GB2071849A - Improvements in ultrasonic testing - Google Patents

Abstract

The sensitivities of measuring channels associated with test heads Pk1 to Pk3 are adjusted according to a basic sensitivity determined for the installation as a whole, after comparing the outputs of the measuring channels derived from test flaws 12 in a test piece 1. These outputs represent the respective maximum values of echo signals arising from the test flaws 12 during a calibration procedure in which test piece 1 is rotated about axis Z relative to the test heads. <IMAGE>

Description

SPECIFICATION Improvements in ultrasonic testing In the non-destructive testing of workpieces in the form of tubes or rods, it is known to use a testing installation including a plurality of test heads, each associated with a respective measuring channel. A workpiece is introduced into the installation, and has transmitted into it an ultrasonic signal, in the form of pulses. Echos are reflected from discontinuities, including the external surface of the workpiece and the internal surface if the workpiece is in the form of a tube, as well as from flaws.

The ultrasonic echos are picked up by the test heads, each of which incorporates a transducer which is used to convert the ultrasonic signal into an electrical echo signal. This signal is processed by the associated measuring channel and used to provide information concerning the workpiece.

In order to ensure accurate measurement, it is necessary to make adjustments to balance the measuring channels, ensuring that the sensitivity of all measuring channels corresponds, or otherwise to take into account differences in sensitivity when evaluating the results obtained from the different measuring channels.

A method of making such adjustments is proposed in Federal German specification DT-AS19 10 750. As proposed in this earlier specification, a stationary test piece in the form of a tube having predetermined test flaws is introduced into the testing installation and the measuring channels activated one after the other. The signal obtained from each measuring channel is compared with a reference signal. In the event that the signal derived from any of the measuring channels differs from the reference signal, a damping element, or attenuator, is adjusted to result in the relevant measuring channel exhibiting the desired sensitivity.

In order for this known method to be put into effect, it is necessary for the input signal to each measuring channel to be present at a constant magnitude for the duration of the step of balancing the measuring channel. The test piece must therefore first be rotated, or the test heads made to revolve about the test piece, until a suitable signal is established, and the balancing operation is then carried out. Particularly in the case of an installation for testing thick-walled tubes of large diameter, this operation is very cumbersome, time consuming, and leads to relatively inaccurate measurement results.

In order to overcome these problems, it is proposed herein that during the test, the test piece should be made to revolve about its axis relative to the test heads; or the test head made to revolve around the test piece, and that during each such relative revolution of the test piece and test head, the maximum value of an echo signal arising from a specific type of test fault, in or on the test piece, is measured for each measuring channel and stored. The stored signals are subsequently used to establish a basic sensitivity value for the installation, and correction factors are determined for all the channels from the differences between the said basic sensitivity value and the respective echo signals.

The drawing shows a diagramatic representation of an ultrasonic test installation, to illustrate the performance of the proposed method of adjusting the sensitivity of the installation.

Referring to the drawing, the ultrasonic test installation depicted includes three ultrasonic test heads Pk, to Pk3, all of which are intended to operate in accordance with the pulse/echo technique. Each test head includes a transducer, to which electrical pulses are supplied in order to cause ultrasonic pulses to be propagated through the workpiece. Ultrasonic echos arising from discontinuities in the workpiece are picked up by each test head and converted into an electrical echo signal which is evaluated to provide information concerning the discontinuities, this information being displayed by an indicator 11.

Each test head is associated with a measuring channel, that of test head Pk, alone being shown for reasons of clarity. The remaining two test heads are associated with respective identical measuring channels. Before the testing installation may be used to test production workpieces, it is necessary to balance the measuring channels, in order to compensate differences in sensitivity. For this purpose, use is made of a test piece 1 having test flaws 1 2.

In operation, the test piece is rotated about its axis, and electrical pulses are supplied to the test head Pk, under discussion by a transmitter 2 at a frequency IFF of, for example, 360 pulses per revolution. As an alternative to rotating the test heads, the installation may rotate about the axis. The ultrasonic echo pulses received by the test head Pk, from the external and internal surfaces of the tube and flaws 1 2 are converted into an electrical echo signal which is supplied to an amplifier 3, the amplified output signal from which being input to an analogue/digital (A/D) converter 4.

