evenly distributed, such as on a linear or logarithmic scale. Advantageously, adjusting the number and distribution of bins allows additional control over pattern matching.

[0279] In block 2006, one or more pattern lengths l∈[m, m+b] are selected, where m and b are any positive integers. The signal data is divided into multiple subsequences of length l. In some embodiments, the signal subsequences are overlapping windows of length l. In other embodiments, the signal subsequences may be non-overlapping.

[0280] In block 2008, a membership function is selected to compare patterns between multiple subsequences. The membership function takes a subset S of the signal data and calculates whether it is a sufficiently similar pattern or a new pattern. The membership function may return a binary membership (1, 0) or a scalar value representing the amount of membership.

[0281] In some embodiments, multiple sets of signal data are acquired and membership functions may compare whether a pattern in one signal is found in the other signal. In one example, the membership function may compare patterns found in adjacent elements of a multi-element sensor array.

[0282] In one example, subsequences of signal data may be matched if the amplitude values ​​at each point in the subsequence match exactly. Because this approach matches only exactly equal values, careful selection of bin digitization and signal coarsening can have a strong effect on the results.

[0283] In another embodiment, permutation entropy is selected as the membership function, where two subsequences of signal data match if their amplitude values ​​match exactly at every point, but the membership function allows for multiple permutations of the order in which the subsequence values ​​appear.

[0284] In another embodiment, approximate entropy is selected as the membership function. Approximate entropy is a variation that applies a tolerance to membership using a scalar similarity based on the vector distance between two signal sequences. Two subsequences are considered a match if the vector distance between the two subsequences is less than a specified threshold.

[0285] In another embodiment, a gradient entropy function is selected as the membership function. In gradient entropy, two signal subsequences are considered a match based on the algebraic gradient between successive magnitudes of the signal data. In one example, all signal gradients between approximately 0 and 15 degrees may be considered identical. In another example, all signal gradients between approximately 45 and 80 degrees may be considered identical. Advantageously, gradient entropy can ignore small changes in signal magnitude but identify abrupt changes that may represent MTR characteristics within a subject. For a detailed description of one method for calculating the gradient entropy membership function, see Cuesta-Frau, D. Slope Entropy, "A New Time Series Complexity Estimator Based on Both Symbolic Patterns and Amplitude Information." Entropy 2019, 21, 1167.

[0286] In another embodiment, fuzzy entropy is selected as the membership function. Fuzzy entropy is a modification that calculates scalar similarity based on a function of the vector distance between two signal sequences rather than calculating binary membership. For an expanded description of the fuzzy entropy membership function, see W. Chen, Z. Wang, H. Xie, and W. Yu, "Characterization of Surface EMG Signal Based on Fuzzy Entropy," in IEEE Transactions on Neural Systems and Rehabilitation Engineering, Vol. 15, No. 2, pp. 266-272, June 2007, and H. Azami, P. Li, S. E. Arnold, J. Escudero, and A. Humeau-Heurtier, "Fuzzy Entropy Metrics for the Analysis of Biomedical Signals: Assessment and Comparison," in IEEE Access, Vol. 7, pp. 104833-104847, 2019. Advantageously, in fuzzy entropy, calculating similarity based on a function of vector distance may allow inspection system 100 to more precisely control the scalar similarity between subsequences.

[0287] In one example, the fuzzy entropy similarity may be calculated using a linear function that takes values ​​between 0 and 1, corresponding to a linear scaling of the distance between the subsequences.

[0288] In another example, fuzzy entropy similarity may be calculated using an exponential decay function such that subsequences with small vector distances are highly similar, but membership decreases rapidly as distance increases.

[0289] It is contemplated that other related membership functions, such as variance entropy, fluctuation variance entropy, sample entropy, cosine similarity entropy, or phase entropy, may be advantageously applied to examine MTR characteristics.

[0290] It is contemplated that any method that quantifies the likelihood of classifying and counting patterns in a time domain signal may be suitable as a membership function.

[0291] At block 2010, a membership function is calculated for each subsequence of the signal data. A distribution p, which represents the commonality of each pattern in the signal, may then be generated based on the membership function. In some embodiments, this distribution p is then calculated based on a scalar value φ, which represents the distribution of values. m are summed up to

[0292] In some embodiments, the weighting function g is optionally represented by the formula φ m =Σg(p i ) According to the formula φ m is applied to the probability distribution before summing to

[0293] In block 2012, a signal intensity is generated based on this distribution. In some embodiments, the signal intensity I is calculated using the formula I=-logφ m According to φ m It is generated from the negative logarithm of

[0294] In other embodiments, the signal entropy is compared for two pattern lengths, which may allow improved identification of one or more MTR characteristics. In such embodiments, the signal intensity I is calculated as a function of φ m+b and φ m the negative logarithm of the difference between I=-(logφ m+b -logφ m ) can be generated from

[0295] In other embodiments, the signal entropy may be compared for more than two pattern lengths. In such embodiments, the signal intensity I is calculated using the formula I=-(logφ m+b -logφ m + _) According to φ m , φ m+1 , φ m+b The logarithm of the difference between

[0296] The equation used in block 2012 is reproduced in Figure 20B.

[0297] The calculation of I is repeated at multiple locations throughout the subject to generate a spatial intensity map representative of one or more MTR characteristics within a region of the subject, or aggregated into one or more response scores representative of the MTR state of the subject. An MTR score indicative of the MTR present in the subject 102 may then be calculated for the subject 102 based on the response scores and / or the spatial intensity map. Signal peaks

[0298] 21 illustrates a method 2100 for analyzing signal data based on one or more spectral descriptors of the signal. This method requires a frequency spectrum, which may be obtained through any of the methods described herein. In some embodiments, method 2100, or portions thereof, are implemented or performed using inspection system 100 of FIG. 1A. For example, controller 108 of FIG. 1A may perform or cause one or more steps of method 2100 to be performed.

[0299] In block 2102, signal data indicative of interactions between the subject and acoustic waves, electromagnetic waves, and / or induced currents is obtained using one or more sensors, such as the acoustic transducer 106 and / or near-surface sensor 134 described with reference to FIG. 1A.

[0300] At block 2104, one or more metrics are calculated that quantify the peakiness or flatness of the signal data. As used herein, the terms "peakiness" and "flatness" describe the tendency of a signal to have abrupt changes. A low "peakiness" signal can be understood as a high "flatness" signal, and so these terms can be used interchangeably.

[0301] In some examples, the peakiness / flatness of the signal data may be calculated by one or more functions of spectral kurtosis, spectral skewness, spectral crest, spectral falloff, spectral slope, and spectral roll-off point. For a mathematical description of these functions, see Peeters, G., "A Large Set of Audio Features for Sound Description (Similarity and Classification) in the CUIDADO Project." Technical Report; IRCAM: Paris, France, 2004.

[0302] In other examples, other functions can be used to quantify signal peakiness, including spectral flatness and spectral entropy. For mathematical descriptions of these functions, see J. D. Johnston, "Transform coding of audio signals using perceptual noise criteria," in IEEE Journal on Selected Areas in Communications, Vol. 6, No. 2, pp. 314-323, February 1988, and H. Misra, S. Ikbal, H. Bourlard, and H. Hermansky, "Spectral entropy based feature for robust ASR," 2004 IEEE International Conference on Acoustics, Speech, and Signal Processing, Montreal, QC, Canada, 2004, pp. 1-193.

[0303] In other examples, signal peakiness may be quantified by calculating the number of extrema in the signal data, the amplitude of extrema in the signal data, the width of extrema in the signal data, and the prominence of extrema in the signal data.

[0304] At block 2106, one or more intensity values ​​indicative of the MTR present in the subject are generated based on a metric representative of the peakiness of the signal.

[0305] In another example, the intensity values ​​are generated from a distribution of results obtained by calculating the spectral entropy within a window of signal data at multiple locations on the subject. The spectral entropy for the waveform data is calculated as: I=H spectral =-ΣP(m)log2P(m) It can be calculated by: where I is entropy and P(m) is a measure of the power spectrum of the waveform data, obtained by a Fourier or wavelet transform. It is contemplated that P(m) may be weighted or normalized prior to calculation of entropy. Note that while spectral entropy shares a similar name with other signal entropy methods detailed herein, this approach is concerned with the regularity / peakiness of the signal data rather than finding specific patterns within the signal data.

[0306] In some approaches, the calculation of I is repeated at multiple locations throughout the subject to generate a spatial intensity map representative of one or more MTR characteristics within a region of the subject, or summarized into one or more response scores representative of the MTR state of the subject. An MTR score indicative of the MTR present in the subject 102 can then be calculated for the subject 102 based on the response scores and / or the spatial intensity maps. Spatial Segmentation

[0307] 22 illustrates a method 2200 for quantifying feature size, shape, and texture of an intensity map by segmenting and measuring distinct regions. Advantageously, the spatial segmentation methods described herein quantify spatial features of the object 102 (FIG. 1A) and allow for improved discrimination between objects that may have similar intensity maps. Also advantageously, the spatial segmentation methods allow for fusion of intensity maps from multiple analysis methods or sensor configurations. Method 2200 is one exemplary approach for quantifying the size, shape, or texture of an intensity map generated using one or more of the approaches described herein.

[0308] In block 2202, the inspection system 100 (FIG. 1A) receives as input one or more intensity maps associated with the object under test 102, where the intensity maps represent one or more MTR characteristics. In some embodiments, the intensity maps may be generated using one of the analysis methods described herein. In other embodiments, the intensity maps may include one or more of the following: signal magnitude, signal time of flight, signal attenuation, signal phase, signal impedance, signal voltage, and signal frequency.

[0309] At block 2204, one or more intensity maps are segmented to generate segmented intensity maps. That is, the intensity maps are used to segment the subject 102 into separate regions. In some embodiments, the intensity maps may be segmented by application of one or more of an Otsu threshold, a Voronoi-Otsu threshold, a Gaussian mixture model, a Chan-Vese model, a watershed, or a marching squares algorithm. The segmented intensity maps may optionally be filtered to remove outlying features not characteristic of MTR. Signal data is associated with the segmented regions to generate segmented data.

[0310] In block 2206, one or more spatial statistics are calculated for features in the segmented intensity map, in some embodiments, for each feature. The spatial statistics may quantify the shape, size, or texture associated with the intensity map. The spatial statistics may include at least one of size, shape, aspect ratio, area, perimeter, volume, homogeneity, intensity, surface area, or angle. In some embodiments, signal energy characteristics are aggregated across groups of distinct regions, and one or more of skewness, roll-off, entropy, kurtosis, or prominence of the aggregated signal energy characteristics are calculated.

[0311] As used herein, feature homogeneity is defined as a measure of the distribution of intensity values ​​within a feature. A highly homogeneous feature has little variation in values ​​within the feature, while a less homogeneous segmented feature may have gradients or other small variations in its internal intensity.

[0312] In some embodiments, coherent features of the signal data are calculated and statistics for the coherent features are determined.

[0313] At block 2208, the subject is classified based on spatial statistics of the segmented intensity map features. In some embodiments, the distribution of spatial statistics that best represents one or more MTR characteristics within the subject is used to classify the subject 102. In some embodiments, one or more response scores indicative of the MTR within the subject are generated based on the statistics of the coherent features.

[0314] In one example, the intensity map corresponds to a shift in the peak frequency of the bulk acoustic wave. The area and aspect ratio of each feature is calculated and compared to a reference distribution, and if the area and aspect ratio distribution is within an acceptable range, the object 102 is graded as acceptable.

[0315] In some embodiments, the subject's MTR score is determined based at least in part on spatial statistics that quantify the shape, size, and / or texture of the intensity map. Co-occurrence matrix

[0316] FIG. 23 illustrates a method for quantifying feature size, shape, and texture of one or more intensity maps into one or more response scores using a co-occurrence matrix method. Advantageously, the co-occurrence matrix method enables fusion of intensity maps from multiple analysis methods or sensor configurations. Also advantageously, the co-occurrence matrix method incorporates an offset vector to allow for sensitivity to the spatial size and spatial orientation of MTR features within the subject. It is contemplated that the offset vector may be adjusted by the inspection system 100 (FIG. 1A) based on the manufacturing process in question. Method 2300 is one exemplary approach for quantifying the size, shape, or texture of intensity maps generated using one or more of the approaches described herein.

[0317] At block 2302, the inspection system 100 receives as input one or more intensity maps representing one or more MTR characteristics to be generated. In some embodiments, the intensity maps may be generated using one of the analysis methods described herein. In other embodiments, the intensity maps may include one or more of the following: signal magnitude, signal time of flight, signal attenuation, signal phase, signal impedance, signal voltage, and signal frequency. In some embodiments, the intensity maps are generated based on one or more parameters associated with the eddy current inspection data.

[0318] In an example where the first intensity map and the second intensity map have different coordinate systems, an affine transformation may be used to bring both intensity maps into a common coordinate system.

[0319] In block 2304, the intensity map is digitized into a number of bins in the range [0,...i,...N]. In some examples, the bins may be linearly distributed within the range of intensity values, while in other examples, the bins may be non-linearly distributed, for example, on a logarithmic scale. Advantageously, adjusting the digitization bins allows additional control over which features of the intensity map are classified as equivalent.

[0320] It will be appreciated that in embodiments where multiple intensity maps are used, the i-th bin of each map may represent a different scale or value.

[0321] In block 2306, one or more vector offsets r are selected. The vector offsets may be in distance, angle, time, or any other dimension of the intensity map. In some embodiments, the vector offsets r may correspond to intensity maps of the same subject coordinates, but measure separate metrics of MTR intensity.

[0322] In one such example, the first intensity map can be the time of flight of the acoustic surface wave and the second intensity map can be the complex impedance of the eddy current coil. In another such example, the first intensity map can be the phase response of the acoustic surface wave and the second intensity map can be the attenuation of the longitudinal wave.

[0323] It will be appreciated that the mathematical formulation of the offset vectors and co-occurrence matrices may be adjusted to match the dimensionality of a selected intensity map in embodiments where the signal data and intensity map may have dimensions not otherwise described herein.

