Method for determining hard spots in metal test objects, and inspection device and inspection-device assembly therefor

Abstract

The invention relates to a method for determining hard spots in metal test objects, in which method two magnetic-flux-leakage data records of a test region of the test object are generated by means of at least one inspection device via two magnetic-flux-leakage measurements, wherein, furthermore, an eddy-current data record of the test region is created using an eddy-current unit by means of an eddy-current measurement upstream or downstream of the first magnetic-flux-leakage measurement, and subsequently the three data records are analyzed with respect to their signal profile in the test region, the signal profile of the eddy-current data record being used in addition to the signal profiles of at least one of the magnetic-flux-leakage data records in order to analyze the test region for hard spots. The invention also relates to an inspection device for determining hard spots of a test object, which inspection device is designed in particular as a pipeline pig having at least one and preferably at least two propulsion means, and to an inspection-device assembly comprising an inspection device and an EDP device.

Description

[0001] Method for determining hard spots in metallic test objects as well as inspection device and inspection device arrangement therefor

[0002] The invention relates to a method for determining hard spots in metallic test objects, particularly in pipelines, tank walls, and metal sheets. Furthermore, the invention relates to an inspection device for determining hard spots and an inspection device arrangement comprising an inspection device and a computer device.

[0003] Local hardening in metallic test objects, referred to as "hard spots," occurs during the production or subsequent processing of the metallic test objects, for example, in the rolling mill during production or during subsequent repair and welding processes on the test object. Strong local temperature gradients during quenching or tempering of the test object cause microstructural changes that generally lead to undesirable local hardening of the carbonaceous material of the test object. Hard spots can occur simply due to the quenching of the base material during the production of the test object, regardless of the initial temperature.

[0004] During tempering, steel is heated to a low temperature and exposed to this temperature for a specific period of time in order to specifically adjust properties such as hardness and toughness. During tempering, energy is introduced to specifically initiate phase transformations, e.g. from retained austenite to a martensite form. The general purpose of tempering is to relieve stresses built up by prior hardening of the material. To do this, the metal is heated to a temperature below the pearlite point (723 °C). The higher the tempering temperature above 150 °C, the softer the steel becomes and the toughness increases. Transformation processes from martensite / retained austenite to softer microstructures take place.

[0005] The microstructure of the test object influences its hydrogen resistance, which is usually made of steel. This is particularly relevant given the foreseeable increase in the use of existing pipelines for the transport of hydrogen. Hard spots in pipelines or other metallic test objects therefore pose a potential threat to pipeline operation. The failure of hard spots in pipelines is often associated with hydrogen-induced cracking (“HIC”). For example, problems can arise if the hard spot is located beneath a defective coating and the pipe is protected by cathodic protection and is also located in groundwater. Undesirable transformations or effects occur on the metal surface, during which atomic hydrogen can be absorbed into the metal, which can be further enhanced by the cathodic protection.As described above, tempering of the martensitic hardspot results in microstructural transformations that can be assigned to an energy class. These tempered microstructures can have varying susceptibility to HIC.

[0006] From the publication "Multiple Data Inspection of Hard Spots and Cracking",

[0007] Proceedings of the 2014 International Pipeline Conference, Calgary, an approach for determining hard spots is known in which several data sets comprising a low-MFL data set and a high-MFL data set are used to determine hard spots. While the second magnetization for the magnetic flux leakage measurement ("MFL measurement") aims to achieve magnetization in the area of ​​maximum permeability, particularly in the axial direction, the first, higher magnetization aims to saturate the test object in the test area. However, it has been shown that a number of hard spots with a relevant hardness of more than 320 Brinell cannot be detected using the state-of-the-art method. Furthermore, it has been shown that the assumptions to be used for determining the hardness of hard spots found can be improved.A better knowledge of the hard spots enables a better assessment of the test object, of defects that may develop from the hard spots, and a proactive repair of individual hard spots to avoid future critical conditions.

[0008] It is an object of the present invention to improve the state of the art with regard to the aforementioned disadvantages.

