Processing signals acquired during guided wave testing

Abstract

Processing signals acquired during guided wave testing A method of processing signals acquired during guided wave testing of an elongate member (2), such as a pipe, in which at least one guided wave (7) is generated in the elongate member, the at least one guided wave is reflected by reflectors (8) in the elongate member and reflected guided waves (9) are detected. The method comprises determining at least one reflection coefficient or a parameter for calibrating a guide wave test in dependence upon reflections from the reflectors which include at least one multiple reflection. The reflections may include a single reflection from a first reflector, a single reflection from a second reflector and a multiple reflection from the first and second reflectors.

Claims

1. A computer-implemented method of processing signals and calibrating a guided wave test, the method comprising: receiving signals acquired during guided wave testing of an elongate member, the guided wave testing including first and second guided waves generated in the elongate member on first and second, opposite sides respectively of a transducer ring, the first and second guided waves propagated in first and second opposite directions respectively, the guided waves reflected by reflectors in the elongate member and reflected guided waves detected; determining at least one reflection coefficient or a parameter for calibrating a guided wave test in dependence upon reflections from the reflectors which include a single reflection from a first reflector on the first side, a single reflection from a second reflector on the second side and a multiple reflection from the first and second reflectors, wherein determining the parameter for calibrating a guided wave test comprises: calculating an amplitude, A.sub.0, of the first and second guided waves, wherein:
A.sub.0 =sqrt((A.sub.1).sup.2A.sub.2/A.sub.121) wherein: A.sub.1 is a peak value of a peak corresponding to the single reflection of the first guided wave from the first reflector, A.sub.2 is a peak value of a peak corresponding to the single reflection of the second guided wave from the second reflector, and A.sub.121 is a peak value of a peak corresponding to a triple reflection of the first guided wave from, in order, the first reflector, the second reflector and the first reflector; and calibrating the guided wave test on the elongate member by fitting a distance amplitude correction (DAC) curve using the amplitude, A.sub.0.

2. A computer-implemented method according to claim 1, wherein the first and second guided waves comprise respective pulses and the amplitudes are amplitudes of peaks corresponding to respective reflections of the pulses.

3. A computer-implemented method according to claim 1, comprising: transmitting signals to at least two sets of transducers spaced apart along the elongate member; receiving signals from at least two sets of transducers spaced apart along the elongate member; and extracting signal contributions corresponding to even-numbered reflections.

4. A computer-implemented method according to claim 3, the method comprising: displaying a plot using the extracted signal contributions.

5. A computer-implemented method according to claim 1, wherein the second reflector used to determine the reflection coefficient or parameter is located outside a dead zone or near field and the first reflector used to determine a reflection coefficient or parameter is located inside the dead zone or near field.

6. A computer-implemented method, comprising: generating the first and second guided waves in the elongate member; acquiring signals from transducers coupled to the elongate member; and processing the signals according to claim 1.

7. A non-transitory computer-readable medium storing a computer program which, when executed by a computer, causes the computer to perform a method according to claim 1.

8. Apparatus configured to perform the method according to claim 1, the apparatus comprising: transducers for generating at least one guided wave in an elongate member, and instrumentation for generating a set of signals for driving the transducers and receiving signals generated by the transducers.

9. Apparatus according to claim 8, wherein the computer and instrumentation are unitary.

10. A computer-implemented method according to claim 1, wherein the parameter for calibrating the guided wave test is the amplitude, A.sub.0, of first and second guided waves.

11. A computer-implemented method of processing signals and calibrating a guided wave test, the method comprising: receiving signals acquired during guided wave testing of an elongate member, the guided wave testing including first and second guided waves generated in the elongate member on first and second, opposite sides respectively of a transducer ring, the first and second guided waves propagated in first and second opposite directions respectively, the guided waves reflected by reflectors in the elongate member and reflected guided waves detected; determining at least one reflection coefficient or a parameter for calibrating a guided wave test in dependence upon reflections from the reflectors which include a single reflection from a first reflector on the first side, a single reflection from a second reflector on the second side and a multiple reflection from the first and second reflectors, wherein determining the parameter for calibrating a guided wave test comprises: calculating an amplitude, A.sub.0, of first and second guided waves, wherein:
A.sub.0=A.sub.1A.sub.2/A.sub.12 wherein: A.sub.1 is a peak value of a peak corresponding to the single reflection of the first guided wave from the first reflector, A.sub.2 is a peak value of a peak corresponding to the single reflection of the first guided wave from a second reflector, and A.sub.12 is a peak value of a peak corresponding to a double reflection of the first guided wave from the first reflector and then the second reflector; and, calibrating the guided wave test on the elongate member by fitting a distance amplitude correction (DAC) curve using the amplitude, A.sub.0.

12. A computer-implemented method according to claim 11, wherein the first and second guided waves comprise respective pulses and the amplitudes are amplitudes of peaks corresponding to respective reflections of the pulses.

