GB2026163A - Method for automatically obtaining test results in the non-destructive testing of materials using ultrasonic pulses - Google Patents

Abstract

A method for automatically testing materials using ultrasonic pulses compares the variation between two or more successive digitalised A images, obtained by varying the position of a test head arrangement with reference variations for known defect types. A test head arrangement (10) comprising several different probes (Figure 1, not shown) is guided over the surface of a test piece (1) by a manipulator (11) controlled by a computer (13) by way of a control device (12) in accordance with a set test programme. The pulse for the computer and the test head arrangement is produced by the ultrasonic device (14). The images produced are stored and variations in echo amplitude are compared separately in the X and Y directions in the computer, with reference variation values associated with given classes of defect. The method is applicable to monitoring seam welds. <IMAGE>

Description

SPECFICATION Method for automatically obtaining test results in the non-destructive testing of materials using ultrasonic pulses This invention relates to a method for automatically obtaining a test result in the non-destructive testing of materials using ultrasonic pulses, by evaluating two or more successive digitalised A images, obtained by varying the position of the test head arrangement which consists preferably of several test heads of different radiation characteristics.

It is known to feed the sound pulse from the test transmission head into the test piece at a given angle to the normal to the surface. The sound pulse is reflected at a reflector and then converted into an electrical signal in the test receiver head, this reflector being indicated as a defect only if it has been classified as "harmful". The test transmission and receiver heads can also be in the form of a single unit, in which the same electroacoustic transducer is used for sending and receiving the ultrasonic pulses. The distance between the test head and reflector is determined from the propagation time of the ultrasonic pulse, using the known velocity of sound for the material of the test piece. The echo height is under given conditions a measure of the size of the reflector.

If the echo height is measured as a function of the propagation time and is reproduced on the time vector of a cathode ray oscilloscope, in which the propagation time is proportional to the reflector distance as determined by the velocity of sound, then this is known as a A image. (See Krautkramer, Material Testing using Ultrasonics, 3rd edition, page 253). This term is also used if the variation of echo height with propagation time is fed in digitalised form to an evaluation apparatus, e.g. a computer, instead of being reproduced on a cathode ray oscillograph.

In order to subject an unknown reflector to sound in an optimum manner, it is known to feed the sound into the test piece at various input angles, in which case it can be advantageous to use separate test transmission and receiving heads. If it is preferred not to determine the depth of the reflector from the ray geometry and propagation time, several test receiving heads can be disposed one behind the other, so that for a fixed distance apart of the test heads within the test head arrangement, each test head indicates reflectors in a given depth zone. This method is known as the tandem method, see J. and H. Krautkramer, Material Testing with Ultrasonics, Springer-Verlag, 3rd edition 1975, pages 473 and 527.

Likewise, it is possible to obtain useful information regarding the type of defect, e.g. whether in the form of slag streaks, blind defects, pores etc., by the test superintendent continuously observing the cathode ray oscillograph and displacing the test head arrangement over the test piece both in the direction towards the reflector and also parallel to the reflector. The variation in the A image due to the movement of the test head arrangement is known as the echo dynamic range.

If, e.g. when testing a weld seam, the echo indications fluctuate between a maximum value and a threshold value, e.g. the zero line, as the test head arrangement is displaced parallel to the weld seam, then this in an indication of separated reflectors. In this case, the fluctuations in the amplitude as the test head arrangement is displaced parallel to the welding seam provide the test superintendent with information regarding the interruptions in the reflector. In addition, the test superintendent can obtain information regarding the geometrical shape of the reflector from the envelope curve of the echo dynamic range, which is obtained as the test head arrangement is displaced.

If the echo indications remain within a definite fluctuation in their amplitude as the test head arrangement is displaced parallel to the weld seam over a long region, and if the echo envelope curve rises sharply and then falls off as the test head arrangement is moved towards the weld seam and then away from the weld seam, this indicates an extensive defect in the weld seam direction.

