WO2014045007A1

WO2014045007A1 - Ultrasonic non-destructive testing of solid objects

Info

Publication number: WO2014045007A1
Application number: PCT/GB2013/051954
Authority: WO
Inventors: Frederic Bert CEGLA; Jonathan Mark Allin
Original assignee: Permasense Ltd
Priority date: 2012-09-21
Filing date: 2013-07-23
Publication date: 2014-03-27
Also published as: GB201216883D0; GB2506145A

Abstract

Ultrasonic non-destructive testing of a solid object makes use of a transducer which is removably and relatively lightly held against a solid object to be tested without the use of an intervening couplant. By exciting the transducer using a Golay or similar pulsed sequence, high performance can be achieved simultaneously with low coupling forces, low contact pressures, low voltages, low power requirements and low detection times.

Description

ULTRASONIC NON-DESTRUCTIVE TESTING OF SOLID OBJECTS

The present invention relates to ultrasonic non-destructive testing (NDT) of solid objects. It is particularly although not exclusively concerned with the testing of solid objects such as pipes and other fixed installations, The invention preferably finds particular application in environments that are hostile or where access is difficult.

Ultrasonic testing of the properties of a liquid is relatively straightforward due to the high longitudinal wave coupling that can easily be achieved between a transducer and the liquid under test. Longitudinal waves from the transducer couple into and out of the liquid with quite small losses, which means that low signal inputs can be used. When testing solid objects, however, there is inevitably a large acoustic impedance mismatch between the solid material of the transducer and the test structure itself, due to the air that is present between the asperities via which the two surfaces contact. This means that there is a very large reflection of ultrasound back to the transducer, and only a small proportion of the incident signal enters the structure to interrogate it, For this reason dry coupling (direct contact) needs a very high voltage transducer to supply a suitably large input signal and is not used in practice.

A variety of different approaches have been used to ameliorate this difficulty, most of which involve attempting to improve the coupling between the transducer and the solid object under test. The most common approach is to use some sort of coupling liquid such as mineral oil or (less commonly) a low loss rubber. These provide greater coupling efficiencies but at the expense of convenience in use. Also, since such couplings are poor at transmitting shear waves they are typically useful only in applications where the testing can be carried out using longitudinal waves.

An alternative approach to achieving improved coupling is to bond the transducer permanently to the test object, for example by means of an adhesive or soldering. While this approach is convenient for some types of long term monitoring it cannot be used in situations where it is impossible or undesirable to make permanent structural connections to the object being tested. It is also very inconvenient to have to wait for adhesive to cure or to take soldering equipment on site, since ease of installation is very often important. 54

Improved coupling can also be achieved by pressing the transducer very hard against the test object. This is normally done by clamping, and to achieve significantly higher efficiencies requires a clamping force which approaches the yield point of the surfaces which are being forced together. If a sufficiently high clamping force is used along with a lower (but still quite high) pulse, it is possible to measure a returning signal from the object without the use of any couplant. Of course, there are many situations in which high contact forces may not be desirable or even possible.

Yet another approach is to polish the respective surfaces to a mirror finish.

Regardless of the way in which the transducer is attached to the object under test, lower voltage pulses are possible by taking multiple readings and then averaging the results. Such averaging techniques typically extend the time needed for accurate measurement to minutes or more. Averaging requires significant power since during the entirety of that period all of the necessary electronics have to be active.

In the present invention a high voltage heavily clamped (or permanently secured) transducer is done away with in favour of a low voltage device which is merely pressed against the test object with a relatively low force. The transducer emits a specific coded signal which could be a random or pseudo random sequence but desirably is a Golay coded sequence of the type sometimes used for analysing liquid or semi-liquid samples. Examples of this code being used in the analysis of water, silicone oil and aerated toothpaste are described in APY Chang, R E Challis, V G Ivchenko and AN Kalashnikov, "A field programmable gate array-based ultrasonic spectrometer " Meas. Sci. Technol. 19(2008) 045802 (13pp); and R E Challis and V G Ivchenko "Sub-threshold sampling in a correlation-based ultrasonic spectrometer" Meas. Sci. Technol. 22(2011) 025902 (12pp).

