EP3596455A1

EP3596455A1 - Apparatus and method for ultrasonic testing

Info

Publication number: EP3596455A1
Application number: EP18724479.3A
Authority: EP
Inventors: Johannes Vrana; Matthias Goldammer; Hubert Mooshofer
Original assignee: Siemens AG
Current assignee: Siemens AG
Priority date: 2017-04-28
Filing date: 2018-04-25
Publication date: 2020-01-22
Also published as: CN110678748A; WO2018197529A1; US20210116421A1; DE102017207269A1

Abstract

The invention relates to a method and an apparatus for ultrasonic testing by means of a selection of probes, wherein a computer device is used for determining minimum required waiting times between two consecutive pulses for all possible pulse sequences (S1) and then determining an optimized pulse sequence (S2) of the shortest possible test cycle for the probes.

Description

description

Apparatus and method for ultrasonic testing In the ultrasonic examination is placed a test head on one side of the component, in particular in the pre ^¬ the side, and a short pulse insonified. This pulse is reflected by discontinuities or errors and by the back wall representing the back. Reflected pulses travel back to the probe after reflection, which is used as a receiver after sending the short pulse, and can therefore be made visible. However, the reflected signals are reflected back into the component just as they impinge on the component surface and thus pass through the component a second, third and so on. For each ping-pong, the probe records a signal again. This Sig ^¬ nal is always more weakened depending on the material until it goes down in the noise after a few ping pong cycles. During the ultrasonic inspection, components are scanned. That is, it carried out many such pulses can also be as shots ^¬ records, one after the other. After sending another pulse you can see in the received time signal, in addition to the actual signal, as well as not enough attenuated, in particular multiply reflected, late incoming repeat echoes of one of the earlier pulses. This would then lead to bogus announcements or phantom echoes that would be misinterpreted as real mistakes. Therefore you have to wait between two pulses long enough for the repeat echoes to be sufficiently decayed. From this War ^¬ tezeit the pulse repetition rate is obtained. Since the repeat echoes decay at different speeds with complex specimen geometry, the pulse repetition rate must be set to the latest echoes.

In the automated test, several test heads are sometimes used in parallel or one test head several times, for example with different reinforcements for different depths. che and the same, used. If multiple probes are ver ^¬ turns, one speaks of a plurality of real channels, using the test instrument applies a test head at the multiple ver ^¬ is of multiple virtual channels. However, any real or virtual channel can create bogus displays in any other channel. Phased Array (PA) probes have multiple resonators arranged in an array that may be one-dimensional or even two-dimensional. By delayed pulses and receiving the individual elements can be the

Electronically controlling the insonification angle, the focal point of the

Focusing the beam electronically to a certain depth, moving the cone of sound linearly, etc. Each of these delay settings is referred to as the "Focal Law." Often, however, the phased array test does not trigger a single angle, but the sound beam is panned but a pan that the delay for a particular angle is set, the probe gefeu ^¬ ert, waiting for the answer and then the delay for the next angle is adjusted, etc. thus, the probe must pulses N times, by an angle swivel with N different Similarly, as with automated testing, the various real or virtual channels, each pulse of the probe may also cause a dummy indication in another pulse. Array probes can also be used in automated testing In this respect, the pulse repetition rate can also be influenced by the aspects mentioned above.

Full-matrix capture (FMC) or Total Focussing Method (TFM) uses a phased array probe. In these methods, one is usually pulsed with one element and received with all the elements, then pulsed with the next element and received again with all, etc. The data thus obtained is then billed to an event image. But also here every pulse of the test head can have a fake cause a different pulse, which can ultimately negatively affect the calculated result image.

In the Synthetic Aperature Focussing Technique (SAFT), the data from several probe heads positions, which may be provided in a classical or phased array fashion, and possibly from multiple real or virtual channels, are billed together. Again, the above conditions for avoiding bills of late repeating echoes should be considered.

The setting of a suitable waiting time from one pulse to the next, ie the setting of the pulse repetition frequency, must be made before the test. This is currently done manually by the examiner. This is quite simple for a one-channel test. The examiner may, and starting from a very large value, the waiting time as much ver ^¬ cut that barely any bill on the A-image will appear.

With multiple real and / or virtual channels, with multiple focal laws or when using FMC / TFM, manual adjustment becomes an extremely time-consuming procedure. If the waiting times are set too long, but this affects the test time. Therefore it must be tried to optimize the waiting times.