If, as the test head arrangement is displaced parallel to the weld seam, the echo heights move within a definite region, i.e. the indications do not fall back as far as a threshold value, for example the zero line, then this represents an accumulation of reflectors in the weld seam direction.

This manual testing is obviously very subjective, and places high requirements on the test superintendent.

An attempt can be made however to substantially schematise the testing procedure, see Krautkramer Special Publication 187 and First/Alpine Report Ph 12/69.

There remains however the drawback that as a subjective element the test superintendent represents a substantial factor in the determination and evaluation of reflector indications, in which fluctuations and uncertainties in interpreting the echo indications are unavoidable. The factors acting to the disadvantage of automation are the necessarily high test capacity and the related high costs, the high cost of training test personnel and the lack of comparability in evaluating the results. Under such conditions, a repeated test is not always comparable with a test previously carried out, i.e. the tests are little reproducible. In addition, the time required for such manual testing is substantially greater than the time required for mechanical testing, however in this latter case the advantageous information received from the echo dynamic range has to be dispensed with.

According to the present invention there is provided a method for automatically obtaining a test result in the non-destructive testing of materials using ultrasonic pulses, wherein a digitalised A image is produced for a first position using a test head arrangement, and is stored in a suitable store, the position of the test head arrangement is varied in the X and/or Y direction with reference to coordinates of the test piece surface, and thus further A images are produced and stored, the variations in echo amplitude of the successive images are compared separately in the X and Y directions in a computer in accordance with a given operational sequence, with like dimensional reference values representing the echo amplitudes and the change in position of the test head arrangement and associated with given classes of defect, which are indicated, recorded and/or used as actuating variables for a subsequent stage of operation.

It is further advantageous if further data regarding the test piece can be present on the computer as a decision aid in forming the results.

The invention shall now be described in greater detail hereinafter, with reference to the accompanying drawings in which: Figure la shows a plate-shaped test piece with a plurality of test heads (test head arrangement) of different positions and sound radiation paths mounted on its surface, and which contains a spherical reflector.

Figure Ib shows a similar test piece which contains a flat reflector; Figure 2a, 2b is a graphical representation in which each upper half shows echo signal values (A image) as a function of the position of a determined test head for types of reflector in weld seams.

Figure 3 is a graphical representation of a succession of echo indications for a test piece with spherical reflectors.

Figure 4 is a graphical representation of a succession of echo indications for a test piece with flat reflectors.

Figure 5 is a block diagram of a device for carrying out the method using the computer.

A description will first be given of the relationship between the A images obtained during dynamic testing and the movement of the test head arrangement, in which the following abbreviations will be used: EK = individual test head TK = tandem arrangement of test heads A = echo height (echo amplitude) The indices indicate the angle between the noise pulse and the normal to the surface, e.g. EK70 = 700 test head angle.

Example 1. If a reflector 2 in the work piece 1 is subjected at different angles to sound radiation by the test heads EK70 etc., Figure 1 a, and if the echo amplitudes A of the various test heads are approximately of equal height, assuming uniform indication sensitivity, and if there is also no substantial difference in the echo indication when the reflector is subjected to sound from the opposite side, then the reflector must have a spherical boundary. It can be said that this is then a spherical reflector. The information that the amplitude heights for the various sound angles are approximately the same can be represented symbolically as follows: AEK70 Z AEK45 Z AEKO = ATK45 This is valid for sound radiation whether fed from the right or from the left.

Example 2. In the case of a reflector 3 of extensive area, Figure 1 b, various echo heights will be received from the various test heads, i.e. from the test heads feeding at different angles, and these can be symbolically represented as follows: AEK70 AEK45 + AEKO 8 ATK45 This is valid for sound radiation whether fed from the right or from the left.

Different radiation angles will obviously be chosen in order to identify defects in each angular position.