[0324] At block 2308, the inspection system 100 calculates the probability p that a point P1 having an intensity in bin i has a neighboring point P2=P1+r having an intensity corresponding to bin j. ijThis calculation is repeated for each combination of bins and stored in a matrix G. If a single two-dimensional intensity map is used, this approach corresponds mathematically to a gray level co-occurrence matrix.

[0325] This process in block 2308 is optionally repeated for a set of multiple offset vectors r. Advantageously, using a set of offset vectors may enable the inspection system 100 to determine an offset vector that best represents one or more MTR characteristics of the subject.

[0326] In block 2310, the inspection system 100:

[0327]

number

[0328] and

[0329]

number

[0330] Calculate the matrix G, which represents the joint probability of observing an event due to a reference region (e.g., a pixel) with one particular value A i and the offset region has a second specific value B j This process is repeated for all bins of data, resulting in a matrix representing the joint probability distribution of values ​​at P0 and P1.

[0331]

number

[0332] Create a.

[0333] In block 2312, a weighting function w(ij) is optionally applied to the matrix, such as a co-occurrence matrix. Applying a weighting function may allow the inspection system 100 to emphasize particular combinations of intensity bins, allowing additional discrimination regarding particular MTR characteristics.

[0334] The matrix generated via blocks 2302-2312 serves as a quantification of the texture of the intensity map. One or more response scores for the subject may then be generated based on this quantification of the intensity map.

[0335] At block 2314, the values ​​in the matrix are aggregated into one or more response scores. A weighting function may be applied to the matrix to aggregate the values ​​of the matrix into one or more response scores. For example, the values ​​in G are aggregated into one or more response scores R that indicate one or more MTR characteristics across the region of the subject. In one example, the response scores are calculated using optional weighting and scaling functions according to the formula R= h(ΣΣ(w(p ij )), p ij ∈G The data is compiled according to the following:

[0336] In one example, the weighting function w(ij) is w(i,j)=p ij *log(p ij ), which is similar to the Shannon entropy.

[0337] In another example, the weight function w(ij) is w(i,j)=p ij *|ij| and the scaling function is

[0338]

number

[0339] is.

[0340] In some embodiments, the calculation of R may be repeated for multiple values ​​of the offset vector, and the change in R may be calculated for multiple positions and feature orientations. The change in R may be used to calculate a characteristic length, characteristic orientation, characteristic size, characteristic shape, or characteristic texture of the intensity map.

[0341] In some embodiments, multiple matrices are used to determine or represent the likelihood that two energy signals separated by a particular offset distance have the same characteristics. For example, multiple matrices can be calculated or generated from multiple intensity maps or different locations.

[0342] The testing system 100 may then treat the subject according to the value of the one or more response scores. An MTR score, indicative of the MTR present in the subject 102, may then be calculated for the subject 102 based on the response scores.

[0343] Method 2300 illustrates one approach for characterizing the size, shape, and / or texture of the intensity map described herein. In some approaches, the size, shape, and / or texture of the intensity map may be quantified by selecting two or more characteristics of the signal data used to generate the intensity map. Each of the two or more characteristics is then divided into ranges of levels. A matrix containing the joint probability of observing each combination of the ranges of levels is then calculated. Weights and adjustment functions are then applied to the matrix, and one or more MTR scores may be generated for the subject based on this matrix. Intensity map aggregation

[0344] FIG. 24 illustrates an additional method 2400 for aggregating intensity maps to generate one or more response scores indicative of one or more MTR characteristics across regions of the intensity map.

[0345] In block 2402, the inspection system 100 (FIG. 1A) receives one or more intensity maps associated with the specimen 102 (FIG. 1A). In some embodiments, the intensity maps may be obtained from any combination of the sensor configurations and analysis methods herein. In other embodiments, the intensity maps may be obtained from any combination of the sensor configurations described herein and any of the analysis methods detailed in U.S. Non-Provisional Application No. 18 / 678,918, the entirety of which is incorporated herein by reference.

[0346] In block 2404, the inspection system 100 selects one or more regions in the intensity map. In some embodiments, the regions may overlap, such as filters. In other embodiments, the regions may not overlap.

[0347] In block 2406, the inspection system 100 aggregates the intensity map within each region by applying a function to the intensity data corresponding to the selected region. One or more response scores may be generated based on this aggregation. In some embodiments, the aggregation function receives an input array of data and returns one or more scalar response scores. In other embodiments, the aggregation may be applied via a window across the intensity map, which may advantageously generate a gradient representing the spatial distribution of one or more MTR characteristics across the subject.

[0348] In some embodiments, this aggregation may be calculated from one or more of the following basic statistical measures: mean, median, skewness, kurtosis, moments, variance.

[0349] In other embodiments, this aggregation may be calculated from a measure of statistical entropy, such as Shannon entropy, Renyi entropy, Tsallis entropy, etc. Renyi entropy is calculated using the formula

[0350]

number

[0351] can be defined via

[0352] where R is the response score, α is the variable coefficient, and p i is the probability that the intensity map has a value in the ith bin. In the limit α → 1, this is identical to the Shannon entropy.

[0353] In one example, the intensity map corresponds to the time of flight of a surface wave inspection: the object under test is divided into overlapping blocks and the variance of the time of flight is calculated for each block.

[0354] In another example, the intensity map may be the phase response of an eddy current inspection. The system is divided into non-overlapping blocks, and the Renyi entropy is calculated for each block. Advantageously, the parameter α may be optimized to provide improved correlation with the MTR characteristics.

[0355] In block 2408, the testing system 100 treats the subject 102 according to the value of the one or more response scores. Beam Steering

[0356] 25 illustrates a method 2500 for identifying one or more MTR characteristics by inspecting an object under test at multiple angles using one or more phased array transducers. Advantageously, because MTR characteristics can be sensitive to the angle at which the sound beam arrives, comparing a range of inspection angles can enable inspection system 100 (FIG. 1A) to make additional identifications.

[0357] In block 2502, the inspection system 100 controls one or more phased array transducers to generate acoustic waves at multiple inspection angles. In some embodiments, the angle of incidence can be within a range of ±5° with respect to the surface of the object being inspected. In other embodiments, the angle can be in a range of about ±10° to about ±25°.

[0358] In some embodiments, the system may control the angle of refraction within the object instead of the angle of incidence at the surface of the object.

[0359] In some embodiments, the inspection system may control a phased array transducer to focus an ultrasound beam into a subsurface region and focus at a selected depth and angle, hi other embodiments, the inspection system may control a transducer to generate a plane wave within the object and propagate at a selected angle.

[0360] In one example, a plane wave is used and the refraction test angle varies between approximately -3° and 3° within the subject.

[0361] In block 2504, the inspection system 100 records signal data for each inspection angle as each acoustic wave interacts with the object under test.

[0362] In block 2506, the inspection system 100 calculates one or more intensity maps and / or response scores for each inspection angle using one or more of the analysis methods described herein.

[0363] In block 2508, the inspection system 100 calculates the change in the intensity map and / or the response score as a function of the inspection angle. An MTR score indicative of the MTR present in the object 102 may then be calculated for the object 102 based on the response score and / or the spatial intensity map. Fusion of data from multiple sensors, modalities, or domains

[0364] 26A, 26B, and 26C illustrate methods of testing a subject using at least two combinations of one or more test sensors and one or more regions of the subject to ascertain details about one or more MTR characteristics. Advantageously, the combination of sensors and / or regions may enable the test system 100 to provide a procedure that is more representative of the MTR characteristics within the subject than a test utilizing only one test sensor, analytical method, or test region.

[0365] FIG. 26A illustrates a method 2600 for combining near-surface inspection data with bulk inspection data.

[0366] In block 2601, one or more surfaces of a specimen are inspected using a near-surface inspection device (e.g., surface wave UT (SWUT) or eddy current inspection (ECT)) and a bulk inspection device (e.g., volume UT (VUT)). The near-surface inspection device may be based on Rayleigh surface waves generated by a transducer or transducer array (e.g., surface wave ultrasound (SWUT), scanning acoustic microscope (SAM)). The excitation for the Rayleigh surface waves may be in the range of 1 MHz to 100 MHz using pulses, chirps, tone bursts, or coded excitation. In some embodiments, the near-surface inspection device may be based on acoustic waves generated by laser pulses (e.g., laser ultrasound or spatially resolved acoustic spectroscopy). In some embodiments, the near-surface inspection device may be based on bulk longitudinal or shear acoustic waves focused near the specimen surface. The excitation for the bulk acoustic waves may be in the range of 1 MHz to 100 MHz using pulses, chirps, tone bursts, or coded excitation. In some embodiments, the near-surface inspection device may be based on induced electric fields (e.g., eddy current inspection) or polarized waves (e.g., polarized light microscopy). The excitation for the induced current may be in the range of 100 KHz to 50 MHz using a single frequency, multiple frequencies, or a frequency sweep, such as a pulse, chirp, tone burst, or coded excitation.

[0367] In block 2602, the bulk inspection method may inspect one or more surfaces of a specimen utilizing a bulk inspection device. The bulk inspection device may include any ultrasonic sensor configuration, including the transducer and sensor arrangements described herein and illustrated in embodiments of FIGS. 2-7B.

[0368] In some embodiments, the bulk inspection device may utilize an ultrasonic sensor configuration as shown in FIG.

[0369] At block 2604, one or more scores are determined and assigned based on a combination of bulk and near-surface measurements. The near-surface measurements may be calculated using one or more of the following methods: entropy methods as previously described herein, joint probability or gray-level co-occurrence methods as previously described herein, comparisons between time-frequency spectra of near-surface data across multiple locations or multiple samples, and / or statistics of near-surface data such as amplitude, impedance, time-of-flight, etc.

[0370] In some examples, the entropy method may include the entropy by analysis method described in U.S. Non-provisional Application No. 18 / 678,918, which is incorporated herein by reference in its entirety. A variety of entropy methods may be used.

[0371] For example, one or more scores

[0372]

number

[0373] is the expression

[0374]

number

[0375] It may be calculated via Shannon e(H) via where A is the area of ​​the material sample or subarea of ​​the material sample over which data is collected, and p i is the probability that a point in region A has a value equal to the i-th level in Y'. One or more scores are calculated from the value of H for each region in an acoustic scan performed by the inspection device as described herein.

[0376] Another example of an entropy measure is one or more scores

[0377]

number

[0378] is the expression

[0379]

number

[0380] can be calculated by the Renyi entropy via

[0381] where α is an arbitrary constant between 0 and ∞. For α=0, α=1, and α→∞, H(A,α) can be evaluated by taking the limit. For α=1, the limit converges to the same metric as the Shannon entropy (H).

[0382] In some examples, the joint probability or gray level co-occurrence method may include the joint probability or gray level co-occurrence method as described in U.S. Non-provisional Application No. 18 / 678,918, the entirety of which is incorporated herein by reference.

[0383] For example, a gray-level co-occurrence matrix (G) is calculated from the standardized surface acoustic wave data (Y'). The gray-level co-occurrence matrix (G) is a measure of the intensity relationship between point P1 = (x, y, t) and its neighbors P2 = (x + Δx, y + Δy, t), where t is one or more times at which the data is sampled. It is contemplated that this approach may be applied to raw surface wave acoustic data, standardized acoustic wave data, or other transformations of near-surface measured acoustic wave data.

[0384] In one example, the gray level co-occurrence matrix (G) is a square matrix whose values ​​p i,j is an expression

[0385]

number

[0386] contains the joint probability that P1 has intensity i and P2 has intensity j, as defined by where N is a constant that normalizes the probability, A is a subset that includes some or all of the data in Y', x and y are the coordinates of a point in region A, Δx and Δy are specified distance offsets, and i and j are the intensity levels of the data in Y'.

[0387] One or more scores

[0388]

number

[0389] This second estimator of is given by the formula

[0390]

number

[0391] The joint probability p i,j may be derived from where W(i,j) is a weighting function based on i and j, an example is W(i,j)=|ij|, h is an adjustment function, an example is

[0392]

number

[0393] In another example, the weighting function is

[0394]

number

[0395] and the adjustment function is h(x)=x. In another example, the weighting function is W(i,j)=1 and h(x)=h(ΣΣx)=ΣΣx 2 Numerous other weighting and adjustment functions may also be used depending on the characteristics of the material sample and the inspection system.

[0396] In block 2606, the bulk measurements (e.g., volumetric UT (VUT)) are compared to near-surface measurements (e.g., surface wave UT (SWUT) or eddy current testing (ECT)) of one or more surfaces of the subject. In some embodiments, this may include comparing spatial profiles from one or more surfaces of the subject.

[0397] In block 2608, the compared bulk and near-surface measurements are analyzed and interpreted to determine a volumetric measurement. In some embodiments, the volumetric measurement may include a linear profile (e.g., of the MTR region) as shown in FIG. 26B or a surface profile (e.g., of the MTR region) as shown in FIG. 26C.

[0398] Advantageously, surface-based methods are more sensitive to changes in microstructure and may therefore provide a calibration for the interpretation of bulk measurement methods. Using macroslices to calibrate volume measurements

[0399] 27A-27F illustrate a method 2700 in which the analysis of an inspection of an in-production specimen is calibrated by inspection of an adjacent test coupon.

[0400] In block 2702, one or more witness macroslices are obtained from material similar to the specimen 102 during manufacturing. The macroslices may be from material directly adjacent to the specimen, near the specimen, or from another location in the same heat or forge. In one example, a titanium billet may be forged to a final diameter, and multiple cross sections may be cut from the billet. One side of each cross section may have a macroslice removed for analysis. FIG. 27B shows a macroslice 2720 of the specimen 2719.

[0401] In another example, the specimen may be die forged from a nickel billet to form the forged shape. During the forging process, a coupon may be punched from the center of the part and saved for use as a witness coupon.

[0402] In block 2704, the macro-slice 2720 may optionally be cut to expose or uncover a cross section. In some embodiments, the macro-slice 2720 may be polished. Advantageously, in other embodiments, the macro-slice 2720 may be inspected in a ground, machined, and / or peened surface state, thereby allowing for faster and / or cheaper manufacturing of the specimen.

