Patent Application
20190369059
Reference is made to German Application No. 10 2018 208 824.1 filed Jun. 5, 2018, which is incorporated herein by reference in its entirety.
The invention pertains to a method for the nondestructive examination of a test specimen by use of ultrasound, in which ultrasonic waves are coupled into the test specimen with a plurality of ultrasonic transducers and ultrasonic waves reflected within the test specimen are received by a plurality of ultrasonic transducers and converted into ultrasonic signals, wherein the ultrasonic signals are stored and subsequently divided into a plurality of discrete signal information by use of a propagation time-based, phase-corrected superposition in the course of ultrasonic signal data processing, and wherein each discrete signal information is respectively assigned to a volume element within the test specimen, referred to as a voxel hereinafter.
The procedure in the nondestructive examination of a test specimen by use of ultrasound, for example for the purpose of testing materials with respect to defects such as cracks, inclusions or other material inhomogeneities, comprises the coupling of ultrasonic waves into the test specimen, the detection of the ultrasonic waves that are transmitted through the test specimen or reflected, diffracted, at least one of scattered and refracted within the test specimen, as well as the evaluation of the ultrasonic waves that are converted into ultrasonic signals.
The above-described and generally known examination method makes it possible to detect and evaluate the respective transmission or reflection properties of a test specimen for ultrasonic waves. Due to a corresponding evaluation of the received ultrasonic signals, defects located within a test specimen such as material cracks, foreign material inclusions or material boundaries are in this method mapped in the form of regions with different reflection properties. The position, shape and size of the defects can be three-dimensionally displayed with high spatial resolution.
This method has diverse fields of application. For example, the method is used for testing and measuring homogeneity or strength properties of structural components (concrete walls, ceiling or wall elements, etc.) or for detecting cracks, that is in railway wheels or aircraft components. In this context, the use of the method for testing composite materials is of particular interest.
At the present time, ultrasonic material testing methods are typically characterized by the use of multi-transducer measuring systems or so-called transducer arrays, by use of which the internal structure of a test specimen including relevant material discontinuities or material defects can be visualized in conjunction with suitable signal processing and image reconstruction techniques.
Known defect detection techniques that are based on the coupling, detection and evaluation of ultrasonic waves, particularly so-called transducer array techniques, respectively concern the evaluation of ultrasonic wave amplitude information and the analysis of the transmission characteristics of the test specimen by use of the transmission technique or the analysis of the reflection characteristics of the test specimen by use of pulse-echo technique.
A typical test assembly for ultrasonic wave material testing has an ultrasonic transducer or preferably multiple ultrasonic transducers, which couple the ultrasonic waves into the test specimen from one side thereof and receive the returning ultrasonic echo signals, that is the reflected ultrasonic echo signals. The ultrasonic waves reflected in the interior of the test specimen are detected as echo signals and recorded. In this case, only the signal amplitudes of the recorded echo signals are evaluated. The amplitude of an echo signal is directly proportional to the reflection capacity or reflectivity of the reflector and depends on its size, shape, orientation and acoustic properties. In this context, the term acoustic properties primarily refers to the acoustic impedance, which is the product of the density and the acoustic velocity in the material. The greater the difference between the acoustic impedance values within the test specimen and the reflector, the greater the reflectivity of the reflector. Larger reflectors with comparable acoustic properties generate stronger echo signals than smaller reflectors.
One known method for the nondestructive ultrasonic examination of a test specimen is described in EP publication 1 820 012 B1. In this method, ultrasonic waves are emitted into the test specimen by use of a plurality of ultrasonic transducers, which are arranged on a test specimen surface in a linear or array-shaped manner and also capable of detecting the ultrasonic wave components reflected within the test specimen, as well as converting these reflected ultrasonic wave components into echo signals. When the measurement is carried out, different emitter/receiver constellations are activated and measuring signals are respectively recorded. Preferably all ultrasonic transducers detect ultrasonic signals and forward the detected ultrasonic signals to be recorded in order to respectively receive the reflected ultrasonic wave components. Depending on the test specimen and the metrological objectives, the examination of the test specimen takes place in the course of a plurality of individual measuring cycles, in which a corresponding number of measuring signals is respectively obtained and stored.
The evaluation of the ultrasonic signals takes place offline, after the coupling and detection of the ultrasonic waves has taken place, based on a reconstruction algorithm, which is selected in dependence on a virtually pre-definable acoustic irradiation angle or a virtual focusing of the ultrasonic waves coupled into the test specimen and applied to the stored ultrasonic signals. Three-dimensional images of the respective transmission and reflection properties of the test specimen can be calculated from the stored ultrasonic signals with the aid of such reconstruction algorithms without requiring any additional ultrasonic measurements. This reconstruction principle is based on the use of Synthetic Aperture Focusing Technique (SAFT), which projects all received ultrasonic signals on a common time axis. In this case, all ultrasonic signals reflected by a particular reflector can be added up in an equiphase manner. A subsequent reconstruction of arbitrary acoustic irradiation angles results from a phase-shifted addition of the reception signals of different ultrasonic receivers. This generally known offline evaluation makes possible synthetic reconstruct at nearly any wave angle and to thereby perform an ultrasonic sweep through the stored data set.