13. A computer-implemented method according to claim 11, comprising: transmitting signals to at least two sets of transducers spaced apart along the elongate member; receiving signals from at least two sets of transducers spaced apart along the elongate member; and extracting signal contributions corresponding to even-numbered reflections.

14. A computer-implemented method according to claim 11, wherein the second reflector used to determine the reflection coefficient or parameter is located outside a dead zone or near field and the first reflector used to determine a reflection coefficient or parameter is located inside the dead zone or near field.

15. A computer-implemented method, comprising: generating the first and second guided waves in the elongate member; acquiring signals from transducers coupled to the elongate member; and processing the signals according to claim 11.

16. A computer-implemented method according to claim 11, the method comprising: displaying a plot using the extracted signal contributions.

17. A non-transitory computer-readable medium storing a computer program which, when executed by a computer, causes the computer to perform a method according to claim 11.

18. Apparatus configured to perform the method according to claim 11, the apparatus comprising: transducers for generating at least one guided wave in an elongate member, and instrumentation for generating a set of signals for driving the transducers and receiving signals generated by the transducers.

19. Apparatus according to claim 18, wherein the computer and instrumentation are unitary.

20. A computer-implemented method according to claim 11, wherein the parameter for calibrating the guided wave test is the amplitude, A.sub.0, of first and second guided waves.

21. A computer-implemented method of processing signals and calibrating a guided wave test, the method comprising: receiving signals acquired during guided wave testing of an elongate member, the guided wave testing including first and second guided waves generated in the elongate member on first and second, opposite sides respectively of a transducer ring, the first and second guided waves propagated in first and second opposite directions respectively, the guided waves reflected by reflectors in the elongate member and reflected guided waves detected; determining at least one reflection coefficient or a parameter for calibrating a guided wave test in dependence upon reflections from the reflectors which include a single reflection from a first reflector on the first side, a single reflection from a second reflector on the second side and a multiple reflection from the first and second reflectors, wherein determining the at least one reflection coefficient includes: calculating a reflection coefficient, R.sub.1, of the first reflector and/or a reflection coefficient, R.sub.2, of the second reflector, wherein:
R.sub.1=exp(2?|z.sub.1|) sqrt(A.sub.121/A.sub.2)
R.sub.2=exp(2?|z.sub.2|) sqrt(A.sub.121 A.sub.2/(A.sub.1).sup.2) wherein: A.sub.1 is a peak value and z.sub.1 is the peak position of a peak corresponding to the single reflection of the first guided wave from the first reflector, A.sub.2 is a peak value and z.sub.2 is the peak position of a peak corresponding to the single reflection of the second guided wave from the second reflector, A.sub.121 is a peak value of a peak corresponding to a triple reflection of the first guided wave from, in order, the first reflector, the second reflector and the first reflector, and ? is an attenuation coefficient having a value zero or a positive non-zero number; and, calibrating the guided wave test on the elongate member by fitting a distance amplitude correction (DAC) curve using the reflection coefficient, R.sub.1, of the first reflector and/or the reflection coefficient, R.sub.2, of the second reflector.

22. A computer-implemented method of processing signals and calibrating a guided wave test, the method comprising: receiving signals acquired during guided wave testing of an elongate member, the guided wave testing including first and second guided waves generated in the elongate member on first and second, opposite sides respectively of a transducer ring, the first and second guided waves propagated in first and second opposite directions respectively, the guided waves reflected by reflectors in the elongate member and reflected guided waves detected; determining at least one reflection coefficient or a parameter for calibrating a guided wave test in dependence upon reflections from the reflectors which include a single reflection from a first reflector on the first side, a single reflection from a second reflector on the second side and a multiple reflection from the first and second reflectors, wherein determining the at least one reflection coefficient includes: calculating a reflection coefficient, R.sub.1, of the first reflector and/or a reflection coefficient, R.sub.2, of the second reflector, wherein:
R.sub.1=exp(2?|z.sub.1|)A.sub.12/A.sub.2
R.sub.2=exp(2?|z.sub.2|)A.sub.12/A.sub.1 wherein: A.sub.1 is a peak value and z.sub.1 is the peak position of a peak corresponding to the single reflection of the first guided wave from the first reflector, A.sub.2 is a peak value and z.sub.2 is the peak position of a peak corresponding to the single reflection of the second guided wave from the second reflector, A.sub.12 is a peak value of a peak corresponding to a double reflection of the first guided wave from the first reflector and then the second reflector, and ?is an attenuation coefficient having a value zero or a positive non-zero number; and, calibrating the guided wave test on the elongate member by fitting a distance amplitude correction (DAC) curve using the reflection coefficient, R.sub.1, of the first reflector and/or the reflection coefficient, R.sub.2, of the second reflector.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) Certain embodiments of the present invention will now be described, by way of example, with reference to FIGS. 2 to 10b of the accompanying drawings in which:

(2) FIG. 1a is a plot of pulse-echo signal amplitude against distance for test of a pipe having a square end cut, two welds and a defect including 0 dB and ?16 dB distance amplitude correction (DAC) curves fitted to the peak amplitude of the peak corresponding to reflection from the weld;

(3) FIG. 1b is the same plot as that shown in FIG. 1a but which includes 0 dB and ?16 dB DAC curves fitted to the peak amplitude of the peak corresponding to reflection from the square cut end;

(4) FIG. 2 illustrates a guided wave inspection system in accordance with the present invention;

(5) FIG. 3 is a schematic block diagram of a computer system used in the guided wave testing system shown in FIG. 2;

(6) FIG. 4a illustrates propagation of guided waves in a pipe in which a guided wave is reflected multiple times between two reflectors and the reflections are received from the same side of the pipe into which the outgoing waves are sent;

(7) FIG. 4b schematically illustrates a plot of signal amplitude against distance for the reflected waves shown in FIG. 4a;

(8) FIG. 5a illustrates propagation of guided waves in pipe in which a guided wave is reflected multiple times between two reflectors and the reflections are received from the opposite side of the pipe into which the outgoing waves are sent;

(9) FIG. 5b schematically illustrates a plot of signal amplitude against distance for the reflected waves shown in FIG. 5a;

(10) FIG. 6 is a process flow diagram of a method of determining the size of a reflector;

(11) FIG. 7a is a linear plot of pulse-echo signal amplitude against distance for a guided wave test of a pipe, the plot showing reflections which are received from the same side of the pipe into which outgoing waves are sent;

(12) FIG. 7b illustrates the same data shown in FIG. 7a except using a logarithmic scale;

(13) FIG. 8a is a plot of signal amplitude against distance for a guided wave test of another pipe, the plot showing reflections which are received from the same side of the pipe into which outgoing waves are sent;

(14) FIG. 8b is a plot of signal amplitude against distance for a guided wave test of another pipe, the plot showing reflections which are received from the opposite side of the pipe into which outgoing waves are sent;

(15) FIG. 9 illustrates propagation of guided waves in pipe where a row of transducers acts as a reflector;

(16) FIG. 10a illustrates propagation of guided waves in pipe in which a guided wave is reflected multiple times between two reflectors on the same side of a transducer ring; and

(17) FIG. 10b schematically illustrates a plot of signal amplitude against distance for the reflected waves shown in FIG. 10a.

DETAILED DESCRIPTION OF CERTAIN EMBODIMENTS

(18) Referring to FIG. 2, a system 1 for inspecting a pipe 2 or other elongate member using guided waves is shown. The system 1 includes a transducer ring 3 having at least two rows or sets 4.sub.1, 4.sub.2 of transmitting/receiving transducers 5 spaced apart along the longitudinal axis 6 of the pipe 2. In this example, the transducers 5 are discrete, piezoelectric transducers and the transducers 5 in each set 4.sub.1, 4.sub.2 are angularly-spaced around the perimeter of the pipe 2 in a similar way to that described in WO 96 12951 A ibid. However, the transducers 5 need not be discrete and other types of transducers can be used.

(19) The transducer ring 3 is used to excite guided waves 7 in the form of pulses in the pipe 2 at a propagating frequency, typically in the range between about 10 and about 500 kHz, which propagate along the pipe 2. Discontinuities 8, such as defects, welds or other types of interface (herein collectively referred to as reflectors) reflect some or all of the outgoing waves 7 as reflected waves 9. At least two sets 4.sub.1, 4.sub.2 of transducers 5 are used so that the direction of propagation of reflected waves 9 can be determined. As will be explained later, an outgoing wave 7 can be reflected once, twice, three times or more. Moreover, a wave 7 which is multiply reflected between a pair of reflectors 8 can be used to help determine the size of reflectors 11 more accurately. Reflected waves 9 are recorded in pulse-echo and pitch-catch arrangements using the sets of transducers controlled by guided wave instrumentation 10 and a computer 11. The guided wave instrumentation 10 is a microprocessor-controlled system which sends electrical signals in the form of tone bursts (or wavepackets) to the transducers 5 in the transducer ring 3. The transducers 5 convert the electrical signals into mechanical vibrations and excite guided waves 7 in the pipe 2. The transducers 5 also convert the mechanical vibration of reflected waves 9 into electrical signals which are fed to the guided wave instrumentation 10 where they can be amplified and/or filtered and digitised. The electrical signal is a function of the mechanical vibrations. The guided wave instrumentation 10 and/or computer 11 can record the electrical signals and can process the signals to select waves of a given directionality and/or modes. For example, this may involve filtering out or subtracting contributions to a signal from unwanted reflections and/or modes. The instrumentation 10 or computer 11 can convert a time-varying electrical signal which can include contributions from different reflections into an amplitude-distance plot using the equation z=?.Math.t/2 where ?, is the velocity of the guided wave, t is time and z is distance. The signal trace is often referred to as an A-scan type plot and shows the envelope of the rectified electrical signal.