[0403] In block 2706, the macro-slice 2720 is inspected using one or more inspection techniques: surface wave ultrasound, eddy current, longitudinal wave ultrasound, shear wave ultrasound, Lamb wave ultrasound, polarized light microscopy, or electron backscatter diffraction to obtain macro-slice signal data. Figure 27C shows the macro-slice 2720 being inspected by a sensor 2722.

[0404] In block 2708, the signal data is processed to generate macro-slice quantification signal data. The macro-slice quantification signal data may quantify at least one of a spatial intensity map, a gradient, one or more different regions within the macro-slice 2720, a distribution of spatial values, or one or more response scores. Figure 27D shows an intensity map 2724 for the macro-slice 2720.

[0405] In some embodiments, the macro-slice inspection data may be quantified using any of the signal data analysis methods described herein that are appropriate for the selected near-surface sensor.

[0406] In other embodiments, the macroslice examination data may be quantified using one or more of the methods contained in US Non-Provisional Application No. 18 / 678,918, which is incorporated herein by reference in its entirety.

[0407] At block 2710, the specimen under production is inspected using any combination of sensor arrangements and / or analytical methods described herein to obtain specimen signal data.

[0408] It is contemplated that other analytical techniques common in the industry, such as attenuation, transmission amplitude, and backscatter amplitude measurements, may also be used in conjunction with macroslice examination.

[0409] At block 2712, the object is characterized based on the macro-slice quantification signal data and the object signal data. The macro-slice quantification signal data is used by the controller 108 to guide the interpretation of one or more additional inspection methods. Advantageously, when the macro-slice 2720 is derived from a material representative of the object, the microstructure observed in the macro-slice 2720 may contain information useful for improving the accuracy and discrimination capabilities of other inspection methods.

[0410] In one example, a spatial intensity map with local phase discontinuities is generated from surface wave ultrasound testing of macro-slice 2720. The object in production is then inspected using shear wave ultrasound testing, and method 1600 of FIG. 16 is used to create an intensity map of the object. The object's intensity map is then compared to the intensity map of the macro-slice, and inspection system 100 then adjusts its interpretation of the object's intensity map based on the values ​​of the intensity map. FIG. 27E shows raw intensity map 2726 of object 2719. FIG. 27F shows intensity map 2728 of object 2719 after calibration with macro-slice 2720. As a result, inspection system 100 may treat the object based on the intensity map, which represents one or more MTR characteristics. Additional analytical methods

[0411] In some embodiments, the method of analyzing signal data includes analyzing the amplitude of one or more peaks in a waveform representing an acoustic wave, the analyzed acoustic wave being a transmitted acoustic wave propagating through a volume of a subject via transmission.

[0412] Acoustic waves are transmitted from a single-element acoustic transmitter to an acoustic receiver array along a transverse direction across a segment of the subject. In some approaches, the acoustic waves may propagate through different inspection zones 164 ( FIG. 1A ) of the subject (see also inspection zone 664 in FIG. 6 ). For example, the acoustic waves propagate across various segments along the transverse direction, e.g., at different depths. FIGS. 6A-6B illustrate how acoustic waves may propagate across different segments (e.g., strings) of the subject to inspect different inspection zones 164. In this manner, the acoustic waves inspect or scan different locations or inspection zones 164 of the subject. In some examples, the acoustic waves propagate across the string path of a billet, and the acoustic waves may propagate along various string paths. In some examples, the controller 108 ( FIG. 1A ) causes the acoustic waves to propagate from the single-element acoustic transmitter along a transverse direction across the subject. The segment (eg, chordal path) traversed by the acoustic wave can range from near the outer diameter of the object to across the centerline of the object.

[0413] In some approaches, the acoustic waves propagated across the segment of the subject include at least one of shear waves or longitudinal waves. In one example, longitudinal waves are propagated across the segment of the subject adjacent to the centerline of the subject. In another example, shear waves are propagated across the segment of the subject away from the centerline.

[0414] The peak amplitude of each acoustic waveform is determined at each location or examination zone 164 on the subject (see also examination zone 664 in FIG. 6). When the acoustic waves are ultrasound waves, the peak amplitude of an A-scan may be determined. An A-scan refers to a representation of an individual waveform taken at a single location on the subject, which may be a single pulse measurement or an average of multiple pulses. An A-scan includes a time-amplitude representation showing acoustic echo amplitude as a function of time. Controller 108 (FIG. 1A) may determine the peak amplitude of each acoustic waveform.

[0415] Based on the peak amplitudes, statistics of peak amplitudes across the subject are determined. In some approaches, statistics are calculated across different locations or examination zones 164 (see also examination zones 664 in FIG. 6 ) where an acoustic scan was performed. In one example, the peak amplitude is determined for each examination zone 164 of the subject, and then an average peak amplitude for all examination zones 164 of the subject is determined. The controller 108 may determine the statistics of peak amplitudes for the subject. The statistics of peak amplitudes may include one or more of the mean, median, median absolute deviation, interquartile range, or percentile values. It is also contemplated that a measure of statistical entropy, such as Shannon entropy or Renyi entropy, may be calculated for the distribution of peak amplitudes or peak amplitude statistics.

[0416] One or more response scores indicative of the MTR are generated for the subject based on the average peak amplitude. In some examples, the controller 108 generates a response score for the subject based on the average peak amplitude across all subjects. The arrival time of the peak amplitude is determined, and the statistics determined in block 104 are applied to that measurement, which can then be correlated with an MTR score.

[0417] In some embodiments, the method of analyzing signal data includes analyzing a signal envelope of signal data representing acoustic waves. The analyzed acoustic waves may be acoustic waves propagating through a volume of a subject via transmission.

[0418] Signal data representing acoustic waves received throughout a volume of the subject is collected. In some examples, the signal data includes a time-domain waveform. The signal data is collected by one or more transducers in a pulse-echo configuration and may represent acoustic waves scattered or reflected from the subject. In some embodiments, the acoustic waves are shear waves. The controller 108 may be configured to collect or receive the signal data.

[0419] A signal envelope is calculated for the signal data. In one example, when the waveform data is from an ultrasound examination, the signal envelope is calculated for an A-scan. In some examples, the signal envelope is calculated using at least one of a Hilbert transform, a Hilbert-Huan transform, a rolling root-mean-square filter, or fitting an interpolating spline to the signal peaks. The controller 108 may be configured to calculate the signal envelope for the waveform data.

[0420] One or more metrics indicative of the acoustic wave scattering attenuation pattern or A-scan backscatter profile are calculated based on the signal envelope. Such metrics include, for example, time vs. peak, slope, concavity, etc. In some examples, the controller 108 is configured to calculate the metrics.

[0421] In some approaches, signal statistics are determined from multiple signal envelopes. In one example, the time to peak of each individual waveform is measured with a metric assigned based on one or more statistical measures of the resulting distribution of values ​​(e.g., mean, median, interquartile range, percentiles, median absolute deviation). In another example, multiple waveform envelopes are combined into a composite waveform, and a metric is assigned based on characteristics of the composite waveform. In another example, the composite waveform is obtained via a weighted average of multiple waveform signal envelopes, which correspond to multiple locations or test zones 164 on the sample (see also test zones 664 in FIG. 6).

[0422] One or more response scores indicative of the MTR for the subject are generated based on the metrics. In some examples, the controller 108 is configured to generate the response score for the subject based on the metrics. A material with a high MTR within the subject may result in a signal envelope having a different shape than a signal envelope for a material with a low MTR because the MTR causes sound to scatter and attenuate.

[0423] FIG. 28 illustrates signal envelopes for example waveforms. The signal data shown in FIG. 28 was collected using a transducer in a shear wave pulse-echo configuration. In particular, the signal data was collected using the transducer arrangement shown and described with reference to FIG. 5 . Graph 2810 illustrates an acoustic waveform 2812 for a first MTR case, labeled MTR Case 1. Line 2814 is the signal envelope for acoustic waveform 2812 calculated using a Hilbert transform. Graph 2816 illustrates an acoustic waveform 2818 for a second MTR case, labeled MTR Case 2. Line 2820 is the signal envelope for acoustic waveform 2818 calculated using a Hilbert transform. In MTR Case 2, the subject under investigation exhibited a higher level of MTR, with a different average envelope shape.

[0424] In some embodiments, a method of analyzing signal data includes analyzing a waveform representing an acoustic wave using statistical properties or attributes of the waveform. The analyzed acoustic wave can be a scattered or reflected acoustic wave.

[0425] Signal data representing acoustic waves propagating through a volume of the subject is collected. The signal data is collected by one or more transducers in a pulse-echo configuration and may represent acoustic waves scattered or reflected from the subject. In some embodiments, the acoustic waves are shear waves. The controller 108 may be configured to collect or receive the signal data.

[0426] The signal data is then processed. In some embodiments, the signal data is processed using a Hilbert transform. It is also contemplated that in other embodiments, raw, unprocessed signal data may also be used. In some examples, the controller 108 may be configured to process the signal data.

[0427] One or more metrics indicative of the acoustic wave scattering attenuation pattern are calculated. Metrics indicative of the acoustic wave scattering attenuation pattern may include statistics such as root-mean-square amplitude, number of peaks, peak amplitude distribution, statistics of peaks with a particular prominence, height, and / or width, etc. Note that peak amplitude is different from peak amplitude distribution. Peak amplitude generally refers to the largest observed amplitude, as well as the sorted amplitudes of the second highest peak, the third highest peak, the Nth highest peak, etc. However, peak distribution may refer to filtering of peaks based on prominence, height, and / or width, or the use of signal envelopes. In some approaches, signal data can be used to calculate a measure of information entropy. In one example, signal entropy is calculated from time-series signal data using methods such as approximate entropy, variance entropy, fluctuation variance entropy, spectral entropy, fuzzy entropy, gradient entropy, and phase entropy. In another example, information entropy is calculated from distribution values ​​of waveform or peak data, such as C-scan images, using methods such as Shannon entropy or Renyi entropy. It is contemplated that any metric of signal or information entropy can be used to generate a response score.

[0428] The data is used to generate one or more response scores indicative of the MTR within the subject. For example, a material with a higher MTR may exhibit a different pattern of peaks as the acoustic waves scatter and attenuate. In some examples, the controller 108 is configured to generate one or more response scores indicative of the MTR.

[0429] In some embodiments, the method of analyzing signal data uses wavelet analysis to analyze waveforms representing acoustic waves. The analyzed acoustic waves can be scattered or reflected acoustic waves.

[0430] Signal data representing acoustic waves propagating through a volume of the subject is collected. The signal data is collected by one or more transducers in a pulse-echo configuration and may represent acoustic waves scattered or reflected from the subject. The acoustic waves may be longitudinal, shear, or mixed-mode waves. The controller 108 may be configured to collect or receive the signal data. In some embodiments, when the acoustic waves are ultrasound waves, the signal data is in the form of an A-scan.

[0431] The acoustic signal data is modified by convolution or deconvolution with a signal wavelet. In one example, the wavelet is a Gaussian pulse corresponding to a characteristic of a transducer pulse, e.g., a center frequency of the transducer. In another example, the wavelet is characteristic of a feature of the object, e.g., a wavelet corresponding to an acoustic wave scattered from a particle feature. In some examples, the controller 108 (FIG. 1A) is configured to convolve or deconvolve the wavelet.

[0432] The convolved or deconvolved signal data is analyzed. In some examples, the analysis is based on characteristics such as amplitude, statistical distribution of signal peaks, etc. In other examples, the analysis is based on frequency domain characteristics of the convolved or deconvolved signal data. In some examples, the controller 108 (FIG. 1A) is configured to analyze the signal data and compare the deconvolved wavelets.

[0433] In another example, spectrograms are generated via wavelet analysis. The time-frequency distribution of the signal data may indicate one or more MTR characteristics within the subject. The spectrograms may be compared for internal variation by comparing multiple spectrograms from multiple locations within the subject. The spectrograms may also be compared with different subjects or calibration standards to assess MTR characteristics.

[0434] One or more response scores indicative of the MTR are generated for the subject based on the wavelet characteristics. In some examples, the controller 108 is configured to generate the response scores for the subject.

[0435] In some embodiments, a method for analyzing signal data involves analyzing a signal response from a waveform representing an interaction with an internal or back-wall surface within a subject. This method may also be applied to other metrics representing attenuation within a sample, such as a signal propagating through the subject, or a signal propagating and then reflected back to a sensor. A signal may propagate through the subject to a back-wall reflector and then be reflected back to the reflector. This method may analyze the ratio of the back-wall amplitude to the root mean square (RMS) of scattering as determined from a waveform representing the acoustic wave. The RMS of scattering is the RMS of the amplitude of the scattered acoustic wave (e.g., the scattered signal). The acoustic wave analyzed via this method may be a scattered or reflected acoustic wave. The RMS is one measure of the amplitude of the scattered acoustic wave. The maximum amplitude and the 75th percentile amplitude are also measures of the scattered acoustic wave. It is contemplated that the RMS may be applied whenever amplitude data, either raw or signal envelope, is used.

[0436] Signal data representing acoustic waves propagating through a volume of the subject is collected. The signal data is collected by one or more transducers in a pulse-echo configuration and may represent acoustic waves scattered or reflected from the subject. In some approaches, the acoustic waves are longitudinal waves and / or shear waves. A controller 108 (FIG. 1A) may be configured to collect or receive the signal data.

[0437] An attenuation metric based on statistics of the back-wall reflected signals from the longitudinal waves is calculated. Examples of attenuation metrics include mean amplitude, median amplitude, etc. In some examples, a ratio of attenuation to scattering is also calculated, where scattering is obtained from scattering of the longitudinal waves or shear waves. In some examples, the controller 108 (FIG. 1A) is configured to estimate the attenuation.

[0438] One or more response scores indicative of the MTR are generated for the subject based on the attenuation metrics and / or the attenuation / scatter ratio. In some examples, the controller 108 (FIG. 1A) is configured to generate a response score for the subject.