Figure 5 shows a block diagram for carrying out the method. The test head system 10 is guided over the surface of the test piece 1 by the manipulator 11 in accordance with the normal coupling methods, wherein the manipulator is controlled by the computer 13 by way of a control device 12 in accordance with a set test programme, so that the coordinates of the test head arrangement are known to the computer. In the case of manual operation, the manipulator transmits its actual coordinates to the computer by way of the control device. The pulse for the computer and test heads is produced by the ultrasonic device 14 in accordance with known computer techniques. The ultrasonic device also comprises an amplifier for the received pulse, evaluation units and A/D conversion for digitalising the A image, and these are not shown separately.The normal peripheral units such as a printer 15, a data store (tape recorder) 16 and input equipment and a display unit 17 are associated with and coupled to the computer.

If in example 1 a single echo is received when the test head arrangement is displaced both in X and in the Y direction, and this echo during the displacement increases from the threshold height, for example the zero line, and then falls back, this indicates a single pore (region a in Figure 2a).

If a sequence of separate individual echoes is received, i.e. during the displacement of the test head in one coordinate direction an echo increases and then fades within a set displacement stroke, and then within a further displacement stroke of the test head arrangement a new echo arises which fades again within this displacement stroke etc., this represents an accumulation of several pores (row of pores) (region bin Figure 2a). In both cases the echo dynamic range shows a maximum over the given range t of the time vector.

The echo dynamic range corresponds to the echo height amplitudes. The echo dynamic range indicates the range of variation of the envelopes of the succession of echo amplitudes.

If the A image gives an overlapping of several individual echoes which cannot be resolved (region c of Figure 2a), this is classified as a pore bunch in accordance with the given criteria. The echo dynamic range within the region At is substantially smaller than in the region a orb.

If in example 2 a single echo is obtained as the test head arrangement is displaced both in the X and Y direction (region a in Figure 2b), this echo being present only within a range At, this represents an approximately flat reflector surface with a longitudinal extension which is similar to or smaller than the sound beam cross-section. However, if this indication remains of approximately constant height as the arrangement is displaced in the X direction (region b of Figure 2b), this represents one reflector which extends in the X direction. If a succession of separate individual echoes is indicated during displacement in the X direction (region c, Figure 2b), this indicates several individual reflectors.

If when displacing in the X direction an indication is obtained having a substantially smaller dynamic range in a region At than in the region d, then this represents an accumulation of flat reflectors.

Likewise, superposed reflectors, i.e. reflectors in different depths of the test piece, can be classified by displacing the head arrangement in the Y direction.

Decision values obtained from previous experience are preset in a computer, which using the coordinates for the movement of the test head and the echo height in the A image classifies the reflectors. For this purpose, it is necessary for the A image to be fed in digitalised form to the computer in accordance with known methods. A store connected to the computer can store the data relative to the respective echo height and the subsequent classification of the reflectors. In the case of a repeat test, using the addresses of the surface coordinates, the results can be compared with each other and any result modification can be automatically determined.

The test is reproducible provided the other relevant data for the test are stored.

One example of the method is represented by classification into an individual pore, a pore accumulation, a pore bunch, a flat short single reflector (reflector extension smaller than the sound beam cross-section), a long flat individual reflector and an accumulation of flat reflectors.

Firstly, the computer determines whether the condition AEK70 t AEK45 z AEK0 a ATK70 z ATK45 etc.

or the condition AK70 * AEK4# * AEKO f ATK70 f ATK45 f etc.

is fulfilled. By this means, the computer immediately determines whether it is dealing with a 1. "Spherical reflector" class or 2. "Flat reflector" class.

1. Spherical reflector. In the case of a spherical reflector, in order to make any further decision a root width on the time vector must be set, preferably not in time units but in length units of the coordinate system. This root width, e.g. Ax, depends on the test equipment used. It must be somewhat greater than the root width of an echo from a small test reflector, and can be determined by measurements on a test reflector. Thus a test reflector, e.g. a bore, must be fed with sound. When displacing the test head arrangement in the X direction, an echo will now appear from the time vector, will grow and will then fade away (region a of Figure 2a). This region of the time vector has been indicated in the previous discussion as At.This region will now be known as Ax, as it relates to the displacement of the test head arrangement, with Ax = c. At, c being the velocity of sound. A characteristic quantity k is fixed, where k is slightly greater than Ax (Figure 2). The same applies in principle for the displacement of the test head arrangement in the Y direction. In this example, the same factor k is used for the Y direction. Whether the same factor is used in both directions as the decision aid for the computer depends on the radiation characteristics of the test heads, together with considerations of desirability. It is not the height of this decision factor which is important, but the fact that it is formed.