The A/D converter 4 converts to digital form those analogue echo signals within a flaw expectation range determined by a discriminator window circuit 83. Since, in the testing of tubes, it is to be expected that flaws occur primarily in the vicinity of the inner and outer tube surfaces, the circuit 83 is adapted to cause the A/D converter 4 to discriminate against signals produced by echos arising elsewhere within the test piece and to convert to digital form only the echo signals of these test flaws. Operation of the system takes place under the control of release signals S, to S5 and S7, to be described later.

The digital signals pertaining to flaws in the internal surface of the test piece are stored in a memory 6, while the digital signals pertaining to flaws in the external surface are stored in a memory 7. Each of the memories 6 and 7 is a maximum value memory, in the sense that the memory stores only the maximum signal occurring in each revolution of the test piece relative to the test installation. After one relative revolution of the test piece and test installation, the next test head Pk2 is brought into operation and the contents of the memories 6 and 7 relating to the signals derived from the test head Pk, are transferred into a principal memory 81 of an evaluation unit 8.

Thereafter, the information stored pertaining to the results obtained from test head Pk2 are transferred to the principal memory 81, and test head Pk brought into operation.

After the digital values for the internal and external test flaws as determined by all three measuring channels have been transferred into the principal memory 81, a basic sensitivity factor for the installation as a whole is determined by means of a computer 82. If necessary the sensitivity of one or more of the measuring channels is automatically adjusted in relation to this basic sensitivity by changing the gain of the amplifier 3 using for this purpose a signal generated by computer 82.

The basic sensitivity of the installation may be determined and expressed in accordance with various criteria. One such criterion is to determine the arithmetic mean value of the stored signals. Thus, considering say internal flaws, and assuming that the memory 81 contains the stored signals which can be expressed as having magnitudes of 20, 24 and 22 for the three measuring channels, respectively, the basic sensitivity is taken to be 22.

On this basis, the gain of the amplifier of the third measuring channel is left unchanged, whereas that of the first channel is raised according to the difference (22-20), and that of the second channel reduced according to the difference (24-22).

Alternatively, the basic sensitivity may be taken to be the minimum stored value. On this basis, in the example given above, the basic sensitivity would be taken to be 20, with the result that the gain of the amplifier of the first measuring channel is left unchanged, whereas that of the second measuring channel is reduced according to the difference (24-20), and that of the third channel according to the difference (22-20). Other concepts for determining a basic sensitivity may be employed.

After the basic sensitivity has been determined and the sensitivity of the channels whose sensitivity differs therefrom has been adjusted, the test piece having standard test faults is removed and the installation used to evaluate unknown flaws in production workpieces passing through the installation in direction Z.

Instead of adjusting the gain of the amplifier in order to vary the sensitivity of the measuring channel, it is possible for each measuring channel to include a damping element or attenuator as described in the abovementioned specificatiori, and for the attenuating factor of this element to be adjusted by the computer.

As a further alternative, instead of the computer generating an adjustment signal or correcting the gain of the amplifier 3 or varying an attenuation factor, the arrangement may be such that the sensitivities of the individual measuring channels are left unchanged, and the memory 81 or another memory is used to store a correction factor. In the subsequent testing of tubes with unknown external and internal faults, the correction factor pertaining to the or each measuring channel whose sensitivity differs from the basic sensitivity is added or applied to the measured echo signal value, before all the signals are evaluated, the computation being carried out by means of the computer 82 for example.

The discrimination window may be determined by means of a transit time measuring device 5, to which signals S3 and S4 indicating transit times T, and Te for internal and external faults respectively are determined before the sensitivity balancing operation described above. Each transit time constitutes the centre of a respective window. In the circuit 83, each transit time is modified by a constant which is characteristic of the desired width of the window i.e., the time intervals T, + C T + C are formed. This constant C is, among other things, a function of the wall thickness d of the tube and may be chosen for example equal to d/2. The values for the window are stored in the circuit 83 and during the sensitivity balancing are applied at the correct time to the A/D converter 4.