Alternatively, the input signal could take any other desired form, including but not limited to a multi-frequency signal, a chirp, pseudo-white noise, or a sequence of pulses or spikes.

According to the present invention there is provided a method of ultrasonic testing of a solid object comprising applying a force to hold a transducer directly against a surface of the object, driving the transducer with a coded input signal to generate an ultrasound signal which passes into the object, and detecting an output signal comprising the reflected or transmitted ultrasound signal after passage through the object.

In some embodiments, the signal need not be coded. While the transducer is preferably operated in non-resonant mode, this is not a necessary condition. Embodiments of the invention work very well in conjunction with conventional resonant transducers that are often highly damped.

By exciting the transducer using a coded signal, high performance can be achieved simultaneously with low coupling forces, low contact pressures, low voltages, low power requirements and low detection times. The time required to take a measurement on a solid object of a given thickness, such as a pipe, may be lower than in prior systems.

The present invention will now be described, by way of example, with reference to the accompanying drawings, in which:

Figures la, lb and lc show a first embodiment of a transducer assembly suitable for use with the invention; Figures 2a, 2b and 2c show a second embodiment of a transducer assembly suitable for use with the invention;

Figures 3 a, 3 b and 3 c show a third embodiment of a transducer assembly suitable for use with the invention;

Figures 4a, 4b and 4c show a fourth embodiment of a transducer assembly suitable for use with the invention;

Figures 5a and 5b show embodiments in which a transducer assembly is pressed against a pipe to be tested;

Figures 6a and 6b show embodiments in which transducer assemblies are clamped to a pipe to be tested; P T/GB2013/051954

Figure 7 shows an experimental setup using a load of around 300N, and a graph showing the signal if processed in the conventional way; the signal on the graph shows the electric breakthrough of the send signal and some ringing of the transducers. There is no trace of the ultrasonic signal that has travelled through the test sample and no signal is received after the ringdown (after όμε).

Figure 8 shows a similar set up to Figure 7 but with a load of around 50N, and a graph showing the signal when processed using a Golay code; the signal now shows the initial breakthrough but it also shows a series of consecutive echoes, reflections within the test sample walls. The temporal separation between consecutive echoes widens as the test sample thickness is increased. This clearly shows that the echoes are ultrasonic reverberations within the test sample that were not detectable with the conventional way of testing. Figure 9 shows an experimental setup for testing the transmission through a dry coupled contact under varying loads (the sending transducer is dry coupled and the receiving transducer is bonded);

Figure 10 shows the transmitted signal for the transmission of shear waves using the set up of Figure 9; and

Figure 1 1 shows how the transmitted signal amplitude in Figure 9 varies with compressive load. As will be described in more detail below, embodiments of the present invention comprise a transducer which is brought up to the surface of the solid object to be tested and is removably pressed against it, without the need for any intermediate coupling material between the transducer and the surface. No specific surface preparation need be carried out apart from any necessary cleaning and (if the application so requires) removal of surface corrosion and the like. Although the transducer should be able to sit on a solid and clean part of the object to be tested, no polishing for example is needed. The surface, for this purpose, may be the surface of a paint or coating layer. 13 051954

Turning to Figure 5, the transducer or transducers 10 to be used in the testing of a solid body such as a pipe 20 are held against the surface by a force 30. The force could be provided in any convenient manner including (but not limited to) a magnetic or electrical coupling, gravity, or even by manual pressure. Figure 6 shows an alternative arrangement in which the force 30 is provided by a jubilee band 40 which is removably fitted around the pipe 20. More generally, any type of clamp or clamping means may be used removably to attach the transducer to the object under test. Either the object under test or the clamp may act as a ground. Alternatively, the device may be ground internally via a coaxial cable.