If during the manual or computer-controlled displacement of the test head arrangement, the amplitudes of the successive echoes fluctuate by more than a dynamic value w (region b of Figure 2a) measured in dB, a pore accumulation is present by definition. However, if the echo exceeds the w value only once during the displacement of the test head over the distance k, this indicates a single pore (region a of Figure 2a). The quantity w is preset as a dynamic range dimension according to the test problem. The size of w as a further decision factor is likewise not important, but what is important is that this quantity is formed.

This is further clarified in Figure 3.

In displacing the test head arrangement in the X direction from position X1 to X2, echoes appear, the indications of which within the displacement stroke k always return to the zero line, i.e. the dynamic range value w is exceeded within the region k. This is classified as a row of pores. Between positions X2 and X3, there is no indication during the displacement through k. This region is free of reflectors. In position X3, an echo is present which further fades within the region k. A single pore is to be associated with this echo. There is no indication as far as position X4. This region is free from reflectors. Between X4 and X5, echo indications are present which fluctuate to a smaller degree than the value w, but the displacement distance of the test head arrangement over which echoes appear is greater than k, so that this is to be classified as a pore bunch.

Between X5 and Xe, which represents a distance greater than the value k, there is no indication. Thus this region is to be evaluated as free from reflectors, and in other words this represents the distance of separation from the next pore bunch between X6 and X7, as further indications having a dynamic range smaller than w are present between position Xe and X7. If the distance between X5 and X6 which showed no indication had been smaller than the dimension k, then both bunches would have had to be classified by the computer as a continuous bunch. The displacements of the test head arrangement in the Y direction give indication regarding the extension of the reflector regions in the direction of the weld seam width.It is advantageous, but not absolutely necessary in any individual case, to use the same decision factors k, w forthe direction as for the X direction.

In addition, in making the classification: "single pore bunch"/"separated pore bunches", it is not absolutely necessary to use the factor k obtained from the root width, but instead a different factor can be preset.

2. Flat reflector. If the computer has decided that it is dealing with a flat reflector (Figure 4), then the computer has to decide in accordance with the following criteria: With reference to Figure 4, the test head arrangement is again moved, for example from left to right, i.e. in the X direction. In position X1, an echo occurs having the characteristic that it is only present for a test head displacement smaller than the factor k.

Thus this indication is associated with a flat individual reflector smaller than the sound beam cross-section.

No indication occurs as far as position X2. This region is free from reflectors. In X2, an echo occurs which remains approximately constant in its indication as far as position X3, i.e. the dynamic amplitude range is smaller than a given value v, which as in the case of the value w is determined from test considerations and is shown in Figure 4 by two horizontal lines as a dynamic range. The computer classifies this indication as a longitudinally extending reflector. No indications occur between position X3 and position X4. This region is free from reflectors. Between X4 and the position X5, indications appear having a height which fluctuates more than the dynamic range value w, so that this region is classified by the computer as a row of reflectors.

Between X5 and X6 there is no indication. This region is evaluated as free from reflectors, provided the displacement is greater than k. Between Xe and X7 an indication appears having local amplitude fluctuations smaller than the value w but greater than the value v, and this is classified by the computer as a reflector bunch.

Between X7 and X8 no indication appears. Between Xe and Xg an indication appears having the same characteristics as when moving the test head arrangement from X6 to X7, SO that this is classified as a further reflector bunch, provided the displacement stroke X7 to Xe is greater than k. If the displacement stroke X7 to Xe is smaller than the value k, then the computer has to evaluate the second reflector bunch together with the first reflector bunch as one larger continuous reflector bunch between X6 and Xg.