(20) Referring also to FIG. 2, the computer system 11 is shown in more detail. The computer system 11 includes one or more processors 21, memory 22 and an input/output (I/O) interface 23 operatively connected by a bus 24. The I/O interface 23 is operatively connected to, among other things, an instrumentation interface 25, a display 26 (which can take the form of a touch screen), user input device(s) 27 (such as a keyboard and a pointing device or touch screen), and storage 28 (for example in the form of a hard disk drive or non-volatile memory).

(21) Computer program code 29 is held in storage 28 and loaded into memory 22 for execution by the processor 21.

(22) The computer system 11 carries out a sizing and calibration method based on parts or components of the signals received from the transducers 5 corresponding to reflections from the reflectors which include multiple reflections between reflectors. In particular, the method is based on parts or components of the signals received from the transducers 5 corresponding to single reflections received from one or both of two reflectors and a multiple reflection received after reverberating between the two reflectors. In this example, the amplitudes of peaks corresponding to reflected pulses are used. Using the amplitude of the peak corresponding to a reflection from a first reflector, the amplitude of the peak corresponding to a reflection from a second reflector, and the amplitude of the peak corresponding to a multiple reflection between the first and the second reflectors, the sizes of the two reflectors can be determined. This can be further used to determine the size of other reflectors either by calculation or by using distance amplitude correction (DAC) curve(s).

(23) The calibration method provides a way of sizing guided wave signals without the use of artificially-introduced attachments, such as clamps. Nevertheless, clamps can be used to create a reflection for sizing purposes. The use of clamps can be helpful if no significant reflectors can be observed.

(24) In the following, sizing and calibration methods are described which employ signals arising from two reflectors 8 located on opposite sides of the transducer ring 3. Generally, this arrangement occurs most often and the reflections are most easily observed. However, the methods can be based on reflections received from reflectors 8 that are both located on only one side of the transducer ring 3.

(25) Sizing and calibration methods can be based on double or triple reflections.

(26) Referring to FIG. 4a, the pipe 2 is shown having first and second reflectors 8.sub.1, 8.sub.2 located on opposite sides of the transducer ring 3. The transducer ring 3 is used to excite a first guided wave in the form of a pulse which propagates backwards 30 (shown here to be along the negative z direction) along the pipe 2 and a second guided wave in the form of a pulse which propagates forwards 31 having the same amplitude. The guided waves are reflected by the reflectors 8 and the reflected waves are detected by the transducer ring 3.

(27) The transducer ring 3 receives first and second direct pulse-echo (i.e. single reflections) reflected from the first and second reflectors 8.sub.1, 8.sub.2 respectively and having paths z: 0.fwdarw.?z.sub.1.fwdarw.0 and z: 0.fwdarw.z.sub.2.fwdarw.0. The first and second direct pulse-echoes have peak amplitude values A.sub.1 and A.sub.2 respectively.

(28) The transducer ring 3 receives a first triple-reflection which has been reflected first by the first reflector 8.sub.1, then by the second reflector 8.sub.2 and then again by the first reflector 8.sub.1, i.e. a path z: 0.fwdarw.?z.sub.1.fwdarw.z.sub.2.fwdarw.z.sub.1.fwdarw.0. In this case, the measured signal has a peak amplitude A.sub.121 and corresponds to a reflection received by the transducer ring 3 from the direction of the first reflector 8.sub.1.

(29) The transducer ring 3 also receives a second triple-reflection which has been reflected by the second reflector 8.sub.2, then by the first reflector 8.sub.1 and then again by the second reflector 8.sub.2, i.e. a path z: 0.fwdarw.z.sub.2.fwdarw.?z.sub.1.fwdarw.z.sub.2.fwdarw.0. In this case, the measured signal has a peak amplitude A.sub.212 and corresponds to a reflection received by the transducer ring 3 from the direction of the second reflector 8.sub.2.

(30) FIG. 4b is an amplitudedistance plot showing peaks in amplitude corresponding to odd-numbered reflections, in particular single and triple reflections. The characteristic distances at which the peaks appear can be used to identify reverberation echoes.

(31) Referring to 5a, the same pipe 2 is shown. FIG. 5a differs from FIG. 4a in that instead of odd-numbered reflections, even-numbered reflections, in particular double reflections, are shown. However, single reflections are also used when calculating reflection coefficients.

(32) The transducer ring 3 receives a first double-reflection which has been reflected first by the first reflector 8.sub.1 and then by the second reflector 8.sub.2, i.e. a path z: 0.fwdarw.?z.sub.1.fwdarw.z.sub.2.fwdarw.0. In this case, the measured signal has a peak amplitude A.sub.12 and corresponds to a reflected wave received by the transducer ring 3 from the opposite direction of the first reflector 8.sub.1.