[0009] The object is achieved by an article according to claim 1 and by an article according to claim 14 and by an article according to claim 16. In a method according to the invention for determining hard spots in metallic test objects, in particular in pipelines, tank walls and sheets, two MFL data sets of a test area of ​​the test object are generated by means of at least one inspection device via two MFL measurements, wherein in the first MFL measurement with a first magnetization of a first magnetizing device leading in particular to magnetic saturation of the test object, a high MFL data set is generated, and via the second MFL measurement with a magnetization of a second magnetizing device that is lower than the first magnetization, a low MFL data set of the test area is generated, wherein furthermore by means of an eddy current measurement upstream or downstream of the first MFL measurement,in particular via an eddy current unit of the same or another inspection device, an EC data set of the test area is created and then the three data sets are evaluated with regard to their signal curve in the test area, whereby the signal curve of the EC data set is used to evaluate the test area for hard spots in addition to the signal curves of at least one of the MFL data sets.

[0010] The eddy current measurement is then located before or after the first MFL measurement if the eddy current measurement is not carried out in the highly magnetized area. It is carried out outside the highly magnetized area. In particular, the eddy current sensor or the eddy current unit are not located between the

[0011] Magnets of the first magnetizing device provided for magnetization. In particular, the eddy current unit is arranged offset from the first magnetizing device along the longitudinal direction of an associated inspection device. The longitudinal direction runs parallel or congruent with the direction of relative movement between the test object and the inspection device.

[0012] According to a further development of the invention, the MFL measurement is performed in the area of ​​the test object magnetization performed by the second magnetizing device. For this purpose, the EC sensor used for the eddy current measurement can be arranged, in particular, between the longitudinally offset poles of the second magnetizing device. For this purpose, the EC sensor and eddy current sensor are preferably arranged in the same test head of a sensor device.

[0013] Particularly preferably, the eddy current measurement is performed in a low, particularly remanent, magnetic field in the test object, with a magnitude of less than 2 kA / m and even more preferably less than 1 kA / m, so that a material-sensitive eddy current measurement is performed on a magnetically unsaturated material of the test object. A measurement in an unsaturated material, in combination with the MFL measurements, enables improved information acquisition due to the material-sensitive eddy current measurement.

[0014] The invention leverages the discovery that, depending on the tempering temperature and duration, and thus on the energy introduced or removed for the resulting microstructural change, there are hard spots whose permeability curves cannot be sufficiently distinguished from that of the base material at certain magnetizations, and that the permeability curves of hard spots at different tempering temperatures and durations differ characteristically. It was also recognized that in these critical areas, where neither the high nor the low MFL data set provides significant evidence of a hard spot, evidence of hard spots may be present in the eddy current data.

[0015] To determine the hard spot, the measured signal curves are preferably considered relative to the signal of the base material. This means that the stray flux signal of the hard spot measured in the test area is subtracted from the normal or average stray flux signal measured in the base material (either in the test area or outside of it). The base material signal is considered a constant offset. The evaluation is thus based on a relative measurement signal of the stray field or its curve.

[0016] For the first magnetization, field strengths between 9 kA / m and 30 kA / m are used to saturate the test object. For the second magnetization, field strengths of 1 kA / m to 6 kA / m are used. For eddy current measurements, frequencies between 100 kHz and 1000 kHz are preferably used.

[0017] The measurements are carried out in particular at a relative speed of the test object to the pipeline-accessible inspection device of less than or equal to 4 m / s, in particular less than 2 m / s, in order to minimize any speed effects at the low field strengths of the magnetization caused by the second magnetizing device.

[0018] The evaluation is carried out, in particular, in an IT device, i.e., a device for electronic data processing. The evaluation can be carried out at least partially in a control unit of the inspection device, with this control unit then representing at least part of the IT device. The evaluation can also be carried out in an IT device designed separately from the control unit of the inspection device, to which the data records of the inspection device are transferred.

[0019] After the data has been acquired, the method according to the invention can run completely and, in particular, automatically in the computer system. The determination of a test area from the data can also be automated or user-controlled. In particular, areas with relevant, i.e., sufficiently large, signal profiles are automatically defined using computer support.

[0020] Both the MFL data sets and the EC data set are available as data sets with the measurement signal over the location, i.e., over the length and width of the test object. Accordingly, the measurement signal is obtained as a function of the length and width of the selected test area. Alternatively or additionally, the data set to be evaluated can also be simply a signal over the length or width of the test area, whereby measurement signals can be averaged or statistically evaluated over the length or width. In particular, the MFL data sets focus on the relative measurement signal, with the measurement signal being subtracted from the offset for evaluation purposes.