[0439] In some embodiments, the method of analyzing the signal data involves calculating a Fourier transform or short-time Fourier transform of a section of the signal data representing an acoustic wave. The acoustic wave analyzed via this method may be a scattered or reflected acoustic wave.

[0440] Signal data representing acoustic waves propagating through a volume of the subject is collected. The signal data may be collected by one or more transducers in a pulse-echo or pitch-catch configuration and represent acoustic waves scattered or reflected from the subject. In some approaches, the acoustic waves are at least one of longitudinal waves, shear waves, or mixed-mode waves. The controller 108 may be configured to collect or receive the signal data.

[0441] At least one of a fast Fourier transform (FFT) or a short-time Fourier transform is calculated for a section of the signal data. The section of the signal data for which the FFT or short-time Fourier transform is calculated may include a section of backscatter and / or backwall echoes. The controller 108 may be configured to calculate the fast Fourier transform or short-time Fourier transform of the signal data.

[0442] Characteristics of the FFT result are analyzed as the spatial domain and / or time window of the FFT are varied. The analyzed characteristics of the FFT result may include at least one of the peak frequency, the shape of the FFT, the power spectrum of the FFT (e.g., within a specified frequency range), the variance of the FFT result, the spectral density, the cross-spectral density, the coherence, and the entropy of the spectral density. The controller 108 may be configured to analyze the characteristics of the FFT result as the spatial domain and / or the time window of the FFT are varied.

[0443] One or more response scores indicative of the MTR are generated for the subject based on characteristics of the FFT results. In some examples, the controller 108 is configured to generate the response scores for the subject.

[0444] In some embodiments, a method for analyzing signal data involves determining or calculating a measure of information entropy based on signal data representing acoustic waves. The entropy measure may be calculated directly from the signal data or from one or more characteristics of the signal data. The signal data analyzed via this method may represent scattered or reflected acoustic waves, induced currents, or polarized waves.

[0445] Waveform data representing acoustic waves propagating through a volume of the subject is collected. The waveform data is collected by one or more transducers in a pulse-echo configuration and may represent acoustic waves scattered or reflected from the subject. In some approaches, the acoustic waves are at least one of longitudinal waves, shear waves, or mixed-mode waves. The controller 108 (FIG. 1A) may be configured to collect or receive the waveform data.

[0446] The information entropy measure is determined or calculated based on at least one of: A) time-domain signal data, B) frequency-domain signal data, C) the probability of observing a region with certain characteristics (e.g., aspect ratio, maximum number of voxels) when the region is reconstructed from waveform data, and / or D) statistics of the reconstructed shape region and / or volume of the waveform data and / or waveform. Exemplary measures of information entropy include variance entropy, approximation entropy, fluctuation variance entropy, spectral entropy, sample entropy, Shannon entropy, Tsallis entropy, Renyi entropy, gradient entropy, phase entropy, and fuzzy entropy. The information entropy measure may also be calculated using multiscale entropy methods that compare entropy at multiple time, frequency, or location scales. It is contemplated that any metric that measures the entropy of a two-dimensional data set, a high-dimensional data set, a distribution, a time-series waveform, or a frequency-domain waveform may be used. In some examples, the controller 108 is configured to determine or calculate a measure of information entropy.

[0447] In one example, a response score is generated from a distribution of results obtained by calculating the spectral entropy within a window of waveform data at multiple locations on the subject. The spectral entropy for waveform data is calculated using the formula S=-ΣP(m)log2P(m) It can be calculated by:

[0448] where S is entropy and P(m) is a measure of the power spectrum of the waveform data, obtained by a Fourier or wavelet transform. It is contemplated that P(m) may be weighted or normalized before calculating the entropy.

[0449] In another example, gradient entropy may be calculated for each waveform. The waveform data may optionally be normalized, filtered, or centered.

[0450] In another example, fuzzy entropy may be calculated for each waveform. The waveform data may optionally be normalized, filtered, or centered.

[0451] In another example, the determination of entropy may be based on multi-scale entropy.

[0452] In another example, a window may be moved sequentially over the waveform data and local features extracted from the amplitude of the time to peak within the window. In this example, a response score is calculated using the Renyi entropy of the distribution of properties of these local features, such as size, shape, and / or elongation.

[0453] The entropy metric is given by the formula S=-log(B(X)) can be calculated from where B is a basis function based on a summation over a property of the data (eg, rate, time, frequency, or some other property) or the ratio of the summation between one window of data and another window of data.

[0454] One or more response scores indicative of the MTR are generated for the subject based on the measure of information entropy. In some examples, the controller 108 is configured to generate the response scores for the subject.

[0455] In some embodiments, a method for analyzing sensor data involves selecting at least two regions of data, including a reference region and an offset region, selecting data characteristics, calculating a joint probability distribution from the selected data characteristics and the at least two regions of data, and calculating a composite score. The acoustic waves analyzed via this method may be scattered or reflected acoustic waves. This method may also be applied to near-surface inspection data. The at least two regions of data may include individual points, areas, volumes, or windows of time-domain data.

[0456] Sensor data is collected. The sensor data represents at least one of A) acoustic waves propagating through a volume of the subject, B) acoustic surface waves, C) induced currents on the surface of the subject, or D) polarization of light reflected from the surface of the subject. In some examples, the sensor data is collected by one or more transducers in a pulse-echo configuration and may represent acoustic waves scattered or reflected from the subject. In some approaches, the acoustic waves are at least one of longitudinal waves, shear waves, or mixed-mode waves. In some examples, the sensor data is collected by one or more near-surface sensors and may represent acoustic waves scattered or reflected from a near-surface volume of the subject. In some examples, the sensor data is collected by one or more near-surface sensors and may represent eddy current impedance on the subject. In some examples, the controller 108 is configured to collect or receive waveform data.

[0457] Sensor data in a reference region P0 and one or more offset regions (P1...P n) is determined. The offset region is offset by some difference in space, time, or frequency. The reference region and offset region can be individual points, regions, volumes, time windows, or combinations thereof. In one example, the reference region is a pixel on a two-dimensional scan image, and the offset region is an adjacent or nearly adjacent pixel (e.g., one or more pixels away from the pixel in the reference region). In some embodiments, the pixel in the offset region may share a common edge or corner with the pixel in the reference region, or the pixel in the offset region may be one or more pixels away from the pixel in the reference region. For example, the pixel in the reference region may be located at (x, y) and the pixel in the offset region may be located at (x±1, y), or the pixel in the offset region may be located at (x±2, y). In another example, the reference region is a window of time-domain data, such as an ultrasound waveform, and the offset region is a window of time-domain data that appears at some specified offset.

[0458] At least one characteristic of the sensor data is selected and divided into a set of bins. The characteristic includes intensity of the scanned image, signal amplitude, signal phase, signal time of flight, signal peak amplitude / height, signal peak width, signal peak prominence, signal frequency distribution, calculation of an entropy measure, or basis functions suitable for use in the entropy measure. In some examples, the controller 108 is configured to select the characteristic of the sensor data and divide the characteristic of the sensor data into a set of bins.

[0459] In one example, the selected property may be gradient entropy, which may be calculated from a window of ultrasound waveform data acquired at a particular point on the sample.

[0460] In another example, the peak amplitude of an ultrasound waveform is represented as a two-dimensional image, and the selected characteristic is the amplitude at a particular pixel on the image.

[0461] In another example, the eddy current impedance data may be represented as a two-dimensional image, and the selected property may be the magnitude of the impedance at a particular pixel on the image.

[0462] In another example, the selected characteristic may be the magnitude of the impedance of the eddy current coil at one or more frequencies.

[0463] The reference position is shifted across a subregion of the dataset that includes the sensor data, and signal characteristic values ​​in the reference region and the offset region are collected and divided into bins. In some examples, the controller 108 is configured to select a characteristic of the sensor data and divide the characteristic of the sensor data into a series of bins.

[0464] The reference area (e.g., pixels) is one specific value A i and the offset region has a second specific value B j Probability p of having ij This process is repeated for all bins of data to create a matrix G that represents the joint probability distribution of values ​​at P0 and P1.

[0465] The joint probability value is p ij The score may be weighted or adjusted according to the value of , and converted into a subsequent non-destructive examination (NDE) score. G=h(ΣΣ(w(p ij )) It can be calculated by: where w(x) is the weighting function and h(x) is the scaling function.

[0466] In one example, the weighting function w(ij) is w(i,j)=p ij *log(p ij ), which is equivalent to the Shannon entropy.

[0467] In another example, the weight function w(ij) is w(i,j)=p ij*|ij| and the scaling function is

[0468]

number

[0469] is.

[0470] One or more response scores representing the MTR are calculated from the values ​​of G. In some examples, the controller 108 is configured to correlate measures of the matrix G to the MTR scores.

[0471] In some embodiments, a method of analyzing signal data involves comparing patterns between elements of waveform data from an array transducer. The acoustic waves analyzed via this method can be scattered, reflected, or propagated acoustic waves.

[0472] Waveform data representing acoustic waves propagating through a volume of the subject is collected. In one approach, the waveforms may be collected by one or more transducers in a pulse-echo configuration and represent acoustic waves scattered or reflected from the subject. In another approach, the waveform data may be collected by one or more transducers in a pitch-catch configuration and represent acoustic waves propagating through the subject. The controller 108 may be configured to collect or receive the waveform data.

[0473] Patterns in the waveform data between elements of the transducer array are compared. Examples of patterns include patterns of correlation, decay rate, etc. in the waveform data. Elements that may be compared include, for example, adjacent elements in the transducer array and / or elements separated by N other elements.

[0474] One or more response scores indicative of the MTR are generated for the subject based on the pattern of the waveform data between the elements. In some examples, the controller 108 is configured to generate the response scores.

[0475] In some embodiments, the method of analyzing waveform data involves full-matrix capture of waveform data indicative of acoustic waves. The acoustic waves analyzed via this method may be scattered, reflected, or propagated acoustic waves.

[0476] Waveform data is collected using an array transducer by separately pulsing one or more elements of the array and recording the waveform data received by each element of the array. In one approach, the waveform data may be collected by one or more transducers in a pulse-echo configuration and represent acoustic waves scattered or reflected from the subject. In another approach, the waveform data may be collected by one or more transducers in a pitch-catch configuration and represent acoustic waves propagating through the subject. In some examples, the controller 108 is configured to collect or receive the waveform data.

[0477] A path from a region of the subject to one or more elements of the array transducer is reconstructed. The controller 108 can be configured to reconstruct a path from a region of the subject to each element of the array transducer.

[0478] The waveforms may be received by one or more elements of the array. An optional weighting function may also be applied to each waveform. A composite value for each point on the subject is constructed based on the waveforms. In some examples, the controller 108 is configured to compare the waveforms received by one or more elements of the array, apply an optional weighting function to each waveform, and construct a composite value for each point on the subject based on the waveforms.

[0479] Conventional approaches to full matrix capture for defect or inclusion detection are generally based on the propagation time from every element in the array to each point. In the approach to full matrix capture described herein used for MTR detection, waveforms from particular elements in the array can be weighted to better detect MTRs.

[0480] One or more response scores indicative of the MTR are generated for the subject based on the composite value. In some examples, the controller 108 is configured to generate a response score for the subject.

[0481] In some embodiments, the method of analyzing waveform data involves shape and / or area analysis of two-dimensional or three-dimensional representations of waveform data throughout a subject. The acoustic waves analyzed via this method may be scattered or reflected acoustic waves.

[0482] In a first approach, waveform data is collected throughout the volume of the subject. The waveform data is collected at various points or locations on the subject, with sensor arrangements targeting particular examination zones 164 (see also examination zone 664 in FIG. 6 ) and / or depths of the subject at those points. The waveform data may be collected using longitudinal, shear, or mixed-mode acoustic waves. In some examples, the controller 108 is configured to collect or receive the waveform data.

[0483] Voxels in the waveform data may be selected based on one or more criteria. Voxels represent individual points in space on a three-dimensional matrix. The criteria may be based on thresholds for characteristics of the waveform data. In some examples, the criteria may be thresholds for at least one of amplitude, noise, amplitude statistics, frequency, and noise statistics. In some examples, the controller 108 is configured to select voxels based on the criteria.

[0484] Voxels with similar properties (e.g., amplitude, frequency, noise, amplitude statistics, noise statistics) are grouped together. A group of voxels with similar properties may define a characteristic of the subject. In some examples, the controller 108 is configured to group voxels with similar properties.

[0485] Statistics of the grouped voxels are calculated. The statistics of the grouped voxels may then be related or correlated with MTR characteristics. In one example, the volume of the grouped voxels is broadly related to the MTR volume within the subject. In some approaches, the statistics of the grouped voxels may be associated or correlated with the MTR score of the subject. In some examples, the controller 108 is configured to determine or calculate statistics of the grouped voxels and correlate such statistics with MTR characteristics. In some examples, the waveforms from each slice of the subject (e.g., a billet) after thresholding are stitched together to create a complete 3D visualization of MTR-related features within the subject. Such 3D visualization may be made interactive, allowing an operator to virtually "walk through" the subject and identify volumes of concern for further investigation.

[0486] In a second approach, local peaks within the waveform data are identified. In some examples, the controller 108 is configured to identify local peaks within the waveform data.

[0487] The time of flight and amplitude for the local peak is calculated or determined. The controller 108 may be configured to determine the time of flight and amplitude for the local peak.

[0488] The size, shape, and / or duration of the local peak may be determined. The size, shape, and / or duration of the local peak may be related or correlated with an MTR score for the subject. The controller 108 may be configured to determine the size, shape, and / or duration of the local peak.

[0489] MTR detection and scoring for a subject may be achieved by combining waveform data (e.g., acoustic waveform data) and data analysis from multiple examination zones 164 (see also examination zone 664 in FIG. 6 ) and / or from multiple transducer arrangement configurations described herein. Waveform data may be combined, and conclusions from one of the analysis methods described herein may be based on results from another analysis method described herein. For example, if a phase coherence analysis method results in a score of “A” at depth “B” in examination zone 164 “C” with transducer arrangement “D,” that may affect how a score of “E” at depth “F” in examination zone 164 “G” with sensor “H” is interpreted.