It often happens during the measurement that the sound attenuation changes or repeating echoes appear at different times or different intensities as a result of the component geometry, so that at some points in the result image, ^dummy indications can again be found. Then the settings for the waiting times have to be adjusted and the measurement can be restarted.

In addition to the optimization of waiting times is in part also a swap of real and / or virtual channels or Focal Laws makes sense to wait further reduzie ^¬ ren. With a manual setting, but this leads to an even more time-consuming adjustment procedure. It is an object of the present invention to automatically determine a shortest possible test cycle in a combination of different measuring methods. For example, classical probes can be combined with PA probes and / or FMCA PA probes.

The object is achieved by a method according to the Hauptan ^¬ demanding and by means of a device according to the independent claim. According to a first aspect, a method for ultrasound ^¬ testing by means of a selection of probes, proposed by means of a computer means at least required respective waiting times between two consecutive pulses for all possible shot sequences (Sl) and subsequently an optimized shot order (S2) one of the shortest possible test cycle of the probes is determined.

According to a second aspect, an apparatus for ultrasonic testing by means of one of the preceding methods, with a computer device for calculating at least necessary waiting times for all possible firing sequences and subsequently optimized firing sequences for a combination of at least one probe, at least one phased array probe and / or at least an FMC PA test head.

It is proposed to first determine the minimum required waiting times T _w k for each possible shooting sequence Pi with 1 = 1... N and then to determine an optimum shooting sequence.

Further advantageous embodiments are claimed by the subclaims: According to an advantageous embodiment, the combinations of N pulses Pi with N reception settings EEi with i = 1... N can be detected by means of the computer device.

According to a further advantageous embodiment can

for a N × N combination matrix of pulses Pi and reception settings EEi with i = 1... N, a time signal is recorded over a long period, which contains all subsequent echoes with a relevant amplitude.

According to a further advantageous embodiment can

a default for the maximum allowable amplitude of phantom ^¬ echoes defined and made as a receive setting EEi.

According to a further advantageous embodiment, the wait times after the pulses Pi and the minimum cycle duration can be derived from the matrix of N × N time signals and the amplitude specification for possible permutations of the pulses.

According to a further advantageous embodiment, the optimized or optimal pulse sequence can be selected.

According to a further advantageous embodiment, an automatic determination of the length of the recording period is carried out, wherein a decaying exponential function is determined, which represents an envelope of the time signal and it is checked whether the envelope ^falls below a certain value at the end of the recording ^¬ period.

According to a further advantageous embodiment, the waiting times determined can programming directly to the product by the pulses Pi a tester or a test system USAGE ^¬ be det. According to a further advantageous embodiment, discrete optimization techniques can be used instead of the complete calculation for all channel permutations. According to a further advantageous embodiment, a Monte Carlo approach can be combined with the completely permutative approach.

According to a further advantageous embodiment, the time signals for each of the N x N combinations of pulse and

Reception parameters are measured at several positions and then the maximum of the time signals is determined over all positions. According to a further advantageous embodiment, an automatic re-evaluation of the shortest pulse sequence can take place in parallel with a test at regular intervals.

According to a further advantageous embodiment, instead of determining all the time signals for each of the N × N combinations of pulse and received parameters, only a part of the signals must be determined by means of measurement, the other part being able to be replaced by prior knowledge. According to a further advantageous refinement, a plurality of reception settings can be approximately represented by means of a single reception setting for an FMC test. The invention will be described in more detail by means of exemplary embodiments in conjunction with the figures. Show it:

1 shows a first embodiment of a representation of a pulse with subsequent repeat echoes;

FIG. 2 shows an exemplary embodiment of a combination of probes to be optimized; Figure 3 is a representation of the procedure for determining the optimum combination of probes;

FIG. 4 shows a representation of receiver settings EEi;

5 shows a first view of a second execution ^¬ example of a pulse with its Wiederholechos;

FIG. 6 shows a second illustration of the second exemplary embodiment of a pulse with its repeat echoes;

7 shows a third representation of the second game Ausführungsbei ^¬ a pulse with its Wiederholechos; Figure 8 is a fourth illustration of the second game Ausführungsbei ^¬ a pulse with its Wiederholechos;

Figure 9 shows an embodiment of an inventive

Procedure.

1 shows a first embodiment of a depicting lung ^¬ a pulse with subsequent Wiederholechos.

FIG. 2 shows an embodiment of a combination of probes to be optimized.

In a particularly automated test, two classic probes, one PA probe and one FMC PA probe are used. The two classical probes 1 and 2 are connected to the real channel 1 and 2, the PA

Test head on channel 3 and the FMC PA test head on channel 4.