[0021] In particular, opposite polarizations are used for the MFL measurements of the high and low MFL data sets to obtain more precisely defined measurement results, especially in the downstream low MFL measurement. Magnetization in pipelines runs particularly in the longitudinal direction.

[0022] The eddy current measurement is preferably carried out after the low-MFL measurement, whereby a remanent magnetic field from the low-MFL measurement is present in the test object.

[0023] The IT device can comprise one or, if necessary, several computers. The associated computer program can be a single program or it can be a program package comprising a plurality of program modules, which, for example, run in a distributed manner across different parts of the IT device due to resource constraints and can be stored there on respective IT media. A computer has, in particular, the typical means of a data processing unit, such as one or more processors, at least temporary memory (RAM) and generally non-volatile memory, data communication means, in particular for transmitting data from a control unit of the inspection device or for transmitting the evaluated data to an output or further storage unit, and, if necessary, display and / or input units.In accordance with the above, in a further development of the method according to the invention, the EC signal curve is used to locate the hard spot depending on the signal amplitude of the second MFL data set. This is done whenever the amplitude or signal curve in the low-MFL data set is unremarkable, i.e., does not indicate a hard spot, but simultaneously a maximum is present in the EC data set. This applies in particular when a relative signal amplitude of the second MFL data set remains below a low-MFL threshold in the test area.

[0024] The relative signal amplitude is the difference between the offset, formed from the amplitude of the base material from which the pipeline is constructed, minus the amplitude of the measuring signal, particularly in the test area.

[0025] The low-MFL threshold is preferably twice the noise level; in particular, the low-MFL threshold is in a range from 0.1 kA / m to 0.6 kA / m. The range is particularly preferably 0.2 kA / m, which has proven to be a meaningful value for numerous evaluations.

[0026] Preferably, the EC signal curve is used to locate the hard spot, particularly depending on the measurement signal of the first MFL data set, especially if, alternatively or additionally, a relative signal amplitude of the first MFL data set in the test area remains below a high MFL threshold. If the amplitude or signal curve in the high MFL data set in the test area is unremarkable, i.e., does not indicate a hard spot, but at the same time a maximum is present in the EC data set, the presence of a hard spot is inferred.

[0027] The high-MFL threshold is preferably twice the noise level; in particular, the high-MFL threshold is in a range from 0.1 kA / m to 0.6 kA / m. The range is particularly preferably 0.2 kA / m, which has proven to be a meaningful value for numerous evaluations.

[0028] Particularly when considering both the low and high MFL data sets, an extremum, especially a maximum, in the EC data set in the test area, in the absence of significant MFL signals, indicates a hard spot that is not found in the MFL data sets due to the permeability similar to that of the base material. Furthermore, a signal curve of the EC data set above an EC threshold, which is preferably twice the noise level, is preferably used to detect the hard spot. In general, a noise level is the material- and setup-dependent value above which a measurement signal is significant or meaningful.

[0029] Based on the findings explained above and below regarding the different courses of the permeability curves for hard spots that occurred at different temperatures or different temperature levels and different exposure times, and thus due to different energies, a hard spot is identified in particular on the basis of at least one change in the sign of a relative measurement signal between the two MFL data sets. For example, a hard spot is indicated if there is a positive signal in the low MFL data set and a negative (relative measurement) signal in the high MFL data set. A hard spot is also inferred if a positive signal is present in both MFL data sets. A hard spot is also present if there is a negative signal in both MFL measurements.Finally, a hard spot is also present if no significant (relative measurement) signal is present in either MFL data set, but an extremum, particularly a maximum, is present in the eddy current signal. Consequently, hard spots can be identified based on the signal curves of the relative measurement signals in the high and low MFL data sets as well as in the eddy current data set (EC data set).

[0030] In the case of an inspection of a pipeline, the data sets can be generated in particular in the same run or in two or three runs. The data is preferably generated using the same inspection device, which has the sensor units required for the three different data sets. In particular, these are arranged on different segments of the inspection device designed as a pipeline pig, which are connected to one another by couplings, with the eddy current unit required to generate the EC data being arranged upstream or downstream of the first or second magnetizing device. According to a further development of the method according to the invention, the signal curves, in particular the MFL data and the eddy current data (EC data), in the test area are used to determine the hardness of the hard spot. This evaluation also results from the knowledge about the different permeability curves shown further below in Fig. 1.Based on this finding, hard spots generated at different temperatures and over different periods of time can be better classified and determined with regard to their hardness. This is particularly evident from the assignment of the characteristic signal curves of individual hard spots of different material compositions in the high and low MFL data sets as well as in the EC data set to their hardness.