Various exemplary embodiments of the transducer 10 are illustrated in Figures 1 to 4. In each case, (a) represents an isometric view, (b) a section from the side and (c) a section from the front. Other embodiments are not excluded, and it may be noted that the transducer may include separate transmitting and receiving crystal elements, or a common element which is used for both purposes. It would also be possible for one transducer to have a transmitting element and for another, connected elsewhere on the object, to have a receiving element.

As shown in Figure 1 , the transducer has an external body 100 of any suitable structural material, such as for example a hard plastics material. Within the body 100 is a magnet 102 which will in use releasably couple the transducer onto the surface of a pipe (e.g. a metal pipe) or other solid object (not shown). The strength of the magnet may be chosen according to the desired force with which the transducer is to be urged against the object.

Below the magnet are a pair of soft spacers 104 under which there are live electrodes 110, The live electrode is attached to the upper part of a piezo-electric transducer 108 (for example 0.5mm thick), below which there is a ground electrode 106. On the open side of each of the transducers there is a soft spacer pad 112 that shields the active elements from the outside environment.

In use, the transducer is firmly held against a metallic object to be tested by the magnet 102. The resultant force pushes the ground electrodes 106, piezo-electric transducers 108 and live electrodes 110 up against the soft spacers 104 until the pads 112 come to rest against the object surface. In this configuration, the magnet is a known distance away from the surface o the object, and hence applies a known, constant force which holds the transducer in place. The spacers 104 limit transmission of ultrasound into the magnet and any resultant ringing.

Figure 2 shows an alternative arrangement that dispenses with the pads 112 and where only one single transducer is used. In this embodiment, the transducer simply sits on the ground electrode 106.

A further embodiment is shown in Figure 3. Here, the upper part of the body 100 is formed of the magnet 102. The lower part 114 of the body may (but need not) comprise further magnets. Elastic movement is provided by a spring 116 which extends between the magnet 102 and a structural base plate 118 to which the remaining components are mounted. Below the base plate are two soft spacers (104) and the electrodes and piezo-electric device as in Figure 1. Soft spacer pads 112 are also provided. Finally, Figure 4 shows an embodiment having a body similar to that of Figure 3, but with a spring mounted assembly, the component parts of which are similar to those illustrated in Figure 2, and are identified by the same reference numerals. The difference between this embodiment and that of Figure 3 is that only a single PZT transducer is used here. In each of the embodiments suitable electronics (not shown) cause the piezo-electric transmitter to emit a sequence of high frequency ultrasound signals (typically above lMHz) which couple through the adjacent surface into the object under test. Signals which are reflected from or transmitted through the object are received by the same or by another piezoelectric device, and the received signals are analysed to determine the object's physical properties. These might include wall thickness, internal structure, internal composition, occlusions, cracks and the like.

The transmitter generates coded (e.g. a random or pseudo-random) sequences as input signal. In the preferred embodiment the code is a Golay code, although other types of error correcting or non-error correcting sequences could be used instead. Alternatively, the coded input signal could take any other desired form, including but not limited to a multi-frequency signal, pseudo-white noise, or a sequence of pulses or spikes. The code is preferably a binary code, either time or frequency-based. In preferred embodiments the signal is determined by taking the cross correlation between the input signals and the reflected or transmitted signals. Cross-correlation may be performed in any known manner. Alternatively time-averaging of the output signal may be carried out.

More generally, the transducer is driven by a known digital (i.e. coded) multi-frequency input signal, with the frequency or frequencies optionally varying with time. The frequencies may extend over a range, on either side of a central frequency, with the signal taking discrete frequency values.

The returning output ultrasound signal is received, converted into an electrical signal, and analysed both in the temporal and in the frequency domains, for example to determine the impulse response. Finally a test result is determined on the basis of the known input signal and the analysed output signal.

The magnets used in the embodiments of Figures 1 to 4 (or any other means of urging the device against the object's surface) will typically apply forces which are very much lower than has been the case in the prior art. With the devices described there is no longer any need to generate high forces to achieve sensitivity and accuracy in measurement. Typical forces of a device according to the present invention might for example be 500, 400, 300, 200, or 100 N. Indeed, in some models the forces might be as low as 90, 80, 70, 60, 50, 40, 30, 20 or 10 N. The preferable force ranges applicable to the present invention include all ranges defined at one end by any one of the previously mentioned figures, and the other by any other of the previously mentioned figures.