Test head 1 is pulsed with two different settings, namely by means of a virtual channel 1 and a virtual ^¬ len channel 2. Prüfköpf 2 is pulsed with three different settings, by means of the virtual channels 1,2 and 3, the PA probe with three different focal laws or delay settings, for example, using three different angles and the FMC PA probe has four Elements, where each element is individually pulsed and then received with all four elements. Thus, in this example, one cycle is fired twelve pulses. For this situation, it is important to automatically optimize the waiting times and the order. For this, the interaction of the N pulses on the N receive settings must be determined.

FIG. 3 shows a representation of the procedure for determining the optimum combination of probes.

For this purpose, the pulse PI can be started as a first step S1 and with all N, according to Example 12, different reception settings EEi i = 1... N, ie all virtual channels corresponding to the classical probes, all delay settings, and Although the PA probes, and all elements FMC PA are received and recorded. However, since it is only possible to receive on a virtual channel or to receive or record with a delay setting per classic probe or PA probe, multiple pulses must be pulsed for a complete evaluation of the pulse in order to test all virtual channels in succession. In our example, must be at least 3x pulsed ^¬ to, namely black, red and blue in Figure 2 at pulse first

Depending on the recipient's setting, the evaluation of some recipient settings may be omitted, for example if two recipient settings match. However, this requires prior knowledge of the receiver settings.

If FMC depending on the transmitting element has a differing ^¬ che adjustment of receiving elements are used, the receiver settings EEi must also be successively tested here.

FIG. 3 indicates that this process is then repeated for each pulse P2 to PN, in this example N = 12. gets picked. This provides complete information about the interaction of all N pulses with all N receive settings. For each of the N x N combinations there is a time signal that shows the interaction.

Each receiver setting EEi is a certain gain, which may in particular be time-dependent, and associated with one or more time slots in which data is recorded. These time windows each have a start corresponding to the time after the transmitting pulse, a length in which a pulse-like miscalculation or error can be found. In addition, signals are aussa ^¬ gekräftig only above a certain signal level, as the signals otherwise be lost in the noise. Therefore, a signal level must always be defined, from which signals must be evaluated. The signal level together ^¬ men with the time windows or time windows resulting in one or more temporally constant or variable "aperture" per recipient setting. Within that "aperture" may no other pulse start.

4 shows two such "apertures." For the other examples, the sloping "aperture" is used for receiver setting EEI, the rising aperture for receiver ^setting EE2. The aperture indicates the just allowable height of the disturbing repeat echoes and underlying echoes can be accepted. Figure 5 shows the time course of a pulse Pi, the example ^{¬ has} been recorded with the receiver setting EE2.

The time window marked in FIG. 6 by means of the straight line to ti represents the aperture of the receiver setting EEI and not that of the receiver setting EE2. Within this time window from to to ti no further pulse may be allowed. will be started. Another pulse can be started after the time window, after ti.

As already described above, each of the N receiver settings can be assigned an "aperture" or a time range to _k to ti _k . Therefore, it is now necessary to evaluate in which areas a respective receiver setting is suitable at the earliest. In this case, the range should be long enough for the receiver adjustment time slot to fit in and for admissible, in particular, time-dependent signal levels.The sooner the next pulse can be started, the shorter the entire pulse sequence will be.

FIG. 7 shows as an example that a receiver setting EE2 or "aperture" EE2 does not fit in a first gap, but in a subsequent second one

In this way, a time can be determined for each of the N × N combinations that must be maintained between a pulse Pi and a pulse Pi + i. Figures 7 and 8 indicate the N x N

Combinations with a first Empfängerereinstellungskombina ^¬ tion of apertures EE1, EE2 and EE2 in Figure 7 and a second receiver setting combination of apertures EE1 and EE2 in Figure 8 represents.

Thus, for a given sequence of channels for each channel, the subsequent channel is timed to obtain the shortest possible sequence. This procedure can be performed for each possible sequence of individual pulses Pi, where no new measurement is needed, but the recorded echo sequences are considered le ^¬ diglich. This allows a complete calculation of the total time of all permutations. Since the first one is measured again directly on the last channel, this pairing must also be considered. Upon completion of the calculation will result in a

List (Nl)! different total cycle times which can now be sorted in ascending order. This is shown in Table 1: Table 1:

Pulse sequence

10-8-4-3-1-2- 9-12-11-5-7- 6-10 -... 4,39

10-8-4-3-1-2- 9-12-11- 6-7-5-10 -... 4,86

10-4-3-1- 9-2-8-12-11- 6-7-5-10 -... 5.49

It is also examined whether the influence of the penultimate, before ^¬ penultimate pulse etc. could lead to inadmissible late repeat echoes.