[0031] The determination of the hardness of the hard spot is simplified and thus improved, especially when classification is performed according to energy classes, i.e., when energy-specific calibrations or correlations are used. This results from the realization that individual hard spots with identical or at least sufficiently similar, i.e., indistinguishable, hardnesses can be and have been produced over different periods of time at different tempering temperatures.

[0032] Preferably, the hardness of one or more of the hard spots is determined based on energy classes to which one or more signal curves from at least one of the MFL data sets in the test area are assigned. These energy classes are material-specific with respect to the underlying material, particularly dependent on its carbon content. The relationship between hardness, temperature, and time can be described, for example, using the Hollomon-Jaffe parameter.

[0033] The different amplitude curves of the relative measurement signals identified in a large number of investigations thus lead to a clear classification, whereby within a class the hardness is inferred via correlations between hardness and relative or absolute measurement signal.

[0034] For this purpose, different correlations or correlation curves for hard spots that were generated via different energy effects, in particular temperature effects with different durations at different tempering temperatures, are stored in the computer device or a database assigned to it.

[0035] Preferably, at least three, preferably at least five and in particular six energy classes are used to derive the hardness of different hard spots, in particular wherein a correlation of the hardness as a function of the MFL amplitude, i.e. the measured stray field, is used for each energy class. The energy classes can also include a class of hard spots that may have arisen without additional tempering and only through the original quenching of the base material (0°C class). In a first energy class, for example, the signal curves of hard spots are used that were created at tempering temperatures of up to 100°C over a fixed exposure period, e.g. cooling, of, for example, 15 minutes. A second energy class groups together hard spots that were created at tempering temperatures of more than 100°C, in particular from 150°C to 300°C, again over a defined period of time, for example 15 minutes.A third class includes hard spots that have occurred at tempering temperatures of more than 300°C, particularly at more than 350°C, again over a defined period of time, for example, 15 minutes. Alternatively, the temperature classes can be subdivided even more finely, for example for temperatures from 0°C to 100°C, from 100°C to 200°C (average temperature 150°C), from 200°C to 300°C (average temperature 250°C), then further up to 400°C (average temperature 350°C), then further up to 500°C (average temperature 450°C) and another class up to 600°C (average temperature 550°C). In addition, further classes can be defined for longer effective periods with regard to local cooling or local heating.

[0036] Preferably, at least three carbon-specific classes relating to the base material of the test object are used to derive the hardness of different hard spots. These carbon-specific classes are formed for materials with a carbon content of < 0.1%, from 0.1% to 0.2%, and > 0.2%. Within these classes, energy-specific classes are then formed for the individual tempering temperatures and durations. Preferably, the hardness of the hard spot and / or the assignment to one of the energy classes is determined based on the height of a peak-to-peak amplitude of the relative measurement signals in the test area in one of the MFL data sets, in particular in the low-MFL data set. The peak-to-peak amplitude is considered from maximum to minimum. In extensive evaluations, characteristic signal curves of the relative signals were identified, which allow a corresponding assignment.

[0037] Alternatively, or particularly in addition, the hardness of the hard spot and / or its assignment to one of the energy classes is determined based on an internal peak-to-peak width in one of the MFL data sets, especially in the low-MFL data set. In equally extensive evaluations, characteristic signal curves of the relative measurement signal were identified, allowing for a corresponding assignment.

[0038] Alternatively, or again in addition, it has proven advantageous to assign the hard spot to energy classes or determine its hardness based on the average maximum EC amplitude (eddy current amplitude) of the hard spot present in the test area. In particular, the combination of the three indicators peak-to-peak amplitude, peak-to-peak width, and maximum EC amplitude results in high probabilities for assigning the spot to individual energy classes or hard spot types. The probability of detection (POD) for the test areas considered using the method described above is over 80%, which is relevant for pipeline operators.

[0039] Preferably, a hardness gradient is further determined across the wall thickness of the test object based on the signal width of an EC extremum, i.e., eddy current extremum, and the signal width of an MFL signal section. The signal width of an MFL signal section can correspond to the peak-to-peak width of the previously described MFL signal curve. The determination of the hardness gradient is further improved if a calibration for the electrical resistance of the test object is used. This can be done, for example, via an additional eddy current measurement in the region of the first magnetizing device, thus in the region of the magnetically saturated test object, with the measurement data being evaluated as described. This results in an additional correlation to the hardness of the hard spot, which improves the results overall.