In displacing the test head arrangement in the Y direction, either the same decision criteria can be valid, or other decision quantities can be chosen according to the test problem.

AEK 70 # AEK 45 z AEK o # A70 Coincidence: spherical reflectors AAx > w;Ax < k Single pore AAy > w; Ay < k A Ax > w; #x > k Row of pores in X direction AAy > w;Ay < k AAx < w;Ax > k Pore bunch #Ay < w;#y > k AEK 70 AEK 45 f AEK o f ATD f no coincidence: flat reflectors A Ax > w; #x < k short flat single reflector in X direction AAy > w;Ay < k A Ax > w; Ax > k row of flat single reflectors #Ay > w;#y < k in the X direction a x < v#x > k long flat reflector in X direction #Ay > w;#y < k v < A Ax w; Ax > k Accumulation of flat reflectors v < #Ayw;#y > k AEK 70 t AEK 45 z AEK 0#...

existence of ambiguity unclassifiable examination of installation, possibly manually testing or retesting disturbance Table 1 summarises the decision characteristics for the computer. Thus the computer has alternatively to decide in accordance with the plan of table 1 only whether a first condition is fulfilled or not, in order to then test the subsequent conditions in accordance with planned logic, and to print out the test report from the results. Classification of the reflector position, the depth of a reflector in the test piece and information regarding the size of the reflector are obtained from the coordinates of the manipulator, the propagation time and the echo height in known manner.

According to a further embodiment of the invention, the echo amplitudes are compared in known manner with preset values after compensating for depth, the method being only effective for echo amplitudes which exceed the preset value, by which the reflectors which produce these echoes are automatically classified as defects.

Reflectors producing smaller echo amplitudes than this preset value are not subjected to evaluation, i.e.

are not evaluated as defects.

The reflector position is obtained from the propagation time by known methods.

As further automation, in testing weld seams, the position of the weld seam in terms of its coordinates can be fed to the computer, so that the computer only evaluates echoes from this given region. Furthermore, the seam shape can be fed to the computer in the form of corresponding data, in order to evaluate determined classes of reflector differently, so that for example in deciding whether a reflector is to be viewed as a defect or not, the computer evaluates the relationship between the size of defect and the geometry of the seam, and this is used as the decision quantity. Data regarding the welding process can also be fed to the computer, so that reflectors which may be typical of the process are not evaluated as defects.

The method enables the advantages of the defect dynamics in manual testing to be used in mechanical or computer-controlled, i.e. automated, tests with ultrasonic pulses, such that signal values are fed to the computer as decision criteria, the computer thus being compelled to carry out the steps required for classifying the reflectors, and to classify them either as a defect or a non-defect.

Claims (7)

1. A method for automatically obtaining a test result in the non-destructive testing of materials using ultrasonic pulses, wherein a digitalised A image is produced for a first position using a test head arrangement, and is stored in a suitable store, the position of the test head arrangement is varied in the X and/orY direction with reference to coordinates of the test piece surface, and thus further A images are produced and stored, the variations in echo amplitude of the successive images are compared separately in the X and Y directions in a computer in accordance with a given operational sequence, with like-dimensional reference values representing the echo amplitudes and the change in position of the test head arrangement and associated with given classes of defect, which are indicated, recorded and/or used as actuating variables for a subsequent stage of operation.

2. A method as claimed in Claim 1, wherein only echoes having a minimum amplitude are classified by the computer in accordance with set decision qualities.

3. A method as claimed in Claim 1 or Claim 2, wherein the coordinates of a region of a test piece are present in the computer to allow classification of the images in that region only.

4.' A method as claimed in any preceding Claim, wherein the test head arrangement used consists of several test heads of different radiation characteristics.

5. A method as claimed in ay preceding Claim, wherein the given operational sequence is as hereinbefore described in table 1.

6. A method as claimed in ay preceding Claim, wherein the like-dimensional reference values used are electrical voltages.

7. A method for automatically obtaining a test result in the non-destructive testing of materials using ultrasonic pulses, substantially as hereinbefore described with reference to the accompanying drawings.