(33) The transducer ring 3 also receives a second double-reflection which has been reflected by the second reflector 8.sub.2 and then by the first reflector 8.sub.1, i.e. a path z: 0.fwdarw.z.sub.2.fwdarw.?z.sub.1.fwdarw.0. In this case, the measured signal has a peak amplitude A.sub.21 and corresponds to a reflected wave received by the transducer ring 3 from the opposite direction of the second reflector 8.sub.2.

(34) FIG. 5b is an amplitudedistance plot showing peaks in amplitude corresponding to double reflections. FIG. 5c is an amplitudedistance plot showing peaks corresponding to single reflections.

(35) The plot shown in FIG. 5b is obtained by taking into account the directions in which outgoing and reflected waves propagate and extracting signal contributions arising from an even number of reflections. The guided wave instrumentation 10 (FIG. 2) and/or the computer system 11 (FIG. 2) can process the signals received from two or more rings of transducers 4.sub.1, 4.sub.2 (FIG. 2) to produce a signal trace having peaks which arise from an outgoing wave propagating in one direction, i.e. sent out into one side of the transducer ring, and waves propagating in the same direction received on the opposite side of the transducer ring 3 (FIG. 2) and vice versa. Thus, the plot shows peaks arising from double and (if detectable) other even-numbered reflections, but not from odd-numbered reflections such as single and triple reflections. Therefore, this type of plot only shows peaks arising from even-numbered reverberations (i.e. there are no single reflections) and so this type of plot is herein referred to as a reverberation plot. The plot is characterised by being symmetrical about the origin (i.e. z=0). Thus, in some embodiments, only one side of the trace need be plotted.

(36) Referring in particular to FIG. 5b, the negative and positive sides 32, 33 of the plot shown in FIG. 5b can be obtained in a several different ways. The negative side 32 of the plot (i.e. z?0) can be obtained by generating a pulse in the pipe 2 propagating in a negative (or backwards) direction 30, for example, by driving two transducer rings 4.sub.1, 4.sub.2 (FIG. 2) with appropriate phase difference and detecting reflected pulses propagating in the same direction (i.e. backwards) by measuring phase difference or time delay between the signals received from the two transducer rings 4.sub.1, 4.sub.2 (FIG. 2). Likewise, the positive side 33 of the plot (i.e. 0?z) can be obtained by generating pulses in the pipe propagating in the positive direction 31 and detecting the reflected pulses propagating in a positive direction 31. The guided wave instrumentation 10 and/or computer system 11 processes the signals to identify contributions or components (in this case peaks) corresponding to forward and backward propagating pulses and extract only those contributions or components attributable to even-numbered reflections.

(37) In particular, using two sets of transceivers 4.sub.1, 4.sub.2, extraction is achieved using first and second pulse-echo traces (in which a transducer set excites the outgoing wave and the same transducer set detects the reflection) for the first and second transducer sets 4.sub.1, 4.sub.2 and one pitch-catch trace (in which one transducer set excites the outgoing wave and another transducer set detects the reflection). Each measured trace comprises a different mixture of three independent component traces corresponding to reflections travelling in forward and reverse directions arising from outgoing pulses travelling in forwards and reverse directions.

(38) The traces are processed using fast Fourier transforms in the frequency domain to solve three unknowns (i.e. component traces), namely a first component trace corresponding to pulses transmitted in the forward direction and reflected pulses received propagating in the opposite direction (i.e. backwards), a second component trace corresponding to pulses transmitted in the backward direction and reflected pulses received propagating in the opposite direction (i.e. forwards), and a third component trace corresponding to pulses transmitted in one direction (e.g. forwards) and reflected pulses received propagating in the same direction (e.g. in the case of forward transmitted pulses, the received reflected waves also propagate forwards). Thus, the third component trace corresponding to forward- (or backward-) going transmitted pulses and forward- (or reverse-) going reflections can be extracted and used as a reverberation plot.

(39) Methods based on double reflections are generally preferred because the use of double reflections tends to avoid interference of the double reflection with other reflections. For example, when welds are equally spaced, triple reflections can interfere with a weld reflection and so may not be observed. Furthermore, the signal-to-noise ratio of a double reflection signal tends to be better than a triple reflection signal because it involves fewer reflections.

(40) However, methods based on triple reflections are simpler from the point of view that signals can be sent out in one direction and reflected signals are received from the same direction. When double reflections are used, signals are sent out in one direction and reflections are received from both directions, thus requiring more complex signal processing.

(41) Referring also to FIG. 2, the transducer ring 3 picks up and converts mechanical signals into electrical signals which can be filtered, amplified and pre-processed by the guided wave instrumentation 10 and output as time traces to the computer system 11. The computer system 11 can then be used to size the reflectors and to calibrate signals.

(42) Sizing Reflectors

(43) First, sizing of reflectors based on triple reflections will be described.