[0040] The object stated at the outset is also achieved by an inspection device for determining hard spots of a test object, which is designed in particular as a pipeline pig with at least one and preferably at least two propulsion means, comprising at least two magnetizing devices spaced apart from one another in the longitudinal direction of the inspection device, each having at least one magnet designed in particular as a permanent magnet, and at least two MFL sensors spaced apart from one another in the longitudinal direction of the inspection device, each assigned to a magnetizing device, as well as at least one,an eddy current unit spaced longitudinally from the MFL sensors, comprising at least one eddy current sensor and a control unit for storing the MFL data recorded by the MFL sensors in an eddy current measurement in MFL data sets, as well as for controlling the eddy current unit to generate eddy currents and for storing the eddy current data recorded by the eddy current sensor in an EC data set.

[0041] In particular, the eddy current unit is arranged at a distance in the longitudinal direction from the first magnetizing device, which also has poles of different polarity that are spaced apart in the longitudinal direction and directed towards the test object, between which the MFL sensor is arranged. The eddy current unit can also be arranged offset in the longitudinal direction from the second magnetizing device, forwards or backwards with respect to a relative direction of travel of the inspection device. In particular, the eddy current unit is offset from the second magnetizing device towards one end of an inspection device designed as an inspection pig, preferably on a separate segment. Alternatively, the at least one eddy current sensor can be arranged between the poles of the second magnetizing device.In a design as a pipeline pig, the inspection device according to the invention preferably has, viewed in the longitudinal direction running in the direction of a longitudinal axis of the test object designed as a pipeline, a plurality of MFL sensors arranged next to one another in the circumferential direction and a plurality of eddy current sensors arranged next to one another in the circumferential direction.

[0042] By using the same inspection device for the three high-MFL, low-MFL, and EC measurements, the recorded data can be assigned particularly well and the determination of the hardness of individual hard spots is improved. Alternatively, the eddy current measurement can also be performed using a separate inspection device, particularly a separate pipeline pig.

[0043] In particular, part or all of the method described above or below can be carried out in the control unit of the inspection device, for which purpose this control unit then forms the computer device described above.

[0044] Alternatively, only part of the process, in particular concerning the storage of the measurement data in the control unit, can be implemented, while the evaluation of the hard spots is carried out in a separate computer device.

[0045] The object mentioned above is thus also achieved by an inspection device arrangement comprising an inspection device as described above or below, as well as an IT device configured to implement the method according to the invention as described above or below. It is understood that the IT device can be configured as described above or below.

[0046] Preferably, the critical hard spots found in the test object, in particular a pipeline, using the method according to the invention, especially those with a hardness above 320 HV, are compiled and output in a data set. Based on the data set or the hard spots found, to which a position in the test object can be or will be assigned using conventional positioning methods, a repair of the test object, in particular the pipeline, can subsequently be carried out according to a further development of the invention, in which the hard spot is processed, for example, removed.

[0047] Further advantages and details of the invention will become apparent from the following figures. Schematically shown:

[0048] Fig. 1 shows an object according to the invention in a pipeline,

[0049] Fig. 2 shows the relative permeability of different hot spots in a base material,

[0050] Fig. 3 shows an enlarged section of Fig. 2, Fig. 4 shows a high-MFL data set in a test area,

[0051] Fig. 5 a low-MFL data set in the test area,

[0052] Fig. 6 the EC data set in the test area,

[0053] Fig. 7 High and low MFL data sets in a further test area over different temperatures, Fig. 8 the object according to Fig. 7 with values ​​over the center of the hard spot found according to Fig. 7,

[0054] Fig. 8 Low-MFL data over a hard spot (top) and the eddy current signal (bottom)

[0055] Fig. 9 Signals of a hard spot with low magnetization and eddy current signal

[0056] Fig. 10 Data for a correlation of maximum hardness with peak-to-peak amplitude,

[0057] Fig. 11 Data for a correlation of the maximum hardness with the peak-to-peak width, Fig. 12 Data for a correlation of the maximum hardness with the EC amplitude,

[0058] Fig. 13 a schematic representation of a hard spot in a pipeline wall

[0059] Where appropriate, elements with the same function are provided with identical reference numerals. Individual features of the exemplary embodiments described below, in combination with the features of the independent claims and, where appropriate, further claims, can also lead to subject matter according to the invention.