For this you can first consider the complete sequence as a whole. In the best case, there are no annoying repeat echoes in any of the channels. The pulse sequence can be used in this way, whereby the entire test time can be minimized. The algorithm is so abgeschlos ^¬ sen.

Should late repeat echoes be seen in one receiver setting or multiple receiver settings, it should first be identified which predecessor pulse the

Problem has caused. Subsequently, a waiting period should be extended correspondingly between the two pulses.

If, for example, in the pulse sequence 10-8-4-3-1-2- 9-12-11-5-7-6-10 -... a late repeat echo in pulse 10 fin ^¬ det, that caused by pulse 5 can, between the

Pulse 5 and 7, 7 and 6 and / or 6 and 10 additional waiting times are added to match the gaps. Then it should be checked if this was sufficient.

After the waiting times have been changed appropriately and all unwanted repeat echoes have been eliminated, the result is a new, slightly longer total cycle time. When comparing this total ^¬ cycle time with the total cycle times by other pulse sequences may be that the cycle is longer yields, compared to other cycles. In this case, the longer cycle of the first pulse sequence can be accepted as suffi ^¬ accordingly short in the illustrated below Table 2 below. Alternatively, further optimization can be performed for example by means test the second pulse sequence in Table 2 with the vorste ^¬ Henden described method.

Table 2:

Pulse sequence

10-8-4-3-1-2- 9-12-11-5-7- 6-10 -... adjusted 4,89

10-8-4-3-1-2- 9-12-11- 6-7-5-10 -... 4,86

10-4-3-1- 9-2-8-12-11- 6-7-5-10 -... 5.49

In summary, the invention is characterized by the following core features: For a N × N combination matrix of pulses Pi and reception settings EEi with i = 1... N, a time signal is recorded over a long period, which contains all subsequent echoes with a relevant amplitude. It is defined a specification for the maximum amplitude of phantom echoes and taken as a "diaphragm" or or as Emp ^¬ fang setting EEi.

From the matrix of N × N time signals and the amplitude presets, the waiting times after the pulses and the minimum cycle duration are derived for possible permutations of the pulses.

Based on this, the optimized or optimal pulse sequence is selected.

The following variations can occur: Automatic determination of the length of the recording period, which can result in a repetition with a longer recording period. This can be done, for example, by determining a decaying exponential function that represents an envelope of the time signal and is tested. It can be checked, for example, whether the envelope at the end of the recording period unterschrei ^¬ tet a certain value, for example, whether the smallest amplitude setting for phantom echo is not too large.

The determined waiting times after the pulses Pi are used directly for programming a tester or a test system. For large numbers of channels, known discrete optimization techniques can be used instead of the full calculation for all channel permutations.

For very complex systems, the exhaustive search for the shortest latency can be very computationally expensive, hence the number of permutations increases with the faculty of the channel number. In this case, for example, a Monte Carlo approach can be combined with the completely permutative approach.

This can be done as follows:

Instead of completely computing all permutations, a subset of the channels is selected at random and these are completely permuted and optimized on their own. Thereafter, the same procedure is used with the remaining channels, in order to then line up all the channels. This significantly reduces the computation time so that a number of sub-options can be used. Instead of a subdivision into two subsets, a smaller division into three or more subsets is also possible. The total test duration is no longer optimal in this approach, but can be approximated to an optimal test duration. Devices under test with a location-dependent varying material properties or when the geometry of the test object along the scan path is changed, this can be taken into account who ^¬ that the time signals for each of the N x N combinations of the pulse and reception parameters at a plurality of positions are measured, and then the Maximum of the time signals over all positions is determined, and thus the inventive method can be carried out as described. It may also be parallel to an exam, especially at

Test specimens with location-dependent fluctuating material properties, at regular intervals an automatic re-evaluation of the shortest pulse sequence. To determine, instead of all the timing signals for each of the N x N combinations of the pulse and reception parameters, only a part of the signals can also be determined by measuring ^¬ to, the other part can be replaced by prior knowledge or by wei ^¬ more excellent suitable assumptions.