(44) The reflection coefficients R.sub.1 and R.sub.2 for first and second reflectors can be calculated as follows. The amplitudes A.sub.1, A.sub.2, A.sub.121 and A.sub.212 can be expressed as:
A.sub.1=A.sub.0 R.sub.1 exp(?2?|z.sub.1|) (1-1)
A.sub.2=A.sub.0 R.sub.2 exp(?2?|z.sub.2|) (1-2)
A.sub.121=A.sub.0(R.sub.1).sup.2 R.sub.2 exp(?2?{2|z.sub.1|+|z.sub.2|}) (1-3)
A.sub.212=A.sub.0(R.sub.2).sup.2 R.sub.1 exp(?2?{|z.sub.1|+2|z.sub.2|}) (1-4)
where is ? an attenuation factor which can be zero or a positive non-zero number (i.e. 0??) and A.sub.0 is the amplitude of the outgoing wave.

(45) Equations (1-1), (1-2) and (1-3), for example, can be re-arranged as:
R.sub.1=exp(2?|z.sub.1|)sqrt(A.sub.121/A.sub.2) (1-5)
R.sub.2=exp(2?|z.sub.2|)sqrt(A.sub.121 A.sub.2/(A.sub.1).sup.2) (1-6)
A.sub.0=sqrt((A.sub.1).sup.2 A.sub.2/A.sub.121) (1-7)

(46) Equations (1-1), (1-2) and (1-4) can be re-arranged in a similar way.

(47) The reflection coefficient R.sub.1 (and/or R.sub.2) for the first and/or second reflector can be found using equations (1-5) and (1-6).

(48) Referring to FIGS. 2 and 6, a pulse-echo measurement is carried out to acquire time trace data which includes a set of peaks including peaks corresponding to single reflections and peaks corresponding to the triple reflection (step S1). The peak amplitude(s) of single reflection(s) from one or both of two reflectors, i.e. A.sub.1 and/or A.sub.2, is (are) measured (step S2). The peak amplitudes of the triple reflections, i.e. A.sub.121 and/or A.sub.121, are also measured (step S3).

(49) The reflection coefficient for the first reflector, i.e. R.sub.1, is calculated by dividing the amplitude of the peak corresponding to the triple reflection, i.e. A.sub.121, by the amplitude of the peak corresponding to a single reflection from the second reflector, i.e. A.sub.2, and taking the square root (step S4). The attenuation factor, ?, can be set to zero or determined using one of the methods described later (step S4).

(50) A similar calculation can be used to determine the reflection coefficient for the second reflector, i.e. R.sub.2. Also, the values can be used to determine the amplitude of the outgoing wave, A.sub.0.

(51) Calibration involves fitting a distance amplitude correction (DAC) curve to one or more peaks corresponding to reflections from respective one or more reference features (such as welds) which are assumed to have the same reflectivity. The curve exhibits a characteristic fall off in amplitude attributable to attenuation. Thus, two features which have the same reflection coefficient, but which are located at different positions along a pipe, produce peaks having different peak amplitudes.

(52) Once a DAC curve for a given reflectivity has been fitted to peaks whose reflectivity is known or determinable, attenuation can be found and further DAC curves for other values of reflectivity can be constructed. These other curves can be used to determine the reflectivity of other features.

(53) Fitting a curve involves shifting and stretching the curve so that it intercepts the peak amplitudes. This can be done by an operator or by the computer system 11.

(54) The sizing method hereinbefore described is used to calculate a reflection coefficient for a given feature. A DAC curve for the reflection coefficient can then be constructed and then fitted to the feature (step S6). The size of other features can then reported by the software with respect to this curve. Data, such as attenuation coefficients and reflection coefficients can be stored in storage 28 (FIG. 3) (step S7).

(55) FIG. 7 shows the result of a guided wave test using 20 kHz, torsional mode guided waves of an in-service pipe (not shown) having an outer diameter of 114.3 mm and a nominal wall thickness of 6.02 mm, and having a branch and two girth welds.

(56) From the measurement, the following peak amplitudes can be measured, namely A.sub.121=0.77 mV, A.sub.1=18.99 mV, A.sub.2=5.42 mV. The peaks having amplitudes A.sub.1 and A.sub.3 correspond to welds and the peak having an amplitude A.sub.2 corresponds to a branch.

(57) Assuming that ?=0, the reflection coefficient of the first feature (a weld) is R.sub.1=0.14. The distance, measured from (the middle of) the transducer ring 3, of the first feature (a weld) is ?1.32 m and for a second feature (branch) is +1.87 m. The distance to the reverberation is measured as ?4.52 m. The reverberation is 3.20 m behind feature 1, which is approximately equal to the distance between feature 1 and 2.

(58) A.sub.121 and A.sub.212 are related to each other by A.sub.121/A.sub.212=(A.sub.1).sup.2/(A.sub.2).sup.2.

(59) Sizing of reflectors based on double reflections will be described.