[0060] An inspection device arrangement according to the invention according to Fig. 1 comprises an inspection device 4 designed to travel in a test object 2, in this case a pipeline, and an EDP device 6. The EDP device 6 can communicate with a control unit 10 of the inspection device 4 in a wired or wireless manner via a communication connection 8 shown in dashed lines in order to record the measurement data stored in the control unit 10 in the form of MFL data or data sets and in the form of EC data or EC data sets.

[0061] The inspection device 4 has a plurality of propulsion means 12 in the form of cups, which ensure the propulsion of the inspection device, also referred to as a pig, through the pipeline. A first segment 14 is used to record high-MFL data or data sets, a second segment 16 for low-MFL data or data sets, and a third segment 18 for recording EC data and generating corresponding data sets. The respective measurement data are combined into respective data sets, in particular in the control unit 10. The control unit 10 is arranged in a final segment 20 of the inspection device 4. The four segments are movably connected to one another via couplings 22 to enable the best possible pipeline accessibility and improved data alignment.

[0062] The first two segments 14 and 16 each have magnetizing devices 24 and 26, respectively, between which sensor heads 28 comprising MFL sensors are arranged. The MFL sensors are arranged in the sensor heads 28. Both segments 14 and 16 have a plurality of magnetizing devices 24 and 26, respectively, in the circumferential direction, as well as a plurality of MFL sensors in corresponding sensor heads 28, so that the inside of the pipeline is completely covered. Accordingly, segment 18 also has a plurality of eddy current units 30 comprising eddy current sensors in the circumferential direction, only one of which is shown as an example. The MFL sensors, in the form of Hall sensors, like the eddy current sensors, are arranged offset from one another in the longitudinal direction L of the inspection device. The longitudinal direction L runs parallel to or congruent with the direction of the relative movement of the inspection device to the test object.

[0063] In addition to recording the data for generating the corresponding data sets in the control unit 10, the control unit 10 also excites the coils of the eddy current units 30. After completing an inspection run through the pipeline, the data sets from the three measuring units are transferred from the control unit 10 to the computer device 6, where they are preferably evaluated automatically. It is understood that, in general, the data can be segmented in order to evaluate the data of the test object 2 not completely, but rather, for example, segment by segment for individual pipeline sections.

[0064] In general, for example, relevant data areas, i.e. those with a sufficiently large amplitude in at least one of the data sets, which correspond to individual areas of the pipeline in the longitudinal and circumferential directions, can be selected as a test area in the respective data sets by means of a selection algorithm, so that one or more test areas of the test object are defined accordingly, in which the hard spots are then determined according to the method according to the invention.

[0065] The computer device 6 runs programs or program modules through which the evaluation takes place, i.e., which implement the individual evaluation steps of the method. Furthermore, the calibration or correlation data required for the evaluation can be stored in the computer device or in a database linked or connectable to it.

[0066] Depending on the design of the control unit 10, it can alternatively or additionally use its data processing means to carry out at least a partial evaluation of the recorded data, so that after the inspection run has been completed, ready-made information can be obtained from the control unit.

[0067] Essential for the invention is the insight shown in Fig. 2 regarding the relative permeability p r . Shown are the curves of the permeability number or relative permeability p r for an exemplary base material (solid, hardness 175 HV), a pipeline steel with 0.23% carbon content, as well as for hard spots of different hardness as a function of magnetization with a magnetic field strength H:

[0068] - Hardness 400 HV (dashed),

[0069] - Hardness 361 HV (dotted) and

[0070] - Hardness 241 HV (dash-dotted).

[0071] The calculation of the hardness in HV can be determined, for example, using a test force of 10 kiloponds.

[0072] For the individual hardnesses of the hard spots relative to the curve of the base material, negative or positive values ​​result from the difference between the individual curves, which explains why the signal response of a hard spot in a magnetic field is sometimes associated with negative amplitudes, i.e. negative relative measurement signals of the magnetic leakage flux, and sometimes with positive amplitudes (=positive relative measurement signals) of the magnetic leakage flux. These deviations are illustrated by the arrows "negative amplitude" and "positive amplitude". The relative measurement signals are positive when the permeability of the hard spot is lower than that of the base material. Conversely, the relative measurement signals are negative when the permeability of the hard spot is higher than that of the base material. The orientation of the signal amplitude can then be determined for each possible magnetization and for different hardnesses.