For an FMC test, the multiple receive settings can be approximately represented by means of a single receive ^¬ setting. A possible procedure for finding a disturbing predecessor pulse or predecessor pulse may be the following:

If, for example, a late repeat echo can be found in the receiver position 10 in the pulse sequence 10-8-4-3-1-2-9-12- 11-5-7-6-10-000, the chain can be shortened stepwise or extended. This leads to a longer direct result. It is well known that the signal of late repeat echoes is getting weaker and weaker. One that is tried first, the chain 7-6-10, then the chain 5-7-6-10, the chain 11-5-7-6-10 and determines which of the pulses causes the prob lem ^¬. Another possible procedure for checking whether the adaptation of the pulse sequence was sufficient can be a testing of the sub-chains and then of the complete inspection chain. A test of the sub-chains can be carried out such that the sub-chain length is gradually increased, as otherwise ^¬ if the pulse has to be moved further.

Instead of the shortest pulse sequence of Table 1 as a slightly longer pulse sequence can be selected if thereby the residual signals remain below the associated diaphragm and the signal-to-noise ratio is ver ^¬ enlarges in this way.

The inventive step is seen in the following:

In contrast to the prior art, the pulse repetition rate and order of the channels is determined by machine. In the case of an exhaustive search is an optimal short test time guaranteed, while in manual setting a huge effort and a lot of experience are necessary to arrive at vergleichba ^¬ ren results.

The present invention has the following advantages: The test duration can be effectively minimized. The testing costs can be effectively reduced. Optimal use can be made of the test equipment and the test personnel. It can be avoided erroneous checks that need to be corrected because of phantom echoes.

It can also be advantageous in test specimens with location-dependent fluctuating material properties carried out a Prüfzeitoptimierung, since several digits can be considered. FIG. 9 shows an exemplary embodiment of a method according to the invention. By means of a computer device at least required respective waiting times between two aufei ^¬ subsequent pulses for all possible shot sequences in a first step Sl and subsequently determined in a second step S2 an optimized firing order of a shortest possible test cycle of the probes.

Claims

claims

1. Method for ultrasonic testing by means of a selection of probes,

characterized in that

by means of a computer device at least required each ^¬ waective waiting times between two successive pulses for all possible shot sequences (Sl) and subsequently ei ^¬ ne optimized firing order (S2) of a shortest possible test cycle of the probes is determined.

2. Method according to claim 1,

characterized in that

the combinations of N pulses Pi with N reception settings EEi with i = 1... N are detected by means of the computer device.

3. The method according to claim 1 or 2,

characterized in that

for a N × N combination matrix of pulses Pi and reception settings EEi with i = 1... N, a time signal is recorded over a long period of time, which contains all subsequent echoes with a relevant amplitude.

4. The method according to claim 3,

characterized in that

Defines a default for the maximum allowable amplitude of phantom ^¬ echoes and is made as a receive setting EEi.

5. The method according to claim 4,

characterized in that

from the matrix of N x N time signals and the amplitude specification for possible permutations of the pulses, respectively, the waiting times are derived after the pulses Pi and the minimum ^cycle duration.

6. The method according to claim 5,

characterized in that

the optimized or optimal pulse sequence is selected.

7. The method according to claim 3,

characterized in that

an automatic determination of the length of the Aufnahmezeitrau- mes is carried out with a decaying Exponentialfunk ^¬ tion is determined, which represents ^¬ represents an envelope of the time signal and it is checked whether the envelope is below a certain value at the end of the up ^¬ acquisition period.

8. The method according to any one of the preceding claims, characterized in that

, the determined waiting time after the pulses Pi directly ver for programming a tester or a test system ^¬ turns.

9. The method according to any one of the preceding claims, characterized in that

discrete optimization techniques are used instead of the full calculation for all channel permutations.

10. The method according to any one of the preceding claims, characterized in that

a Monte Carlo approach is combined with the completely permutative approach.

11. The method according to any one of the preceding claims, characterized in that

the time signals for each of the N x N combinations of pulse and receive parameters are measured at a plurality of positions and then the maximum of the time signals is determined over all positions.

12. The method according to any one of the preceding claims, characterized in that

an automatic re-evaluation of the shortest pulse sequence takes place in parallel with a test at regular intervals.

Method according to one of the preceding claims, characterized in that

instead of determining all the time signals for each of the N x N combinations of pulse and receive parameters, only a part of the signals is determined by measurement, the other part being replaced by prior knowledge.

14. The method according to any one of the preceding claims, characterized in that

for an FMC test, several receive settings are approximately represented by means of a single receive setting.

15. Apparatus for ultrasonic testing by means of one of the prior methods, with

a computer device for calculating at least neces sary waiting times for all possible shot sequences and subsequently optimized shot sequences for a combination of at least one probe, at least one phased array probe and / or at least one FMC PA probe.