(60) The amplitudes A.sub.1, A.sub.2 and A.sub.12 can be expressed as:
A.sub.1=A.sub.0 R.sub.1 exp(?2?|z.sub.1|) (2-1)
A.sub.2=A.sub.0 R.sub.2 exp(?2?|z.sub.2|) (2-2)
A.sub.12=A.sub.21=A.sub.0 R.sub.1 R.sub.2 exp(?2?{|z.sub.1|+|z.sub.2|}) (2-3)

(61) These can be re-arranged as:
R.sub.1=exp(2?|z.sub.1|)A.sub.12/A.sub.2 (2-4)
R.sub.2=exp(2?|z.sub.2|)A.sub.12/A.sub.1 (2-5)
A.sub.0=A.sub.1 A.sub.2/A.sub.12 (2-6)

(62) FIG. 8 shows the result of a guided wave test using 36 kHz, torsional mode guided waves of an in-service pipe (not shown) having an outer diameter of 406.4 mm and a nominal wall thickness of 6.35 mm, and having, amongst other features, two girth welds.

(63) From the measurement, the following peak amplitudes can be measured, namely A.sub.21=A.sub.12=0.15 mV, A.sub.1=1.34 mV and A.sub.21.25 mV.

(64) Assuming that ?=0, the first reflection coefficient, R.sub.1, is 0.14 and the second reflection coefficient, R.sub.2, is 0.11. The first feature is ?9.96 m from the transducer ring position and the second feature is +2.51 m from the transducer ring. The distance between the first and second features is 12.45 m, which is equal to the distance where the reverberation occurs.

(65) Determining Attenuation

(66) As explained earlier, the attenuation factor, ?, can be set to zero if it is small. The exponential term containing the attenuation factor then has a value of one. This approximation may be appropriate also when the distance between two reflectors is small, e.g. a few meters.

(67) However, the attenuation factor can be found at step S4 (FIG. 6), for example, in one of two ways.

(68) In a first method, the transducer ring 3 is moved a distance, ?z, between two reflectors and the change in scale of the guided wave signals is observed. In order for this approach to be independent of the coupling of the transducer ring at the two test positions, the guided wave signals of two reflectors located on either side of the transducer ring are used:
A.sub.1(0)=A.sub.0 R.sub.1 exp(?2?|z.sub.1|) (3-1)
A.sub.2(0)=A.sub.0 R.sub.2 exp(?2?|z.sub.2|) (3-2)
A.sub.1(?z)=A.sub.0(?z)R.sub.1 exp(?2?(|z.sub.1|+?z)) (3-3)
A.sub.2(?z)=A.sub.0(?z)R.sub.2 exp(?2?(|z.sub.2|??z)) (3-4)
where ?z is the distance between measurements.

(69) Then
[A.sub.1(?z)/A.sub.1(0)]/[A.sub.2(?z)/A.sub.2(0)]=exp(?4? ?z) (3-5)
from which the attenuation, ?, can be calculated.

(70) If the attenuation is found to be negligible and/or the distance between the two reflectors is small, then the attenuation value may be set to zero.

(71) In a second method, the DAC curves are fitted to two or more features with assumed equal reflectivity, normally a series of girth welds.

(72) Attenuation may also be set to a value based on experience.

(73) Clamping Effect of a Row of Transducers

(74) Two rows of transducers 5 can themselves act as clamps on the pipe and, therefore, reflect guided waves. Thus, a guided wave passing through a row of transducers can reflect part of the wave and reduce the transmission amplitude.

(75) Referring to FIG. 9, the effect that a row 4.sub.1, 4.sub.2 of transducers 5 can have is shown. As shown in FIG. 9, a guided wave can be reflected from a first row of transducers. Thus, the row of transducers can serve as a reflector and, therefore, its reflection coefficient can be determined. Notwithstanding this, the effect tends to be negligible for normal applications because reflection is usually small and, therefore, the error due to transmission loss is small. For example, FIG. 8 has a pair of peaks at about ?10 m which is the signal arising from the first reflector and one of the rows of transducers.

(76) Where the reflection from a row of transducers can be neglected, the arrangements can be used to detect and size features, such as defects, that are located in the near field (i.e. in a length of pipe either side of the transducer ring 3 in which the true amplitude of reflections is underestimated) or in the dead zone (i.e. in a length of pipe either side of the transducer ring 3 equivalent to the length of the transmitted signal where reflections cannot be obtained), similar to determining the reflection coefficient of a transducer row.

(77) Other Arrangements

(78) Different reverberation paths than the ones considered above may possibly be used to calculate the reflection coefficients. In particular, these can include paths where both reflectors are on one side of the transducer, rather than described above on opposite sides of the transducer. In tube testing, the presence of two ends with refection coefficients close to 1 leads to characteristic successive reverberations along different paths. These reverberated reflections add constructively as they undergo equal phase changes each time the signal reverberates between the ends and any features in between. Observing these reverberations instead of or in addition to the direct signal can be used to increase sensitivity to small defects. Again, an appropriate calculation using the reverberated signals can be used to calibrate and size the features of interest.