[0490] Waveform data acquired from one or more of the transducer arrangements and one or more of the analysis methods described herein may be used to determine a response score for the subject, which is indicative of the MTR present in the subject and may be used, for example, as a predictor of the level of MTR in the subject.

[0491] It is contemplated that the response score may be used to make a determination regarding the subject. As an example, the subject may be classified or graded based on the response score. In another example, the subject may be rejected or accepted based on the response score. The response score may be compared to a threshold score or threshold range to determine whether to accept or reject the material sample. The threshold score or threshold range may reflect a response score value that is acceptable. A response score that falls within the threshold range may reflect that the material from which the material sample was obtained is suitable for use or further processing. If the response score for the material sample is outside the threshold range, the material sample may be rejected.

[0492] The response score can be related to an MTR score or value. The MTR score can reflect ground truth MTR characteristics, for example, as determined by EBSD, SWUT, polarization, eddy current, or other methods determined to be reliable for MTR determination. The MTR score can be related to the response score and, as a result, can serve as a guide for setting thresholds for accepting or rejecting specimens based on the response score.

[0493] To establish a relationship between response scores and MTR scores, multiple subjects can be evaluated via the acoustic testing approach described herein, thereby determining one or more response scores for the subjects. The multiple subjects can then also be evaluated using a ground truth approach to determine MTR levels, such as EBSD, for the subjects. Relating acoustic scores to ground truth MTR levels can then validate the use of response scores as predictors of MTR and can also guide the selection of thresholds for classifying and / or accepting or rejecting subjects based on response scores.

[0494] In some embodiments, a method is provided for analyzing one or more sets of waveform data to make a subject classification determination.

[0495] The method involves partitioning the specimen into one or more inspection zones 164 (FIG. 1A) (see also inspection zone 664 in FIG. 6). In some examples, the inspection zones 164 are partitioned based on depth into the sample. In some examples, the inspection zones 164 are partitioned based on the geometry of the specimen surface. The inspection zones 164 can overlap or have uninspected space between adjacent zones.

[0496] Each test zone 164 (see also test zone 664 in FIG. 6) is tested, and waveform data is collected for each test zone 164. One or more of the sensor arrangements described with reference to FIGS. 2-7B may be used to test the test zones 164.

[0497] A response score is determined for each test zone 164 (see also test zone 664 in FIG. 6 ) of the subject. The determination of the response score can use data from one or more waveforms. The response score can be determined using one or more of the analysis methods shown and described with reference to FIGS. 8-19 . In some examples, the response score is a function of acoustic waveform data from one or more sensor arrangements used to test a particular test zone 164. In some examples, the response score for a zone is also a function of acoustic waveform data from other test zones 164 of the subject. It is contemplated that any acoustic waveform data from shallower zones may affect the interpretation of acoustic waveform data from testing deeper zones.

[0498] A single composite response score is determined for the subject using waveform data from one or more test zones 164 (see also test zone 664 in FIG. 6).

[0499] Referring briefly to FIG. 29, in a first example, a predictive model is generated from waveform data, and a grading decision for a subject is calculated based on a predicted response score.

[0500]

number

[0501] Equation 1 is an exemplary equation for generating a single composite response score for a subject. Equation 1 provides a predictive value for the response score:

[0502]

number

[0503] is an exemplary formula for determining

[0504]

number

[0505] is the predicted value of the MTR score,

[0506]

number

[0507] is the sensor arrangement configuration m Inspection Zone Z from n (e.g., test zone 164), and n are the coefficients fitted by regression, and h1,...,h n is a function that takes as input the response score of the Nth zone.

[0508] 29, in a second example, a set of predictive models is generated from the waveform data, each test zone 164 (FIG. 1A) (see also test zone 664 in FIG. 6) is assigned a response score, and a grading decision is made taking into account the set of response scores. Equations 2, 3, and 4 are exemplary equations for generating multiple response scores for a subject. Equations 2, 3, and 4 are exemplary equations for a set of predictive models in which each test zone 164 is assigned a response score. In this example, the response score for a first test zone (e.g., test zone 164) is

[0509]

number

[0510] is a function of acoustic waveform data from two different sensor configurations s1 and s2 for the first test zone. In this example, a response score s is assigned to a second test zone (e.g., test zone 164).

[0511]

number

[0512] Alternatively, σ is a function of waveform data from two different test zones and a single sensor configuration.

[0513] A material classification decision is made based on one or more of the determined response scores. Exemplary decisions include accepting the sample, rejecting the sample, and assigning a material grade to the sample.

[0514] In one example, four response scores are measured corresponding to four test zones 164 (FIG. 1A) (see also test zone 664 in FIG. 6). The response scores in the first and second test zones 164 are classified as acceptable, while the response scores in the third and fourth test zones 164 are classified as rejectable. In this example, the sample may be rejected entirely. Alternatively, only the third and fourth test zones 164 may be rejected, and the sample may then be processed to remove the rejected material. By facilitating the scoring of the test zones 164 of material in a specimen, it is contemplated that portions of the specimen 102 (FIG. 1A) or test zones 164 deemed unsuitable for a particular application may be discarded, while other portions of material or test zones 164 may be recovered to reduce waste. Reducing waste may also reduce production costs and material waste. It is also contemplated that tests using multiple test zones 164 have improved sensitivity and capacity.

[0515] In another approach, the MTR level may be characterized by the shape of the acoustic waveform data, which may represent one or more MTR or particle properties of the subject.

[0516] In some embodiments, a method is provided for analyzing one or more sets of acoustic waveforms to characterize MTR levels on a subject.

[0517] The scattered acoustic waveform data may be collected from longitudinal, shear, or mixed-mode waves and may be generated and received using any of the inspection methods shown and described with reference to Figures 1A-7B. In some approaches, the scattered acoustic waveform data may be generated by the inspection system 100 and inspection device 104 of Figure 1A.

[0518] The acoustic data is pre-processed through one or more filtering and / or normalization processes. Pre-processing can involve a variety of operations.

[0519] In one example, the acoustic data may be preprocessed through a Hilbert transform performed on the waveform data at each location on the subject. It is also contemplated that alternative methods, such as fitting a spline to the amplitude or root-mean-square values ​​of the waveform data, may also be used to describe signal strength. Multiple waveforms are then averaged together, and the averaged waveforms are then normalized to have values ​​between 0 and 1.

[0520] The shape of the waveform data is characterized and / or parameterized. In one example, various signal processing parameters may be used, including, but not limited to, peak height, time to peak, peak width, peak prominence, or integrated area. The signal processing parameters may further include scaling of the data to a specified range, one or more frequency-based filters, or one or more statistical-based filters.

[0521] The shape of the waveform data and signal profile characteristics are determined. The shape of the waveform data and signal profile characteristics may be quantified based on a parametric function. The parametric function is fitted to match the waveform data based on a numerical analytical method, including least squares, nonlinear least squares, and principal component regression. It is contemplated that any numerical analytical method for fitting a specified function to data may be used. The function may be fitted to any region of interest within the waveform. The characteristic shape parameters may further include at least one of the height of the signal peak, the position of the signal peak, the width of the signal peak, the prominence of the signal peak, the integrated area of ​​the signal envelope, or the integrated area of ​​the parametric function, to name a few.

[0522] The parametric function may be a polynomial, an exponential function, a Gaussian function, or a nonlinear combination thereof. It is contemplated that additional parametric functions in the form of statistical distributions, such as a Lévy distribution, a lognormal distribution, a Weibull distribution, or similar distributions, may be used. The parametric function may also be related to the probability density function (PDF) or cumulative distribution function (CDF) of the statistical distribution, or to a physics-based model. It is also contemplated that numerical models based on physical principles of acoustic interaction with particles or microtexture features may be fitted in this regard.

[0523] Referring briefly to FIG. 16C , in one example, preprocessed waveform data is fitted to a modified exponential function. Equation 1 and Equation 2 are exemplary equations that may be used to fit the waveform data. Graph 1620 shown in FIG. 16B presents a graph illustrating preprocessed signal data 1622 (e.g., preprocessed waveform data) fitted with a modified exponential equation 1624. The modified exponential equation may include Equations 1 and 2 as shown in FIG. 16C . Information regarding the MTR level in the subject may be derived from the fitted values ​​of one or more function parameters (a, b, c, d, etc.).

[0524] In another example, the data is fitted by a series of splines.

[0525] In yet another example, a linear function is fitted to a domain of waveform data.

[0526] In yet another example, waveform data is preprocessed by averaging over subregions of samples and applying a median filter to each signal. The preprocessed waveform data is then element-wise exponentiated. A spatial map of the subject can then be generated from element-wise differences of cumulative sums of the preprocessed waveform data. The spatial map can be used as a representation of one or more MTR characteristics of the subject or to generate one or more scores. The scores may be indicative of MTR regions at various locations on the subject.

[0527] In another example, the shape of the signal data (eg, waveform data) may be determined by calculating the area under the signal curve of the signal data.

[0528] In another example, the shape of signal data (eg, waveform data) may be determined by calculating the element-by-element difference of a running sum of the signal data raised to a specified exponent.

[0529] The quantified waveform shape data is compared spatially across the subject. The quantified waveform shape data is compared across multiple locations on the subject, which may represent individual locations / points, locally averaged positions, profiles of fit parameters across the subject, etc.

[0530] Subjects can be treated or graded based on the value and distribution of spatially quantified waveform shape data at a single location or across the entire subject.

[0531] In some examples, shear wave ultrasonic testing is performed on a cylindrical component, and the shape of the acoustic waveform is then characterized over the bulk of the subject according to the method shown in FIG.

[0532] A small radial cross section may be machined from the specimen or nearby material having a particular grain or texture feature of interest, and Rayleigh surface wave or eddy current testing may be performed on the cross section. Eddy current testing means that induced currents may be used to detect grain surface or subsurface features.

[0533] Near-surface inspection is performed using Rayleigh surface waves or eddy current inspection. The inspection data is quantified using entropy or joint probability methods based on the spatial distribution of impedance magnitude and phase at one or more excitation frequencies. The near-surface data is then used to generate a spatial intensity profile for the cylindrical component. The spatial intensity profile is graded by the radial attenuation of the MTR level and the integration of the beam geometry with the SWUT predicted MTR profile.

[0534] The ultrasonic wavefront pattern is compared to a predicted radial MTR profile. In some embodiments, the numerical or experimental ultrasonic waveform profile is integrated temporally and spatially over a region of the sample with respect to the radial MTR intensity.

[0535] In some instances, interpretation of the waveform shape may be based on the radial profile, in which case the bulk method is calibrated by the radial cross-section and used to treat larger volumes of similar material without requiring additional cross-sections from the subject.

[0536] In another example, frequency or time-frequency spectra can be compared at multiple locations within a subject or across subjects of different qualities, and the similarity between nearby locations can be assessed based on the similarity or dissimilarity between the frequency or time-frequency spectra.

[0537] In another example, near-surface data can be compared to bulk inspection data to calibrate and guide the interpretation of the bulk inspection data.

[0538] In some embodiments, the one or more inspection sensors are configured to acquire surface or near-surface inspection data of the specimen at one or more inspection frequencies. The one or more inspection sensors may include at least one of an eddy current sensor or a polarized light sensor.

[0539] Surface or near-surface inspection data is received from one or more inspection sensors, and the inspection data is analyzed using at least one of an information entropy determination approach, a joint probability or co-occurrence matrix, or a spatial-temporal and frequency distribution recognition and / or analysis of the surface or near-surface inspection data.

[0540] In some embodiments, the method for analyzing signal data calculates a sensor score by calculating the entropy of one or more characteristics of the signal. The entropy may be calculated for one or more positions, time windows, frequencies, embedding dimensions, and / or scales. In some embodiments, the entropy may be calculated for one or more different time series or sets of signal data. In other embodiments, cross-entropy may be calculated using two or more time series or sets of signal data. In some embodiments, the entropy function may be modified by best-fit parameters obtained through experimentation. In some embodiments, the entropy function is adaptive and may adjust the scale, coarse-graining, and embedding dimensions based on the signal data. In some embodiments, the method, or portions thereof, are implemented or performed using the inspection system 100. For example, the controller 108 may execute or cause one or more steps of the method to be executed.

[0541] As used herein, an entropy function is understood as a function or algorithm within a family of functions that describe the similarity and probability of occurrence for a set of signal data or data. In many embodiments, the entropy function takes the general form of a sum over the logarithm of the probability. When defined for time series data, many entropy functions may be functionally dependent on one or more embedding dimensions and one or more scale parameters, which provides the advantage of adjusting the scale and sensitivity of the output. However, some implementations of entropy, such as Renyi entropy described below, do not follow this formula except in certain limits.

[0542] Some functions in the entropy family are defined for any arbitrary set of data. Other functions are defined specifically for time series or frequency spectra. It is contemplated that entropy functions may be modified with weights or adjustment functions to optimize performance for particular subjects.

[0543] Exemplary functions included in the entropy family include variance entropy, approximate entropy, fluctuation variance entropy, spectral entropy, sample entropy, Shannon entropy, Tsallis entropy, Renyi entropy, gradient entropy, phase entropy, and fuzzy entropy.

[0544] It will be appreciated that in some embodiments, where the signal data is generated by electromagnetic sensors, induced current sensors, or a combination of either of the former with one or more acoustic transducers, these methods may represent extensions and / or improvements of the analysis methods outlined in U.S. Non-Provisional Application No. 18 / 678,918 for signal data collected from the sensor arrangements described herein.

[0545] The object under test 102 (FIG. 1A) is inspected by the inspection device 104 to acquire signal data indicative of the condition of the object under test 102. The collected signal data may be one or more of a time domain signal, a frequency spectrum, one or more peak amplitudes, one or more times of peak occurrence, one or more frequencies of peak occurrence, a signal phase, a signal coherence, a signal impedance, or a sensor voltage.

[0546] One or more characteristics of the signal data are selected. The selected characteristics may be selected based on the particular configuration of sensors in the inspection device 104 (FIG. 1A) and / or characteristics of the specimen 102 (FIG. 1A), such as the material, manufacturing process, and / or surface condition.