The detection capability of the ultrasonic examination is significantly limited, particularly when conventional testing and analyzing techniques are used for examining test specimens that have acoustically inhomogeneous material properties, for example concrete or similar construction materials with an increased acoustic attenuation of the ultrasonic waves and a high level of structural acoustic noise. For example, the conventional amplitude-based signal evaluation does not provide any evaluable or reliable information, for example, on the filling level of a jacket tube of steel in a concrete test specimen. That is whether the jacket tube which is embedded within a concrete matrix is filled entirely, partially or not at all. The amplitude of the metrologically acquired echo signal of the ultrasonic waves, which are reflected on the tube with different filling levels, primarily depends on the geometric dimensions of the tube, as well as the impedance difference between the concrete matrix and the steel material, of which the jacket tube is made. However, information on the filling level cannot be obtained in this way.
DE 10 2014 209 773 A1 discloses a method and a device for examining a component, wherein internal structures of the component such as defects are reconstructed dependent on acquired ultrasonic measuring values, by use of SAFT, and one or more structural elements, as well as their position, are determined in the reconstruction. The ultrasonic inspection is simulated by use of a simulation model for the individual structural elements and the obtained simulation values are compared with the ultrasonic measuring values. Model parameters, on which the simulation is based, are changed in the course of an optimization process in dependence on a difference between the simulation values and the ultrasonic measuring values so that the model parameters describe the constitution of the respective structural element more precisely than prior to the optimization.
DE 10 2013 211 064 A1 describes a SAFT analysis of defects which are near the surface of an object, in which the sound propagation time from a probe positioned at a measuring point to the location of the examined voxel is directly measured and used for determining points in time of amplitudes to be added up in A-image time signals.
DE 10 2012 109 257 A1 discloses a method and a device for generating an ultrasonic image, in which ultrasonic transducer elements of a point focus-controlled 2D ultrasound matrix array emit correlated ultrasonic signals that simultaneously arrive in a focus area due to consideration of corresponding delay times and interfere once at this location. The measuring signals from the focus area are added up with consideration of the respective delay times and a respective amplitude value from the measuring signal sum is assigned in a measuring value matrix for voxels lying in the focus area. The ultrasonic image is generated by using the measuring value matrix.
US published patent application 2012/0055252 A1 discloses a method and a device for detecting radially and axially oriented defects in cylindrical test specimens, wherein a coupling device of the ultrasonic waves has at least one radial and one tangential component. A segment or even the entire volume of the test specimen can be examined for defects by moving the probe along the surface.
WO 2016/087315 A1 discloses an ultrasound imaging method, in which ultrasonic waves are emitted by M ultrasonic transducers and received by N ultrasonic transducers and geometry-dependent artifacts are filtered. For this purpose, the theoretical propagation times Tij(P), which an ultrasonic wave requires from the i-th emitter to the j-th receiver, are calculated for each point P in the area to be examined and for each emitter/receiver pair. In addition, the propagation times Tijgeom for ultrasonic waves, which are reflected on a reflection surface of the test specimen, are calculated for each emitter/receiver pair. Only the amplitudes Sij(t) and the propagation times Tij(P) of which significantly differ from Tijgeom are added up for each point P.
The invention is a method for the nondestructive examination of a test specimen by use of ultrasound, wherein the method makes possible examination of test specimens with acoustically nonhomogeneous materials or material areas, and in particular test specimens with embedded jacket tubes that may have different filling levels, over the internal structure of the test specimen with enhanced detection capability.
The inventive method is based on the realization that the information content of the ultrasonic signals obtained by ultrasonic transducers not only include the time-dependent amplitude information, but also other relevant signal contents such as the wave type, polarization, frequency, modulation, etc., of the respectively received ultrasonic waves. The additional signal contents, which are referred to as signal information contents below, include additional information on the test object to be examined.
With the method of the invention for the nondestructive examination of a test specimen by use of ultrasound, ultrasonic waves are coupled into the test specimen with ultrasonic transducers and ultrasonic waves which are reflected within the test specimen are received by ultrasonic transducers and converted into ultrasonic signals. The ultrasonic signals are stored and subsequently divided into discrete signal information by use of a propagation time-based, phase-corrected superposition in the course of ultrasonic signal data processing. Each discrete signal information respectively is assigned to a volume element (voxel) within the test specimen. The method is characterized by the following steps:
Ultrasonic waves in the form of ultrasonic transverse waves or ultrasonic longitudinal waves are respectively emitted or coupled into the test specimen with a two-dimensionally distributed plurality n of individual ultrasonic transducers, which are preferably DPC (DPC—Dry-Point-Contact) transducers arranged on the surface of a test specimen to be examined. Ultrasonic waves, which are reflected at the interfaces between areas with respectively different acoustic density within the test specimen, are received preferably with all in ultrasonic transducers, arranged on the test specimen surface. The correspondingly converted ultrasonic signals are stored and subjected to subsequent ultrasonic signal data processing.