(79) Reflectors Located on the Same Side of the Transducer Ring

(80) In the examples given earlier, the features (i.e. reflectors) are located on opposite sides of the transducer ring 3 (FIG. 2). The sizing and calibration method can be used even when features are located on the same side of the transducer ring.

(81) Referring to FIG. 10a, the pipe 2 is shown having third and fourth reflectors 8.sub.3, 8.sub.4 located on the same side of the transducer ring 3. The transducer ring 3 is used to excite a guided wave which propagates forwards. The guided waves are reflected by the reflectors 8.sub.3, 8.sub.4 and the reflected waves are detected by the transducer ring 3.

(82) The transducer ring 3 receives third and fourth direct pulse-echo (i.e. single-reflection) signals reflected from the first and second reflectors 8.sub.3, 8.sub.4 respectively and having paths z: 0.fwdarw.z.sub.3.fwdarw.0 and z: 0.fwdarw.z.sub.4.fwdarw.0. The first and second direct pulse-echo signals have peak amplitude values A.sub.3 and A.sub.4 respectively.

(83) The transducer ring 3 receives a triple-reflection signal which has been reflected first by the fourth reflector 8.sub.4, then by the third reflector 8.sub.3 and then again by the fourth reflector 8.sub.4, i.e. a path z: 0.fwdarw.z.sub.4.fwdarw.z.sub.3.fwdarw.z.sub.4.fwdarw.0. In this case, the reverberated signal has amplitude A.sub.434 and is received by the transducer ring 3 from the direction of the third and fourth reflectors 8.sub.3, 8.sub.4.

(84) FIG. 10b is an amplitudedistance plot showing peaks in amplitude corresponding to single and triple reflections. The characteristic distances at which the signals appear can be used to identify reverberation echoes and distinguish them from real reflectors.

(85) The reflection coefficients R.sub.3 and R.sub.4 for third and fourth reflectors can be calculated as follows. The amplitudes A.sub.3, A.sub.4 and A.sub.434 can be expressed as:
A.sub.3=A.sub.0 R.sub.3 exp(?2?|z.sub.3|) (3-1)
A.sub.4=A.sub.0 R.sub.4(1?R.sub.3.sup.2)exp(?2?|z.sub.4|) (3-2)
A.sub.434=A.sub.0 R.sub.4.sup.2 R.sub.3(1?R.sub.3.sup.2)exp(?2?(2|z.sub.4|?|z.sub.3|) (3-3)
where is ? an attenuation factor which can be zero or a positive non-zero number (i.e. 0??) and A.sub.0 is the amplitude of the outgoing wave.

(86) Equations (3-1), (3-2) and (3-3), for example, can be re-arranged as:
R.sub.3=1/sqrt(A*+1) (3-4)
R.sub.4=exp(2?(|z.sub.4|?|z.sub.3|))sqrt(A*+1)A.sub.4/(A.sub.3A*) (3-5)
A.sub.0=exp(2?|z.sub.3|)A.sub.3 sqrt(A*+1) (3-6)
where A*=A.sub.4.sup.2/(A.sub.3A.sub.434). R.sub.3 can be determined without knowledge of attenuation. The reflection coefficient R.sub.3 (and/or R.sub.4) for the third and/or fourth reflectors can be found using equations (3-4) and (3-5). The amplitude of the outgoing wave, A.sub.0, can also be found.

(87) It will be appreciated that many modifications can be made to the embodiments hereinbefore described.

(88) Discrete or segmented piezoelectric transducers need not be used. For example, transducers based on magneto-restrictive or Lorentz force can be used. In the case of magneto-restrictive transduction, the transducer can take the form of a single strip of magneto-restrictive arranged around the circumference of the pipe. In some embodiments, for example using transducers which employ exciting waves via the Lorenz force, the transducers need not be in contact with the pipe.

(89) Higher frequency guided waves, for example up to 2 MHz, can be used.

(90) Other guided waves modes, e.g. longitudinal mode, can be used.

(91) Reflectors need not be adjacent. For example, another reflector (or several reflectors) can lie between two reflectors.

(92) Reflection coefficients (or other parameters) can be found using even-numbered reflections and odd-numbered reflections, and can be combined (e.g. by taking the average) to obtain a combined value, e.g. an average, of reflection coefficient.

(93) Reflection coefficients (or other parameters) can be found using multiple reflections without using any single reflections or using fewer single reflections. For example, instead of using single reflections from the reflectors, triple reflections may be used. These can then be used together with double (or other even-numbered) reflections to obtain reflection coefficients. Alternatively, for example, a single reflection, a triple reflection used instead of a single reflection and a double reflection can be used. This can be used when one or more single reflections cannot be observed or measured.