[0547] One or more entropy functions are selected based on a selection of signal characteristics. In some examples, the selected entropy function may be defined by an arbitrary distribution of values ​​and may be applied to a set of signal characteristics corresponding to one or more positions, times, or excitation frequencies. In other examples, the signal data is an amplitude time series, and the selected entropy function is defined on the time series data. In other examples, the signal data is a frequency spectrum, and the selected entropy function is defined on the frequency domain data.

[0548] In another example, the set of data may be characteristics derived from the signal data, such as the shape, size, orientation, texture, or aspect ratio of a region obtained through a shape segmentation algorithm.

[0549] In one example, the signal characteristic is the phase of an eddy current coil and the entropy function selected is

[0550]

number

[0551] The Rényi entropy can be defined by where α is a variable coefficient and p i is the probability that the impedance of the signal data falls within the i-th bin of impedance values ​​within the region and / or frequency range of the subject.

[0552] In another example, the selected signal characteristic may be the amplitude time series of Rayleigh surface waves, and the selected entropy function is fuzzy entropy. Fuzzy entropy may be understood as the metric by which two signals remain similar over m+1 points. Fuzzy entropy is expressed by the formula H Fuzzy (m, n, r, N)=ln(φ m (n, r))-ln(φ m+1 (n, r) It can be defined by: where m is the embedding dimension, n and r are coefficients, and φ m and φ m+1 is a vector distance metric for time series.

[0553] The derivation and complete equation are shown in Figure 29. It is contemplated that modifications of this equation may be used, such as adding coefficients to optimize model performance. Response scores may be generated from a comparison of fuzzy entropy at one or more locations, times, scales, or frequencies.

[0554] In another example, the selected signal characteristic may be the amplitude time series of a Rayleigh surface wave, and the selected entropy function is the gradient entropy, which may be understood as a function that describes the likelihood that the signal will increase or decrease at a variable rate. In some examples, the gradient entropy is

[0555]

number

[0556] can be calculated via

[0557] In that case,i is the probability that the gradient of the signal matches the i-th pattern in the signal data. Response scores can be generated from a comparison of gradient entropy at one or more locations, times, scales, or frequencies.

[0558] In another example, the selected signal characteristic may be the frequency spectrum of an acoustic surface wave or the frequency dependent impedance of an eddy current coil, and the selected entropy function is the spectral entropy, which is H spectral =-p m logΣ(p m ) can be calculated via where p m refers to the value of each bin in the spectral power distribution of the signal data.

[0559] In another example, a window may be moved sequentially over the signal data and local features extracted from the amplitude of the time to peak within the window. In this example, a response score is calculated using the Renyi entropy of the distribution of properties of these local features, such as size, shape, and / or elongation.

[0560] The entropy is calculated at one or more positions, excitation frequencies, and / or time windows.

[0561] One or more response scores indicative of the MTR in the subject are generated based on a measure of information entropy.

[0562] In some embodiments, a method is provided for analyzing signal data using a co-occurrence matrix by generating a joint probability matrix from at least two marginal distributions that represent characteristics of the signal data. In some embodiments, the method, or portions thereof, are implemented or performed using inspection system 100. For example, controller 108 (FIG. 1A) may perform or cause one or more steps of the method to be performed.

[0563] Using one or more sensors, such as those described with reference to FIG. 27A, signal data indicative of the interaction between the subject and acoustic waves, electromagnetic waves, and / or induced currents is obtained.

[0564] One or more characteristics of the signal data are calculated or determined at one or more of the positions, times, frequencies, or modes of excitation, resulting in at least two distributions of data. In some embodiments, two or more characteristics of the signal data are calculated or determined at one or more of the positions, times, frequencies, or modes of excitation. Examples of characteristics of the signal data that may be calculated include, but are not limited to, time series amplitude, peak amplitude, peak propagation time, impedance, phase, phase coherence, frequency spectrum, or time trace.

[0565] In some embodiments, where the signal data is generated by electromagnetic sensors, induced current sensors, or a combination of either of the former with one or more acoustic transducers, it will be understood that these co-occurrence methods may represent extensions and / or improvements to the analysis methods outlined in U.S. Non-Provisional Application No. 18 / 678,918 for signal data collected from the sensor arrangements described herein.

[0566] A joint probability distribution is created based on selected signal characteristics and / or parameter values, which may be represented as a matrix (G) (e.g., a co-occurrence matrix), and one or more response scores may be developed based on the values ​​of (G).

[0567] In one example, G is first generated by collecting data from a set of data S(r) at multiple locations.

[0568]

number

[0569] A first marginal distribution D0 is generated from the values ​​of S. Then, one or more specified vector offsets

[0570]

number

[0571] The second marginal distribution D1 is defined as

[0572]

number

[0573] The joint probability (G) is then generated from the value of S in G ij =P(S(r)=a i and S(r+Δr)=b j is generated from where a i over the binned values ​​of S(r), and b j spans the binned values ​​of S(r+Δr). In many examples, a i and b j The number of bins for each is the same, resulting in a square matrix (G).

[0574] For a single feature of signal data and a two-dimensional scan grid, (G) takes the form of a matching gray-level co-occurrence matrix.

[0575] In the more general case, a single offset vector

[0576]

number

[0577] The use of (G) is not required, and other marginal distributions used to generate (G) may be derived from multiple characteristics and sets of signal data, multiple observation times / windows, multiple frequencies, and multiple sensors, resulting in matrices of any dimensionality.

[0578] In some examples, after a matrix (G) is created, the elements G ij may be subjected to a weighting function, which provides useful control for emphasizing near-diagonal or off-diagonal elements. Two examples of such weighting functions are w(G ij ) = |i+j| and w(G ij )=i*G ij is.

[0579] Finally, one or more response scores indicative of the MTR in the subject may be generated based on the adjusted matrix G. The one or more response scores may be generated by applying an adjustment function h to the matrix (G). In general form, the response score is Response Score=h(ΣΣw(G ij )) can be generated by

[0580] Two examples of such adjustment functions are h(x)=x N and h(x) = ln(x). It is contemplated that any algebraic function is suitable as a conditioning of (G).

[0581] In one example of applying this method to signal data, the selected signal characteristic is the magnitude of the impedance of a racetrack eddy current coil, and the sensor data is collected via a raster scan. In this example, the marginal distribution is calculated as the offset vector

[0582]

number

[0583] This is generated through the use of a 0.050 inch distance offset and selected in an orientation aligned with the flow of microstructures within the specimen. Beneficially, this approach is used to measure the likelihood that regions within the specimen will return the same measurement, providing a response score that is sensitive to size and shape distribution within the specimen.

[0584] In some embodiments, a method for analyzing signal data based on similarities and / or variations in spectrograms of the signal data is provided. In some embodiments, the method, or portions thereof, are implemented or performed using inspection system 100. For example, controller 108 may perform or cause one or more steps of the method to be performed.

[0585] One or more sensors are used to obtain signal data indicative of the interaction between the subject and the acoustic waves, electromagnetic waves, and / or induced currents.

[0586] As used herein, a spectrogram is understood to be a two-dimensional data set in which one dimension is the frequency spectrum, the second axis is at least one of position and time, and the third dimension is a measure of signal intensity. A spectrogram may include multiple spatial dimensions corresponding to the object of interest.

[0587] A set of spectrograms is collected at one or more locations on the subject. In some examples, the spectrograms may be generated using a short-time Fourier transform. Various windows may be used for the short-time Fourier transform, such as a box window, a Gaussian window, or a discrete prolate spheroidal array window.

[0588] In another example, the spectrogram may be generated using a continuous wavelet transform, for which a variety of wavelets may be used, including both orthogonal and non-orthogonal wavelets, such as the Morlet wavelet, Mexican hat wavelet, generalized Morlet wavelet, etc.

[0589] In another example, a spectrogram may be generated by sweeping the excitation frequency and recording the signal amplitude at each frequency.

[0590] The spectrograms are compared to obtain information regarding the subject's MTR level.

[0591] In one example, spectrograms can be compared by the mean squared error (MSE) between two spectrograms at a variable offset in space or between multiple sensors in an array. The mean squared error can be used as a response score and also as a basic statistic for a set of MSE values. It is contemplated that other measures of image similarity can also be used.

[0592] In another example, the maxima and minima in a set of spectrograms are measured and a response score is calculated based on statistics of these maxima and minima, such as, but not limited to, the count of maxima, the size distribution of maxima, and / or the intensity of maxima.

[0593] In another example, the ridge of the spectrogram is used to generate a response score.

[0594] One or more response scores indicative of the MTR are generated based on the spectrogram.

[0595] In some embodiments, a method for analyzing signal data is based on identifying key components of the signal. In some embodiments, the method, or portions thereof, are implemented or performed using inspection system 100 (FIG. 1A). For example, controller 108 may perform or cause one or more steps of the method to be performed.

[0596] One or more sensors are used to obtain signal data indicative of interactions between the subject and acoustic waves, electromagnetic waves, and / or induced currents. The signal data is collected at multiple locations on the subject. The signal data may also be collected at multiple frequencies or with multiple modalities.

[0597] The signal data is processed into more relevant (e.g., influential) and less relevant components based on principal component analysis and / or empirical / variational mode decomposition. In some embodiments, the signal data is processed into components with one or more levels of relevance, e.g., a first level indicating a more relevant component and a second level indicating a less relevant component. Relevance refers to the degree to which a component correlates with or influences MTR levels in a subject. The level of relevance can be used to determine the components most relevant to detecting MTR.

[0598] One or more response scores for the subject are generated based on the values ​​of the more relevant (e.g., most relevant) components of the signal data. For example, the response score for the subject may be based on the correlation between the most relevant components and the MTR level of the subject.

[0599] In one example, principal component analysis can be used to identify orthogonal components within a set of signal data, remove components that are not correlated with MTR levels, and generate a response score from the components that best correlate with MTR levels in a subject.

[0600] In another example, the signal data is decomposed into subcomponent vibrational modes using at least one of empirical mode decomposition, variational mode decomposition, or multiresolution mode decomposition, also known as the Hilbert-Huang transform. One or more response scores may be calculated from these vibrational modes using methods such as peak amplitude, peak transit time, etc. Additionally, any of the time series analysis methods described herein may be applied to each of the submodes obtained through EMD to generate a response score.

[0601] Further aspects of the present disclosure are provided by the subject matter of the following sections.

[0602] 1. An inspection system for inspecting an object, the inspection system comprising: a transmit sensor array of one or more active elements configured to introduce inspection energy into the object; and a controller in operative communication with the transmit sensor array, the controller configured to: cause the inspection energy to transmit from the transmit sensor array through an incident surface of the object and through a volume of the object; receive signal data indicative of signal energy received from the object at two or more locations of the object; calculate a Fourier transform on the signal data to separate frequency characteristics from phase characteristics of the signal data; calculate a shift in frequency distribution of the frequency characteristics between the signal energy at the two or more locations of the object; generate an intensity map or response score based on the shift in frequency distribution; and determine at least one microtextured region (MTR) score indicative of MTRs present within a subsurface volume of the object based at least in part on the shift in frequency distribution of the signal energy.

[0603] 1. An inspection system for inspecting an object, the inspection system comprising: a transmit sensor array of one or more active elements configured to introduce inspection energy into the object; and a controller in operative communication with the transmit sensor array, the controller configured to cause the inspection energy to be transmitted from the transmit sensor array through an incident surface of the object and through a volume of the object, receive signal data indicative of signal energy received from the object, calculate a distribution of values ​​for a given characteristic of the signal data, calculate a quantified description of the distribution of values ​​for the given characteristic of the signal energy, and determine at least one microtextured region (MTR) score indicative of MTRs present within the subsurface volume of the object based at least in part on the quantified description of the distribution of values ​​for the given characteristic of the signal data.

[0604] 1. An inspection system for inspecting an object, the inspection system comprising: a transmit sensor array of one or more active elements configured to introduce inspection energy into the object; and a controller in operative communication with the transmit sensor array, the controller configured to: cause the inspection energy to be transmitted from the transmit sensor array through an incident surface of the object and through a volume of the object; receive signal data indicative of signal energy received from the object at two or more locations of the object; calculate a Fourier transform on the signal data to separate frequency characteristics from phase characteristics of the signal data; calculate a shift in frequency distribution of the frequency characteristics between the signal energy at the two or more locations of the object; generate an intensity map or response score based on the shift in frequency distribution; and determine at least one microtextured region (MTR) score indicative of MTRs present within a subsurface volume of the object based at least in part on the shift in frequency distribution of the signal energy.

[0605] 1. An inspection system for inspecting an object, the inspection system comprising: a transmit sensor array of one or more active elements configured to introduce inspection energy into the object; and a controller in operative communication with the transmit sensor array, the controller causing the inspection energy to be transmitted from the transmit sensor array through an incident surface of the object and through a volume of the object; the controller configured to receive an intensity map associated with the object, the intensity map being generated based on signal data representing the inspection energy received from the object; quantify at least one of a size, shape, or texture of the intensity map; and determine at least one microtextured region (MTR) score indicative of an MTR within a subsurface volume of the object based at least in part on the size, shape, or texture of the intensity map.

[0606] An inspection system for inspecting an object, the inspection system comprising: a transmit sensor array of one or more active elements configured to introduce inspection energy into the object; and a controller in operative communication with the transmit sensor array, the controller configured to cause the inspection energy to be transmitted from the transmit sensor array through an incident surface of the object and through a volume of the object, receive signal data indicative of signal energy received from the object, decompose the signal data into a plurality of components, select one or more components from the plurality of components indicative of an MTR, and generate at least one intensity map or response score for the object based on the one or more components indicative of the MTR.

[0607] 1. An inspection system for inspecting an object, the inspection system comprising: a transmit sensor array of one or more active elements configured to introduce inspection energy into the object; and a controller in operative communication with the transmit sensor array, the controller configured to cause the inspection energy to be transmitted from the transmit sensor array through an incident surface of the object and through a volume of the object; receive signal data representing the inspection energy received from the object; generate an intensity map or response score for the object, the intensity map representing the signal data arranged corresponding to a spatial arrangement of the signal data; and determine at least one microtextured region (MTR) score indicative of MTRs present within the subsurface volume of the object based at least in part on the intensity map or response score.