These forces are in respect of a pair of coupled piezo-electric crystals of length 13mm, width 3mm and thickness 0.5mm being urged onto the test surface. It will naturally be understood that proportionally different force ranges will apply when the area of the crystal lying adjacent to the surface is, proportionally, greater or less than that specified. The resulting coupling pressures for the contact forces of 500, 400, 300, 200, 100, 90, 80, 70, 60, 50, 40, 30, 20 and 10 N thus are 6.41, 5.13, 3.85, 2.56, 1.28, 1.15, 1.03, 0.9, 0.77, 0.64, 0.51, 0.38, 0.26 and 0.13 MPa for an arrangement consisting of two piezo-electric crystals, pressures on a single crystal will be increased by a factor of 2. T B2013/051954

The transmission frequency will typically be around 2MHz but could be 0.1 , 0.2, 0.25, 0.3, 0.4, 0.5, 0.6, 0.7, 0.75, 0.8, 0.9, 1, 2, 3, 4, 5 or 10MHz. The amplitude of the voltage spikes used in the present invention will typically be lower than 100 volts, with preferable values 90, 80, 70, 60, 50, 40, 30, 20, 1 , 9, 8, 7, 6, 5, 4, 3, 2 or 1 volt. Preferable ranges include those defined at one end by any figure from this list and at the other by any other figure from the list. The input signal may be of longer duration than 5, 10, 20, 50, 100, 200, 500, 750 or 1000 cycles of a central signal frequency,

In contrast with most non-permanently attached conventional devices, the present invention allows the transmission of not only longitudinal but also of shear waves. This provides an extra level of flexibility to the user, depending on the requirements of the particular application.

The size and thickness of the piezo-electric crystals both for transmission and reception can be chosen (by simple experimentation if need be) according to the task in hand. Typically prior art transducers need to be thin (so that they are non resonant, or otherwise significant damping is needed), By using Golay sequences and processing the resulting signal can be shorter in the time domain and this allows the use of thicker, more robust crystals, allowing devices to be made stronger and less fragile. We now turn to some experimental examples illustrating the performance of the system in practice.

Figure 7 shows an experimental set up in which paired piezo-electric crystals 210 provide an ultrasonic input into an underlying metal plate 220. The crystals are evenly loaded by weights (not shown) which bear down on a soft backing material 200.

The figure shows the signal (300 averages) that was recorded when sending a 5 cycle toneburst with 2MHz centre frequency on one dry coupled piezo-electric shear crystal and receiving on an adjacent dry coupled piezo-electric shear crystal (LI 3mm, W3mm, T0.5mm) which were pushed onto a 10mm thick steel plate with a load of -300N. With such a set up, conventional methods do not work at all as the signals are too weak - the signal seen is breakthrough from the excitation and ringing from the initial transducer response. There is no evidence of any ultrasonic signals.

Figure 8 shows the signals that were recorded using Golay coded sequences by sending on one dry coupled piezo-electric shear crystals (LI 3mm, W3mm, TO.5mm) and receiving on an adjacent one. The force that was employed to push them on the 6 and 10mm thick plates was ~50N. Here following the initial sent signal a sequence of echoes is clearly noticeable. The separation of the echoes is larger on a 10mm thick plate compared to a 6mm thick plate, as would be expected due to the longer transit time between plate surfaces in the thicker plate. The ultrasonic signals are therefore clearly visible in this new method, while the conventional method failed.

Figure 9 shows an alternative set up for measuring transmission signals. Instead of the signals being received by reflection, as in the previous set ups, they are in this arrangement received by a separate receiving transducer 230 which is positioned beneath the plate 220.