A propagation time-based, phase-corrected superposition of the ultrasonic signals is obtained from necessary discrete signal informations is carried out and a three-dimensional volume area of the test specimen is reconstructed, which preferably is the entire test specimen volume. Each discrete signal information is respectively assigned to a volume element (a voxel within the test specimen). The volume elements obtained due to the superposition are preferably chosen to be equidimensional, that is spatially identical.
The ultrasonic signals generated in the course of the signal detection, as well as the signal information contained therein are subject to technically related noise, and the average noise level thereof is determined.
Based on the determined noise level, all voxels, which respectively have assigned discrete signal information having a noise level with a signal-to-noise ratio of at least 6 decibel which is the determined average noise level, are furthermore determined. In this way, all voxels with a lower S/N ratio are ignored during the further evaluation because it is assumed that no relevant ultrasonic reflection events have taken place in these test specimen volume areas.
All voxels, having a spatial distance from one another of less than or equal to a wavelength of the ultrasonic waves coupled into the test specimen, which is preferably less than or equal to half this wavelength, are furthermore determined. In this way, voxel groups are formed, which respectively comprise an area within the test specimen that surrounds a spatially coherent reflector.
Subsequently, the signal information assigned to each voxel within a combined voxel group is statistically evaluated. The statistical evaluation is based on at least one of the following signal information contents: polarization, frequency, wave type and wave mode. The statistical evaluation is always based on identical signal information contents of the respectively grouped voxels in order to ultimately determine a degree of similarity between the respectively identical signal information contents of the combined voxels.
In contrast to the previous prior art signal evaluation practice, which is exclusively based on signal amplitudes, the invention performs that the signal evaluation based on the additional signal information contents regarding polarization, wave type, frequency and/or wave mode, which are included in the signal information.
The evaluation of the respective signal information assigned to each voxel is carried so that a degree of similarity between respectively identical signal information contents of the signal information is determined. The examination of the test specimen is based on said degree of similarity. In order to determine, for example, the acoustic density of a test specimen with a heterogeneous material composition, at least the degree of similarity of the polarizations included in the signal information is determined in the course of the evaluation of the discrete signal information within a voxel group, wherein said degree of similarity is proportionally related to the acoustic density of the test specimen within the voxel group. Alternatively or additionally, the degree of similarity of the wave type, wave mode and the frequency, which are included in the signal information of the individual voxels combined into a voxel group, can also be determined in order to obtain information on the acoustic properties of the test specimen to be examined.
The degree of reliability of the statistical evaluation can be enhanced by increasing the number of ultrasonic transducers participating in the data acquisition. With respect to the evaluation of the acoustic density of a test specimen to be examined, it holds that the higher the evaluated degree of similarity between respectively identical signal information contents the voxels combined into a voxel group, the lower the acoustic density of the test specimen within the voxel group in question. Vice versa, volume areas within the test specimen having low acoustic density voxel groups, are respectively associated identical signal information contents that are subject to a greater deviation or a greater divergence, respectively.
In comparison with conventional ultrasonic testing methods, the method of the invention improves detection of defects and provides enhanced information on the spatial acoustic density distribution within a test specimen, preferably for the following test applications:
Exemplary embodiments of the invention are described below with reference to the drawings and without limitation of the general inventive idea. In these drawings:
In
In the course of the evaluation, at least part of the test specimen volume is divided into uniformly dimensioned volume elements or voxels 6. Signal information is respectively assigned to The voxels. The signal information results from the superposition of all ultrasonic signals, which are respectively received from a voxel 6 from the n ultrasonic transducers. Only voxels 6 has the signal information which has a metrologically relevant signal-to-noise ratio That is the signal level deviates from the average noise level by at least 6 dB, which is used for the evaluation. Voxels 6, having a spatial distance from one another which preferably is equal to or less than half the ultrasonic wavelength of the ultrasonic waves emitted into the test specimen 1, are furthermore combined into a voxel group 7 for the evaluation. In the example according to
The polarization information per voxel 6, which is included in each information signal, preferably is used for the signal evaluation from which a degree of similarity between the individual polarization information of the voxels combined into a voxel group 7 is determined. The higher the determined degree of similarity, the lower the acoustic density within the examined volume area of the test specimen 1. In this way, the filling level within the jacket tube 4b can be deduced based on the determined degrees of similarity between the individual voxels 6 regarding the polarization information.
The frequency information or the information on the wave type or wave mode naturally also can be used for correspondingly determining a degree of similarity.
| Number | Date | Country | Kind |
|---|---|---|---|
| 10 2018 208 824.1 | Jun 2018 | DE | national |