[0608] 1. An inspection system for inspecting an object, the inspection system comprising: one or more inspection sensors configured to acquire surface or near-surface inspection data of the object at one or more inspection frequencies, the one or more inspection sensors including at least one of an eddy current sensor or a polarization sensor; and a controller in operative communication with the one or more inspection sensors, the controller configured to receive the surface or near-surface inspection data from the one or more inspection sensors and analyze the surface or near-surface inspection data of the object using at least one of an information entropy determination approach, a joint probability or co-occurrence matrix, or space-time recognition and / or analysis, and a frequency distribution of the surface or near-surface inspection data.

[0609] 1. An inspection system for inspecting an object, the inspection system comprising: an inspection device including an excitation source and a receiver, the inspection device configured to generate at least one of an acoustic wave, an electromagnetic wave, or an induced current; and a controller in operative communication with the excitation source, the controller configured to use the excitation source to generate at least one of the acoustic wave, the electromagnetic wave, or the induced current that interacts with the object, acquire signal data indicative of an interaction between the object and the at least one of the acoustic wave, the electromagnetic wave, or the induced current, and determine at least one response score representative of a microtextured region (MTR) characteristic present for a portion of a volume of the object.

[0610] 1. An inspection system for inspecting a subject, the inspection system comprising: a transmit sensor array of one or more active elements configured to introduce inspection energy into the subject; and a controller in operative communication with the transmit sensor array, the controller configured to: cause energy to be transmitted from the transmit sensor array through an incident surface of the subject and through a volume of the subject; receive signal data indicative of signal energy received from the subject at two or more locations of the subject; generate an intensity map or response score based on the signal data at the two or more locations of the subject; and determine at least one microtextured region (MTR) score indicative of MTRs present within a subsurface volume of the subject based at least in part on variations in the signal energy.

[0611] an inspection system comprising: at least one acoustic transmitter operable to propagate acoustic waves into an interaction volume of a subject, at least a portion of the interaction volume being disposed below a surface of the subject; at least one acoustic receiver positioned to detect the acoustic waves; and a controller in operative communication with the at least one acoustic transmitter and the at least one acoustic receiver, the controller configured to activate the at least one acoustic transmitter to transmit acoustic waves comprising at least one of shear waves, longitudinal waves, or mixed-mode waves; receive signal data indicative of two or more acoustic signals associated with two or more locations in time, space, or frequency, the two or more acoustic signals being received from the at least one acoustic receiver; calculate quantified descriptions of the two or more acoustic signals relative to each other; and determine microstructural properties associated with the subject based at least in part on the quantified descriptions of the two or more acoustic signals.

[0612] 1. A method of inspection, comprising: receiving a first set of signal data from an inspection device, the first set of signal data associated with one or more volumes within the object; receiving a second set of signal data from a near-surface inspection device, the second set of signal data associated with one or more surfaces or cross sections of the object; determining a first response score or intensity map for the object using the first set of signal data; determining a second response score or intensity map for the object using the second set of signal data; and determining at least one microtexture region (MTR) score indicative of MTRs present within a subsurface volume of the object based at least in part on the first response score and the second response score or intensity map.

[0613] 10. The inspection system of any preceding claim, wherein the transmission sensor array comprises an acoustic transducer array including a plurality of acoustic transducers acting as acoustic transmitters and acoustic receivers, and the signal data is received from the acoustic transducer array.

[0614] 10. The inspection system of any preceding claim, further comprising a receiving sensor array of one or more elements configured to receive inspection energy from the specimen, wherein the controller is configured to receive signal data from the receiving sensor array.

[0615] 10. The inspection system of any preceding claim, wherein the inspection energy includes at least one of a longitudinal wave, a shear wave, a Rayleigh wave, or a mixed-mode acoustic wave.

[0616] 10. The inspection system of any preceding claim, wherein the controller is further configured to generate an intensity map based on upshifting or downshifting frequency components of the signal data.

[0617] 10. The inspection system of any preceding claim, wherein the controller is further configured to quantify the intensity map using a cumulative distribution function of the size of regions having uniform frequency shifts.

[0618] 10. The inspection system of any preceding claim, wherein the controller is further configured to quantify the intensity map using a statistical measure of the size of regions having uniform frequency shifts.

[0619] 10. The inspection system of any preceding claim, wherein the controller is further configured to quantify the intensity map using an Otsu threshold.

[0620] 10. The inspection system of any preceding claim, wherein the controller is further configured to generate the intensity map using a Gaussian mixture.

[0621] 10. The inspection system of any preceding claim, wherein the controller is further configured to quantify the intensity map based at least in part on enumeration of spots exhibiting changes in frequency distribution below or above a threshold.

[0622] 10. The inspection system of any preceding claim, wherein the transmit sensor array steers a direction of the inspection energy, and the controller is further configured to compare an upshift or downshift in the frequency distribution with other steered angles.

[0623] 10. The inspection system of any preceding claim, wherein the controller is further configured to quantify the one or more frequency peaks excluding a primary frequency peak, and to quantify a rate of occurrence of the one or more frequency peaks.

[0624] 10. The inspection system of claim 1, wherein the controller is further configured to quantify one or more frequency peaks excluding a primary frequency peak, the one or more frequency peaks having an amplitude above a fixed percentage of the magnitude of the primary frequency peak, and to quantify a rate of occurrence of the one or more frequency peaks.

[0625] 10. The inspection system of any preceding claim, wherein the controller is further configured to form a spectrogram of the frequency distribution of the signal energy.

[0626] 10. The inspection system of any preceding claim, wherein the controller is further configured to generate an intensity map based on a continuous wavelet transform or a short-time Fourier transform over a particular time period.

[0627] 10. The inspection system of any preceding claim, wherein the controller is further configured to generate the intensity map based on at least one of a structural similarity function of the signal data, a cross-coherence of the signal data, or a mean squared error of the signal data.

[0628] 10. The inspection system of any preceding claim, wherein the controller is further configured to generate an intensity map based on the phase shift.

[0629] 10. The inspection system of any preceding claim, wherein the controller is further configured to generate the intensity map by calculating features in the signal data having phase coherence and determining statistics of the features.

[0630] 10. The inspection system of any preceding claim, wherein the controller is further configured to generate an intensity map based on a Pearson correlation coefficient.

[0631] 10. The inspection system of any preceding clause, wherein the controller is further configured to compare the intensity map of the signal energy with a reference intensity map and calculate a difference or variance score.

[0632] 10. The inspection system of any preceding claim, wherein the transmission sensor array is an acoustic transducer array including a plurality of acoustic transducers acting as acoustic transmitters and acoustic receivers, and the signal data is received from the acoustic transducer array.

[0633] 10. The inspection system of any preceding claim, further comprising a receiving sensor array of one or more elements configured to receive inspection energy from the specimen, wherein the controller is configured to receive signal data from the receiving sensor array.

[0634] 10. The inspection system of any preceding claim, wherein the inspection energy includes at least one of a longitudinal wave, a shear wave, a Rayleigh wave, or a mixed-mode acoustic wave.

[0635] 10. The inspection system of any preceding claim, wherein the inspection energy is electromagnetic energy and the transmit sensor array is an eddy current transducer array.

[0636] 10. The inspection system of any preceding claim, wherein the controller is configured to perform a Hilbert transform to isolate amplitude and phase characteristics of the signal data.

[0637] 10. The inspection system of any preceding clause, wherein the controller is further configured to extract a distribution of amplitude values ​​and fit the distribution of signal amplitudes to at least one of a Nakagami distribution, a Gamma distribution, a generalized Nakagami distribution, a Rayleigh distribution, a Homodyne K (HK) distribution, or a Rician distribution, with certain parameters describing the shape of the distribution.

[0638] 10. The inspection system of any preceding claim, wherein the controller is further configured to fit a shape-describing function to the signal data, fit characteristic shape parameters to the signal data, calculate an area under a signal curve of the signal data, or calculate an element-by-element difference of a cumulative sum of the signal data raised to a specified exponent.

[0639] 10. The inspection system of any preceding claim, wherein the shape-describing function comprises a parametric function, the parametric function comprising at least one of a polynomial function, an exponential function, a parametric function having an exponential term and a polynomial term, a parametric function associated with a probability density function (PDF) or a cumulative distribution function (CDF) of a statistical distribution, or a parametric function associated with a physics-based model.

[0640] 10. The inspection system of any preceding claim, wherein the characteristic shape parameter comprises at least one of a signal peak height, a signal peak position, a signal peak width, a signal peak prominence, an integrated area of ​​a signal envelope, or an integrated area of ​​a parametric function.

[0641] 10. The inspection system of any preceding claim, wherein the controller is further configured to extract amplitudes from the signal data and calculate one or more higher order statistics of the amplitudes.

[0642] 10. The inspection system of any preceding claim, wherein the controller is further configured to fit a descriptive characteristic function to the shape of the distribution of values ​​for a given characteristic of the signal energy.

[0643] 10. The inspection system of any preceding claim, wherein the controller is further configured to measure a gradient of a shape of a distribution of values ​​for a given characteristic of the signal energy.

[0644] 10. The inspection system of any preceding claim, wherein the controller is further configured to calculate an envelope for the signal data using at least one of a Hilbert transform, a moving root-mean-square (RMS) filter, an infinite impulse response (IIR) filter, a finite impulse response (FIR) filter, or a spline fit.

[0645] 10. The inspection system of any preceding claim, wherein the controller is further configured to calculate a given characteristic of the signal data based on an envelope for the signal data.

[0646] 10. The inspection system of any preceding claim, wherein the given characteristic of the signal data includes at least one of amplitude, phase, frequency, or shape associated with an envelope.

[0647] 10. The inspection system of any preceding claim, wherein the controller is further configured to calculate a distribution of envelope magnitudes.

[0648] The inspection system of any preceding paragraph, wherein the controller is further configured to generate at least one of an intensity map or a response score for the subject based on a quantified description of the distribution of values ​​for a given characteristic of the signal data, and to correlate the intensity map or the response score with at least one MTR score.

[0649] 10. The testing system of any preceding clause, wherein the controller is further configured to assign an accept or reject identifier to the subject based on at least one MTR score.

[0650] 10. The inspection system of any preceding claim, wherein the controller is in communication with a manufacturing system, the controller further configured to adjust the manufacturing system based on the at least one MTR score.

[0651] 10. The inspection system of any preceding claim, wherein the transmission sensor array comprises an acoustic transducer array including a plurality of acoustic transducers acting as acoustic transmitters and acoustic receivers, and the signal data is received from the acoustic transducer array.

[0652] 10. The inspection system of any preceding claim, further comprising a receiving sensor array of one or more elements configured to receive inspection energy from the specimen, wherein the controller is configured to receive signal data from the receiving sensor array.

[0653] 10. The inspection system of any preceding claim, wherein the inspection energy includes at least one of a longitudinal wave, a shear wave, a Rayleigh wave, or a mixed-mode acoustic wave.

[0654] 10. The inspection system of any preceding claim, wherein the controller is further configured to generate an intensity map based on upshifting or downshifting frequency components of the signal data.

[0655] 10. The inspection system of any preceding claim, wherein the controller is further configured to quantify the intensity map using a cumulative distribution function of the size of regions having uniform frequency shifts.

[0656] 10. The inspection system of any preceding claim, wherein the controller is further configured to quantify the intensity map using a statistical measure of the size of regions having uniform frequency shifts.

[0657] 10. The inspection system of any preceding claim, wherein the controller is further configured to generate the intensity map using an Otsu threshold.

[0658] 10. The inspection system of any preceding claim, wherein the controller is further configured to generate the intensity map using a Gaussian mixture.

[0659] 10. The inspection system of any preceding claim, wherein the controller is further configured to quantify the intensity map based at least in part on enumeration of spots exhibiting changes in frequency distribution below or above a threshold.

[0660] 10. The inspection system of any preceding claim, wherein the transmit sensor array steers a direction of the inspection energy, and the controller is further configured to compare an upshift or downshift in the frequency distribution with other steered angles.

[0661] 10. The inspection system of any preceding claim, wherein the controller is further configured to quantify the one or more frequency peaks excluding a primary frequency peak, and to quantify a rate of occurrence of the one or more frequency peaks.

[0662] 10. The inspection system of claim 1, wherein the controller is further configured to quantify one or more frequency peaks excluding a primary frequency peak, the one or more frequency peaks having an amplitude above a fixed percentage of the magnitude of the primary frequency peak, and to quantify a rate of occurrence of the one or more frequency peaks.

[0663] 10. The inspection system of any preceding claim, wherein the controller is further configured to form a spectrogram of the frequency distribution of the signal energy.

[0664] 10. The inspection system of any preceding claim, wherein the controller is further configured to generate an intensity map based on a continuous wavelet transform or a short-time Fourier transform over a particular time period.

[0665] 10. The inspection system of any preceding claim, wherein the controller is further configured to generate the intensity map based on at least one of a structural similarity function of the signal data, a cross-coherence of the signal data, or a mean squared error of the signal data.

[0666] 10. The inspection system of any preceding claim, wherein the controller is further configured to generate an intensity map based on the phase shift.

[0667] 10. The inspection system of any preceding claim, wherein the controller is further configured to generate the intensity map by calculating features in the signal data having phase coherence and determining statistics of the features.

[0668] 10. The inspection system of any preceding claim, wherein the controller is further configured to generate an intensity map based on a Pearson correlation coefficient.

[0669] 10. The inspection system of any preceding clause, wherein the controller is further configured to compare the intensity map of the signal energy with a reference intensity map and calculate a difference or variance score.

[0670] 10. The inspection system of any preceding claim, wherein the transmission sensor array is an acoustic transducer array including a plurality of acoustic transducers acting as acoustic transmitters and acoustic receivers, and the signal data is received from the acoustic transducer array.