Here, the transmission through the dry coupled joint as a function of load was investigated. In order to increase the signal amplitude and use only one dry coupled contact the receiving transducer (a commercial shear transducer Panametrics VI 54 2MHz) was bonded to the bottom surface of a 10mm thick plate. A shear crystal was placed on top of the plate and pushed onto the plate by various loads. The corresponding transmitted signals were received and amplitude and signal to noise ratio as a function of applied load were deduced. Figure 10 shows the results of the experiment of Figure 9, based upon averages of ten samples.

Figure 1 1 shows how the amplitude of the shear signal varies as a function of compressive load applied to the shear crystal. The change in signal to noise ratio is also shown. Once again, this graph is based on averaging 10 samples per measurement.

Conventional ultrasonic transducers employ damped resonant transducers. Despite all efforts to maximise the damping this usually leads to pulse lengths of several cycles in the time domain. 5 Cycles is not uncommon. This limits the resolution of thickness gauges as the echoes need to be temporally resolved, so that the minimum plate thickness that is measured with a 1MHz 5 cycle toneburst is a thickness of 7.5mm (5 wavelength separation between adjacent echoes, 5*3mm (shear wavelength at lMHz) = 15mm 2 = 7.5mm. (One has to divide by two because the wave transits the material twice, there and back.)

The use of Golay sequences results in much shorter pulse lengths (after the processing step) which in turn allows for lower frequencies to be employed. This is advantageous if one wants to penetrate into highly attenuative materials or limit the effect of roughness on the variability in the ultrasonic measurement.

A particularly significant advantage becomes evident when the transducer is to be permanently installed, for example for structural health monitoring applications. In such applications dry coupling avoids the need for bonding which is often difficult to do on site, and also leads to issues with adhesive degradation, particularly at elevated temperatures.

Claims

1. A method of ultrasonic testing of a solid object comprising applying a force to hold a transducer directly against a surface of the object, driving the transducer with a coded input signal to generate an ultrasound signal which passes into the object, and detecting an output signal comprising the reflected or transmitted ultrasound signal after passage through the object.

2. A method as claimed in claim 1 in which the input signal is a sequence of spikes or pulses.

3. A method as claimed in claim 1 or claim 2 in which the input signal is a random or pseudo-random coded sequence.

4. A method as claimed in claim 1 or claim 2 in which the input signal is a Golay-coded sequence.

5. A method as claimed in any one of the preceding claims including the step of calculating a cross-correlation between the input and output signals.

6. A method as claimed in any one of the preceding claims in which the transducer is driven by voltage pulses of less than 100, 90, 80, 70, 60, 50, 40, 30, 20, 10, 9, 8, 7, 6, 5, 4, 3, 2 or 1 volt,

7. A method as claimed in any of the preceding claims in which the ultrasound signal comprises ultrasonic shear waves.

8. A method as claimed in any one of the preceding claims in which the force is less than 500, 400, 300, 200, 100, 90, 80, 70, 60, 50, 40, 30, 20 or 10 N.

9. A method as claimed in any one of the preceding claims in which the force is provided by a magnet which forms part of a transducer assembly.

10. A method as claimed in any one of the preceding claims in which the force is provided by a clamp.

11. A method as claimed in any one of the preceding claims in which the surface of the object is a paint or a coating layer which overlies a body of the object.

12. A method as claimed in any one of the preceding claims in which the solid object acts as an electrical ground for the transducer.

13. A method as claimed in claim 10 in which the clamp includes or acts as an electrical ground for the transducer.

14. A method as claimed in any one of the preceding claims in which the transducer is driven non-resonantly.

15. A method as claimed in any one of the preceding claims in which the pressure experienced as a result of the force is less than 6.41, 5.13, 3.85, 2.56, 1.28, 1.15, 1.03, 0.9, 0.77, 0.64, 0.51, 0.38, 0.26 or 0.13 MPa.

16. A method as claimed in any one of the preceding claims in which the input signal is of longer duration than 5, 10, 20, 50, 100, 200, 500, 750 or 1000 cycles of a central signal frequency.

17. A method as claimed in any one of the preceding claims in which the signal is a chirp.