In order to ensure that the time and cycle of operation of the method takes place correctly, a unit 9 for cyclical control is provided, which in turn is connected with a pulse generator 10. The unit 9 is arranged to generate at the appropriate time the control signals S1 to S5.

These signals are as follows: S1--Release signal for internal flaw amplitude measurement.

S2-Release signal for external flaw amplitude measurement.

S3-Release signal for internal flaw transit time measurement.

S4-Release signal for external flaw transit time measurement.

S5-Release signal for transmitter 2.

S7-Release signal for evaluation unit 8.

It is advantageous if the test sensitivity of each of the individual measuring channels, i.e. the minimum sensitivity which will allow a flaw signal to be evaluated as such, is estab lished as a function of the basic sensitivity of the installation. For this purpose, the threshold value defining the test sensitivity is adjusted in accordance with changes in the basic sensitivity.

Preferably, the sensitivity balancing is carried out with the test piece 1 displaced automatically through the installation, in the longitudinal direction Z, especially if only one test flaw 1 2 is provided in the longitudinal direction, because hand displacement of the test piece longitudinally can be time-consuming.

For such automatic operation, the test piece is provided with signal points 13, as indicated diagramatically in Fig. 1, and which may be optical or magnetic markings scanned by a sensor 14. This sensor may also serve to transmit pulses indicative of rotation of the test piece, the corresponding signal points around the periphery of the tube not having been shown. Displacement of the test piece can, however, be dispensed with, provided that the test piece has a plurality of identical test flaws 1 2 spaced apart from one another in the longitudinal direction of the test piece at distances corresponding to the spacing of the test heads, the test piece then remaining stationary during sensitivity balancing.

The accurancy of the sensitivity balancing operation can be increased, if the operation of determining the basic sensitivity for the installation is repeated a number of times in succession, the corresponding values are stored, and a mean basic sensitivity calculated and used as the basis for balancing the channels.

By comparing the basic sensitivity values determined by such repeated operations, it is also possible to establish whether random flaws, for example water bubbles between the test head and test piece, are falsifying the basic sensitivity to be determined. Such falsification will be suspected, if, for example, the difference between two successively determined values of the basic sensitivity is larger than a predetermined limiting value.

Claims (11)

1. A method of automatically adjusting the sensitivity of an ultrasonic testing installation for the non-destructive testing of tubes or rods, the installation having a plurality of identical measuring channels each associated with a respective test head, wherein: (a) A test piece is made to revolve about its axis relative to the test heads; or the test heads made to revolve around the test piece, (b) during each such relative revolution of the test piece and test heads, the maximum value of an echo signal arising from a specific type of test fault in or on the test piece is measured for each measuring channel and stored, (c) the stored signals are used to establish a basic sensitivity value for the installation, and (d) a correction factor is determined for the or each channel from the difference between the said basic sensitivity value and the echo signal for such channel.

2. A method according to claim 1, wherein the basic sensitivity of the installation is determined from the arithmetic mean of the stored echo signals.

3. A method according to claim 1, wherein the basic sensitivity of the installation is determined from the minimum stored echo signal.

4. A method according to claim 1 or claim 2, wherein the correction factor is added to echo signal value during subsequent testing of production workpieces to compensate for differences in sensitivity.

5. A method according to any preceding claim, wherein the sensitivity balancing operation is carried out with the test piece moving longitudinally through the installation.

6. A method according to any preceding claim, including making repeated determinations of basic sensitivity, and adopting a mean value therefor.

7. A method according to claim 6, including comparing the values of basic sensitivity obtained by repeated determinations thereby to recognise random errors.

8. A method according to any preceding claim, wherein the test piece has a plurality of longitudinally spaced apart test flaws, at distances corresponding to the spacing of the test heads.

9. A method according to any preceding claim, wherein signal points are mounted on the test piece for automatic position recognition purposes.

10. A method substantially as hereinbefore described with reference to the drawing.

11. A test installation adapted to be adjusted in accordance with claim 1.