[0671] 10. The inspection system of any preceding claim, further comprising a receiving sensor array of one or more elements configured to receive inspection energy from the specimen, wherein the controller is configured to receive signal data from the receiving sensor array.

[0672] 10. The inspection system of any preceding claim, wherein the inspection energy includes at least one of a longitudinal wave, a shear wave, a Rayleigh wave, or a mixed-mode acoustic wave.

[0673] 10. The inspection system of any preceding claim, wherein the inspection energy is electromagnetic energy and the transmit sensor array is an eddy current transducer array.

[0674] 10. The inspection system of any preceding claim, wherein the controller is configured to perform a Hilbert transform on the signal data to isolate amplitude and phase characteristics of the signal data.

[0675] 10. The inspection system of any preceding claim, wherein the controller is further configured to digitize the intensity map into a matrix of values ​​to characterize the size, shape, or texture of the intensity map.

[0676] 10. The inspection system of any preceding claim, wherein the controller is further configured to use multiple matrices that represent the likelihood that two energy signals separated by a particular offset distance have the same characteristics to characterize the size, shape, or texture of the intensity map.

[0677] 10. The inspection system of any preceding claim, wherein the controller is configured to apply a weighting function to the matrix of values.

[0678] 10. The inspection system of claim 1, wherein the controller is further configured to select two or more characteristics of the signal data, divide each characteristic into a range of levels, calculate a matrix containing the joint probability of observing each combination of the range of levels, apply weights and adjustment functions to the matrix, and generate at least one MTR score based on the matrix.

[0679] 10. The inspection system of claim 1, wherein the controller is further configured to quantify the texture of the intensity map by selecting one or more characteristics of the intensity map, dividing each characteristic into a plurality of bins, specifying an offset relationship between adjacent data points, calculating a matrix representing the probability that the point and adjacent points have values ​​within the plurality of bins, applying weights and adjustment functions to the matrix, generating one or more response scores based on the matrix, and generating at least one MTR score based on the one or more response scores.

[0680] 10. The inspection system of any preceding claim, wherein the size, shape, or texture of the intensity map is determined by segmenting the intensity map using at least one of an Otsu threshold, a Voronoi Otsu threshold, a Gaussian mixture model, a Chan Vese model, a watershed, or a marching squares algorithm.

[0681] 10. The inspection system of any preceding claim, wherein the transmission sensor array comprises an acoustic transducer array including a plurality of acoustic transducers acting as acoustic transmitters and acoustic receivers, and the signal data is received from the acoustic transducer array.

[0682] 10. The inspection system of any preceding claim, wherein the inspection energy includes at least one of a longitudinal wave, a shear wave, a Rayleigh wave, or a mixed-mode acoustic wave.

[0683] 10. The inspection system of any preceding claim, wherein the inspection energy is electromagnetic energy and the transmit sensor array is an eddy current transducer array.

[0684] 10. The inspection system of any preceding claim, wherein the controller is further configured to perform at least one of a Hilbert transform to isolate amplitude and phase characteristics of the signal data.

[0685] 10. The inspection system of any preceding claim, wherein decomposition of the signal data into a plurality of components comprises empirical mode decomposition.

[0686] 10. The inspection system of any preceding claim, wherein the decomposition of the signal data into a plurality of components comprises variational mode decomposition.

[0687] 10. The inspection system of any preceding claim, wherein the decomposition of the signal data into a plurality of components comprises a Hilbert-Huan transform.

[0688] 10. The inspection system of any preceding claim, wherein the transmission sensor array comprises an acoustic transducer array including a plurality of acoustic transducers acting as acoustic transmitters and acoustic receivers, and the signal data is received from the acoustic transducer array.

[0689] 10. The inspection system of any preceding claim, further comprising a receiving sensor array of one or more elements configured to receive inspection energy from the specimen, wherein the controller is configured to receive signal data from the receiving sensor array.

[0690] 10. The inspection system of any preceding claim, wherein the inspection energy includes at least one of a longitudinal wave, a shear wave, a Rayleigh wave, or a mixed-mode acoustic wave.

[0691] 10. The inspection system of any preceding claim, wherein the inspection energy is electromagnetic energy and the transmit sensor array is an eddy current transducer array.

[0692] 10. The inspection system of any preceding claim, wherein the controller is configured to perform a short-time Fourier transform or a continuous wavelet transform to calculate time-frequency characteristics of the signal data.

[0693] 10. The inspection system of any preceding claim, wherein the controller is further configured to perform a Hilbert transform to isolate amplitude and phase characteristics of the signal data.

[0694] 10. The inspection system of any preceding claim, wherein the controller is further configured to perform a Fourier transform to isolate frequency and phase characteristics of the signal data.

[0695] 10. The inspection system of any preceding claim, wherein the controller is further configured to use the intensity map to segment the object into separate regions using one or more of a Gaussian mixture model, a threshold, or marching squares / cubes, and associate the signal data with the separate regions to generate segmented data.

[0696] The controller may be further configured to calculate one or more spatial statistics for each feature in the segmented data, the spatial statistics including one or more of size, shape, aspect ratio, surface area, or perimeter.

[0697] The inspection system of any preceding clause, wherein the controller is further configured to calculate coherent features in the signal data, determine statistics of the coherent features, and generate one or more response scores indicative of the MTR in the subject based on the statistics of the coherent features.

[0698] The inspection system of any preceding paragraph, wherein the controller is further configured to: reconstruct paths from a region within the subject to each element of the receive acoustic transducer array; compare waveforms received by each element of the receive acoustic transducer array; apply a weighting function to each waveform; delay the waveforms at each point within the subject; sum the waveforms into a composite value; and generate one or more MTR scores indicative of the MTR within the subject based on the composite value.

[0699] 10. The inspection system of any preceding claim, wherein the controller is further configured to calculate a metric of peakiness of the signal data, the metric comprising one or more of skewness, roll-off, entropy, kurtosis, or prominence.

[0700] 10. The inspection system of claim 1, wherein the controller is further configured to aggregate the signal energy characteristics across the group of segmented regions to generate an aggregated signal energy characteristic and calculate one or more of skewness, roll-off, entropy, kurtosis, or prominence of the aggregated signal energy characteristic.

[0701] 10. The inspection system of any preceding clause, wherein the controller is further configured to segment the signal data across any dimension using one or more of a Gaussian mixture model, a threshold, or marching squares / cubes.

[0702] 10. The inspection system of any preceding claim, wherein the controller is further configured to rescale the signal data to a standardized magnitude or time framework for direct comparison with other signal data.

[0703] 10. The inspection system of claim 1, wherein the intensity map represents one or more of statistical entropy, fuzzy entropy, gradient entropy, fluctuation variance entropy, or phase entropy, which represent the distribution of the received signal energy pattern.

[0704] 10. The inspection system of any preceding claim, wherein the controller is further configured to calculate a measure of information entropy based on at least one of the signal data, the reconstructed shape / area / volume, or statistics of the signal data or the reconstructed shape / area / volume.

[0705] 10. The inspection system of any preceding claim, wherein the controller is further configured to select a window over each vector associated with the signal data, the window length and embedding dimension determining the size of the pattern within each vector.

[0706] 10. The inspection system of any preceding claim, wherein the controller is further configured to calculate a containment metric based on whether the pattern within the window matches a known pattern and count occurrences for the known pattern.

[0707] 10. The inspection system of any preceding claim, wherein the controller is further configured to calculate a probability that the known pattern will appear in the population of received signals.

[0708] 10. The inspection system of any preceding clause, wherein the controller is further configured to generate second intensity maps or second response scores for additional time scales, window lengths, and embedding dimensions.

[0709] 10. The inspection system of any preceding claim, wherein the transmission sensor array is an acoustic transducer array including a plurality of acoustic transducers acting as acoustic transmitters and acoustic receivers, and the signal data is received from the acoustic transducer array.

[0710] 10. The inspection system of any preceding claim, wherein the inspection energy includes at least one of a longitudinal wave, a shear wave, a Rayleigh wave, or a mixed-mode acoustic wave.

[0711] 10. The inspection system of any preceding claim, further comprising a receive sensor array configured to receive inspection energy from the subject, wherein the transmit sensor array and the receive sensor array are acoustic transducer arrays.

[0712] 10. The inspection system of any preceding claim, wherein the inspection energy is electromagnetic energy and the transmit sensor array is an eddy current transducer array.

[0713] 10. The inspection system of any preceding claim, wherein the controller is further configured to calculate phase variation at a particular point in time.

[0714] 10. The inspection system of any preceding claim, wherein the controller is further configured to perform a Hilbert transform to isolate amplitude and phase characteristics of the signal data.

[0715] 10. The inspection system of any preceding claim, wherein the controller is further configured to calculate amplitude variation at a particular point in time.

[0716] 10. The inspection system of any preceding claim, wherein the controller is further configured to compare the signal data to a signal template.

[0717] 10. The inspection system of any preceding claim, wherein the controller is further configured to perform a Fourier transform to isolate frequency and phase characteristics of the signal data.

[0718] 10. The inspection system of any preceding clause, wherein the controller is further configured to correlate the signal data with the signal template and calculate a difference or variation score.

[0719] 10. The inspection system of any preceding claim, wherein the controller is further configured to measure an area coverage of the uniform received energy characteristic.

[0720] 10. The inspection method of any preceding clause, wherein the near-surface inspection device comprises at least one of an ultrasonic sensor configured to generate acoustic surface waves, an ultrasonic sensor configured to interrogate a near-surface volume of the sample, an eddy current sensor, or a polarimetric sensor.

[0721] 2. The inspection method of any preceding claim, wherein analyzing the second set of signal data includes analyzing the second set of signal data using at least one of a measure of entropy determined based on a two-dimensional representation of the second set of signal data across one or more surfaces or cross sections of the object, a measure of entropy determined based on a time domain of the second set of signal data, a measure of entropy determined based on a frequency domain of the second set of signal data, a measure of entropy determined based on an acoustic waveform, or a joint probability determination.

[0722] 2. The inspection system of claim 1, wherein the controller is further configured to analyze the signal data using one or more metrics of statistical properties of the signal data, entropy metrics based on properties of the signal data, joint probability or co-occurrence metrics based on properties of the signal data, phase coherence of the signal data, metrics derived from principal component analysis, variation of the signal data at different locations, times, excitation frequencies or waveforms, or peakiness or flatness of the signal data,...

Claims

1. 1. An inspection system for inspecting a specimen, the inspection system comprising: a transmit sensor array of one or more active elements configured to introduce test energy into the subject; a controller in operative communication with the transmit sensor array, the controller comprising: causing the interrogation energy to be transmitted from the transmit sensor array through an incident surface of the subject and through a volume of the subject; receiving signal data indicative of signal energy received from the subject at two or more locations of the subject; computing a Fourier transform of the signal data to separate frequency characteristics from phase characteristics of the signal data; calculating a shift in frequency distribution of the frequency characteristics between the signal energy at the two or more locations of the subject; generating an intensity map or response score based on said shift in frequency distribution; determining at least one microtextured region (MTR) score indicative of a MTR present within a subsurface volume of the subject based at least in part on the shift in the frequency distribution of the signal energy; The controller and An inspection system comprising:

2. The inspection system of claim 1 , wherein the transmission sensor array comprises an acoustic transducer array including a plurality of acoustic transducers acting as acoustic transmitters and acoustic receivers, and the signal data is received from the acoustic transducer array.

3. 10. The inspection system of claim 1, further comprising a receiving sensor array of one or more elements configured to receive the inspection energy from the object under test, wherein the controller is configured to receive the signal data from the receiving sensor array.

4. The inspection system of claim 1 , wherein the inspection energy comprises at least one of a longitudinal wave, a shear wave, a Rayleigh wave, or a mixed-mode acoustic wave.

5. The inspection system of claim 1 , wherein the controller is further configured to generate an intensity map based on upshifting or downshifting frequency components of the signal data.

6. The inspection system of claim 5 , wherein the controller is further configured to quantify the intensity map using a cumulative distribution function of a size of regions having uniform frequency shifts.

7. The inspection system of claim 5 , wherein the controller is further configured to quantify the intensity map using a statistical measure of the size of regions having uniform frequency shifts.

8. The inspection system of claim 5 , wherein the controller is further configured to generate the intensity map using an Otsu threshold.

9. The inspection system of claim 5 , wherein the controller is further configured to generate the intensity map using a Gaussian mixture.

10. The inspection system of claim 5 , wherein the controller is further configured to quantify the intensity map based at least in part on enumeration of spots that exhibit changes in frequency distribution below or above a threshold.

11. The inspection system of claim 5 , wherein the transmit sensor array steers a direction of inspection energy, and the controller is further configured to compare an upshift or downshift in frequency distribution with other steered angles.

12. The inspection system of claim 5 , wherein the controller is further configured to quantify one or more frequency peaks excluding a primary frequency peak, and to quantify a rate of occurrence of the one or more frequency peaks.

13. 6. The inspection system of claim 5, wherein the controller is further configured to quantify one or more frequency peaks excluding a primary frequency peak, the one or more frequency peaks having an amplitude above a fixed percentage of the magnitude of the primary frequency peak, and to quantify a rate of occurrence of the one or more frequency peaks.

14. The inspection system of claim 1 , wherein the controller is further configured to form a spectrogram of the frequency distribution of the signal energy.

15. The inspection system of claim 1 , wherein the controller is further configured to generate an intensity map based on a continuous wavelet transform or a short-time Fourier transform over a specific time period.

16. 16. The inspection system of claim 15, wherein the controller is further configured to generate an intensity map based on at least one of a structural similarity function of the signal data, a cross-coherence of the signal data, or a mean square error of the signal data.

17. The inspection system of claim 15 , wherein the controller is further configured to generate an intensity map based on the phase shift.

18. The inspection system of claim 15 , wherein the controller is further configured to generate the intensity map by calculating features in the signal data having phase coherence and determine statistics of the features.

19. The inspection system of claim 15 , wherein the controller is further configured to generate an intensity map based on a Pearson correlation coefficient.

20. The inspection system of claim 15 , wherein the controller is further configured to compare the signal energy intensity map to a reference intensity map and calculate a difference or variance score.

Images