[0073] The area marked with a rectangular, dashed box is shown enlarged in Fig. 3.

[0074] The hard spot with a hardness of 361 HV intersects the permeability curve of the base material at approximately 2.2 kA / m. This means that the hard spot of this tempering stage exhibits a positive relative measurement signal at low magnetizations, which decreases with increasing magnetization until it reaches 0 kA at approximately 2.2 kA / m. Then, as the magnetization increases, the signal changes sign and becomes negative, increasing with further magnetization. In the area of ​​the oval shown, the relative measurement signal or the relative signal amplitude is noise or zero, so the corresponding hard spots with a hardness of 361 HV cannot be found in the MFL signal.

[0075] To reliably detect these hard spots, an additional measurement signal is required. For this purpose, a material-sensitive eddy current signal is used, which is measured downstream of the first MFL measurement at remanent magnetization, i.e., at a magnetization of less than 1 kA / m. Exemplary data sets for a high-MFL data set can be found in Fig. 4, with test areas 33 located in the dashed and partially represented oval 32 (cf. Fig. 6). The relative measurement signal is shown in relation to the base material. The high-MFL data do not indicate a hard spot. Oval 32 is also shown in the corresponding low-MFL image according to Fig. 5, although the section viewed is somewhat larger than in Fig. 4. The x-axis represents the extension of the test object in its longitudinal direction, and the y-axis represents the extension of the test object in its circumferential direction.The test areas 33 are partial sections of the section of the test object being examined. In both areas, marked by ovals 32 in Figs. 4 and 5, no significant MFL signals are present. This is different in the eddy current or EC data set shown in Fig. 6. There, characteristic indications of hard spots 34 with correspondingly high amplitude values ​​are found, marked in the data set by black fields within oval 32. A hardness determination of the test area shows corresponding hard spots 34.

[0076] The signal responses of a hard spot generated at different tempering temperatures over a specific period (15 minutes) determined through laboratory measurements are shown in Fig. 7 compared to two low-MFL magnetizations of 2 kA / m and 5 kA / m as well as two high-MFL magnetizations of 10 kA / m and 20 kA / m for different tempering temperatures in a test area. The test setup was based on a measurement situation using a pipeline pig. It can be seen that the relative measurement signals or signal curves differ characteristically from a temperature of 0 °C up to a tempering temperature of 550 °C. This is even clearer in the figure after Fig. 8, which shows signals passing through the O position (y-axis), i.e. the center of the hard spot, in the respective test areas. The columns of the matrix in Fig. 8, with their characteristic signal curves, correspond to the respective energy classes.

[0077] Based on the characteristic signal curves identified by class in Fig. 8, three parameters can be derived that are used to correlate hardness. These are the peak-to-peak amplitude, indicated by a double arrow 36 (Fig. 9), the inner peak-to-peak width 38, and the EC amplitude 40, averaged over a plurality of measurements and otherwise the maximum EC amplitude (Fig. 9, lower diagram).

[0078] For the purposes of the invention, a large number of hard spots were generated and evaluated to map the corresponding correlations. Accordingly, the correlations of maximum hardness with the peak-to-peak amplitude, the inner peak-to-peak width, and the maximum eddy current amplitude shown in Figs. 10, 11, and 12 show an overall POD for the method that is in the range of well over 80% (87%).

[0079] Based on the characteristic signal curves (Fig. 8), energy classes can be derived, so that for the example in Fig. 8, with a total of six energy classes, the aforementioned correlations are implemented and the corresponding measurement signals can be used to detect the hard spots and determine their hardness. The hardness gradient across the thickness of a pipeline wall 42 can also be derived by comparing the signal width 45 of the eddy current or eddy current signal, measured on the inside I of the test object 4, in the area of ​​an eddy current extreme (EC extreme) with the signal width 44 of the low-MF L signal. Based on the observation of the respective widths, the direction of the extension of the hard spot can be defined, whereby the greater width of the MFL signal results in a hard spot that widens towards the outside A.

Claims

Claims 1. A method for determining hard spots (34) in metallic test objects (2), in particular in pipelines, tank walls, and metal sheets, in which two MFL data sets of a test area of ​​the test object (2) are generated by means of at least one inspection device (4) via two MFL measurements, wherein a high-MFL data set is generated by means of the first MFL measurement, which in particular leads to magnetic saturation of the test object (2), with a first magnetization of a first magnetizing device (24), and a low-MFL data set of the test area is generated via the second MFL measurement with a magnetization of a second magnetizing device (26) that is lower than the first magnetization, wherein an eddy current unit (30) is used to generate an EC data set of the test area by means of an eddy current measurement preceding or following the first MFL measurement, and the three data sets are subsequently evaluated with regard to their signal profile in the test area.in particular in an EDP device (6), wherein, in addition to the signal curves of at least one of the MFL data sets, the signal curve of the EC data set is used to evaluate the test area for hard spots (34).

2. Method according to claim 1, characterized in that the EC signal curve is used as a function of the measurement signal of the first MFL data set to localize the hard spot (34), in particular when a particularly relative signal amplitude of the first MFL data set in the test area remains below a high MFL threshold.

3. Method according to claim 1 or 2, characterized in that the EC signal curve is used as a function of the measurement signal of the second MFL data set for localizing the hard spot (34), in particular when a relative signal amplitude of the second MFL data set in the test area remains below a low MFL threshold value.

4. Method according to one of the preceding claims, characterized in that a signal curve of the EC data set above an EC threshold value is used to determine the hard spot (34).

5. Method according to one of the preceding claims 2 to 4, characterized in that a hard spot (34) is determined on the basis of at least one change in the sign of a relative measurement signal between the MFL data sets.

6. Method according to one of the preceding claims, characterized in that the signal curves of the relative measurement signal of the high-MFL data set in the test area, the relative measurement signal in the low-MFL data set and the EC data set are used to determine the hardness of the hard spot (34).

7. Method according to one of the preceding claims, characterized in that the hardness of the hard spot (34) is determined via an energy class to which one or more signal curves are assigned.

8. Method according to claim 7, characterized in that at least three energy classes are used to derive the hardness of different hard spots (34), in particular wherein a correlation of the hardness as a function of the MFL amplitude is used for each energy class.

9. The method according to claim 8, characterized in that the hardness of the hard spot (34) and / or the assignment to an energy class is determined on the basis of the height of a peak-to-peak amplitude (36) of the relative measurement signal in the test area in one of the MFL data sets, in particular in the low MFL data set.

10. Method according to one of the preceding claims 8 or 9, characterized in that the hardness of the hard spot (34) and / or the assignment to one of the energy classes is determined on the basis of a particularly inner peak-to-peak width (38) of the relative measurement signal in the test area in one of the MFL data sets, in particular in the low MFL data set.

11. Method according to one of the preceding claims, characterized in that the hardness of the hard spot (34) and / or the assignment to one of the energy classes is determined on the basis of a particularly mean maximum EC amplitude (40) of the hard spot (34) present in the test area.

12. Method according to one of the preceding claims, characterized in that a hardness gradient over a wall thickness of the test object is determined based on a signal width (45) of an EC extremum and a signal width (44) of an MFL signal section.

13. Method according to claim 12, characterized in that a calibration of the electrical resistance of the test object (2) is used in determining the hardness gradient.

14. Inspection device for determining hard spots (34) of a test object, which is designed in particular as a pipeline pig with at least one and preferably at least two propulsion means (12), comprising at least two magnetizing devices (24) spaced apart in the longitudinal direction (L) of the inspection device (4), each having at least one magnet, in particular designed as a permanent magnet, and at least two MFL sensors spaced apart from one another in the longitudinal direction (L), each associated with a magnetizing device (24), as well as at least one eddy current unit (30) spaced apart from the MFL sensors in the longitudinal direction (L), comprising at least one eddy current sensor,and a control unit (10) for storing the MFL data recorded by the MFL sensors in an eddy current measurement in MFL data sets and for controlling the eddy current unit (30) to generate eddy currents and for storing the eddy current data recorded by the eddy current sensor in an EC data set.

15. Inspection device according to claim 14, characterized in that the inspection device (4) with its control unit (10) is designed to carry out the method according to one of claims 1 to 13.

16. Inspection device arrangement comprising an inspection device (4) according to claim 14 or 15 and an EDP device (6) which is designed to carry out the method according to one of claims 1 to 13.