WO2013017880A1

WO2013017880A1 - Wear measurement

Info

Publication number: WO2013017880A1
Application number: PCT/GB2012/051867
Authority: WO
Inventors: Philip Brian HARPER
Original assignee: Tribosonics Ltd
Priority date: 2011-08-02
Filing date: 2012-08-01
Publication date: 2013-02-07
Also published as: GB201113289D0; GB2493367A

Abstract

A system (10) for measuring wear, the system (10) comprising: a surface boundary (22) comprising a feature (26) having a non-planar impedance discontinuity for indicating wear; a transmitter (30) arranged to emit a wave toward the boundary surface (22) so that the wave (5) comprises a characteristic representative of the feature (26); a receiver (30) for receiving the wave from the boundary surface (22); and a wear determining module (48) configured to measure the characteristic and to determine wear on the boundary surface (22) from the characteristic.

Description

Wear Measurement

Technical Field:

The invention relates to wear measurement, and in particular, but not exclusively, the invention relates to wear measurement using ultrasonic waves. The invention finds particular but not exclusive use in the field of wear measurement in sliding contacts, abrasive contacts and pivoting contacts, such as those found in engines, bearings and flow applications.

Background:

Physical measurement techniques have been used to measure wear. For example in the case of an internal combustion engine liner the internal diameter of the liner/bore can be measured using specially made tools or using non-contact methods such as capacitance or laser reflectometry.

The main disadvantages of this technique are:

• A relatively inaccurate measurement. Piston bore measurements have been known to be ±50 μιη. · An operator has to physically measure the component requiring safe access to the component. This can require shutting down the equipment and partial or complete disassembly which can be difficult to schedule and expensive to implement.

• It is not an on-line measurement.

Time of flight (ToF) ultrasonic measurement techniques rely on measuring the time taken for an ultrasonic wave to travel from an ultrasonic emitter/sensor to a wearing surface and back again. Wear occurring at the surface will result in a reduction in the time of flight. Very accurate timing circuitry is required in order to accurately measure small amounts of wear. Wear of 1 μιη requires a timing resolution of the order of 0.34E-9 seconds. For single shot measurements at single points this requires a digitiser sampling rate of 3 Giga-samples per second (Gsps) or better. This can be expensive to implement from a hardware perspective.

The main problem with this technique is that most materials expand with temperature. This means that the transit time changes as the temperature changes thus affecting accuracy. This introduction of error into the measurement means that the technique is not reliable for accurate measurements where there is any significant change in temperature (for example in an internal combustion engine). This fact is demonstrated by the more or less complete lack of sensors on the market using this technique for detecting wear. The second main problem is the requirement to use very accurate, very high speed electronics to perform the timing measurements. This can be costly and difficult to implement. In addition, temperature measurements need to be made in parallel in order to try and take account of the effect of temperature. The ToF technique relies on having steps or features or thicknesses significantly greater than the wavelength of the ultrasonic wave. Normally a minimum of five wavelengths is required to accurately determine the ToF between two pulses in the time domain.

Electrical loop measurement techniques rely on loops of very fine wire placed very accurately in an insulating material that is mounted in the wearing surface. The fine wire loops are successively removed from the surface to detect set amounts of wear. For example several wire loops could be placed in 5 μιη steps from the surface. As wear occurs the loops get broken and wear can be measured by detecting whether or not an electrical current can be passed through the loop. By detecting which loops can pass a current and which cannot the wear on the surface can be measured. It is an aim of the invention to provide an improved way of measuring wear. The invention also aims to simplify the measurement hardware by reducing the requirement for very high accuracy, high resolution electronics. More particularly, the invention aims to provide more accurate and more reliable wear measurement technology.

Summary of the Invention According to the invention, there is provided an apparatus and method as set forth in the attached claims. Further optional features are seen in the dependent claims and the description which follows.

A method of measuring wear, the method comprising: emitting a wave toward a boundary surface, the boundary surface comprising a feature having a non-planar impedance discontinuity for indicating wear, the feature being arranged to induce a characteristic into the wave so that the wave contains the characteristic; receiving the wave, measuring the characteristic from the wave and determining wear on the component from the characteristic. An apparatus for measuring wear, the apparatus comprising: a receiver for receiving a wave from a boundary surface comprising a feature having a non-planar impedance discontinuity for indicating wear, the feature being arranged to induce a characteristic into the wave; a wear determining module configured to measure the characteristic and to determine wear on the boundary surface from the characteristic.

A system for measuring wear, the system comprising: a surface boundary comprising a feature having a non-planar impedance discontinuity for indicating wear; a transmitter arranged to emit a wave toward the boundary surface so that the wave comprises a characteristic representative of the feature; a receiver for receiving the wave from the boundary surface; and a wear determining module configured to measure the characteristic and to determine wear on the boundary surface from the characteristic.

In this way, an improved way of measuring wear is provided. The measurement hardware is simplified by reducing the requirement for very high accuracy, high resolution electronics. More particularly, more accurate and more reliable wear measurement technology is provided. The non-planar impedance discontinuity may have dimensions smaller than a wavelength of the ultrasonic wave.

In addition to the features above, the following optional features may form part of the invention:

Feature shape · wherein the impedance discontinuity is an acoustic impedance discontinuity. wherein the feature comprises at least two impedance discontinuities arrang interfere with the wave.

• wherein the at least two acoustic impedance discontinuities are displaced from each other with respect to the wave by one or more of a greater than, equal to and sub- wavelength dimension.

• wherein the feature comprises two materials having different acoustic properties.

• wherein the feature comprises two materials having different acoustic properties and the feature is flush with the boundary surface.

• wherein the feature is a non-planar surface.

• wherein the feature comprises a step. • wherein the feature repeats so that there are more than two acoustic impedance discontinuities.

• wherein the feature is formed by one or more of the following techniques: eroding, engraving, micromachining, etching, stamping, casting, cutting, carving and laser ablation.

• wherein the feature is a surface profile.

• wherein the feature is a repeating pattern.

• wherein the feature comprises one or more of peaks and troughs, parallel walls between the peaks and troughs, angled walls between the peaks and troughs, pointed peaks, flat peaks, angled troughs, flat troughs crenelations, concave or convex features.

• wherein the feature includes a combination of two or more profiles.

• wherein the feature comprises a plurality of features, and preferably the plurality of features is arranged in a two-dimensional array, and preferably wherein each of the plurality of features has a three-dimensional profile. Preferably, the three-dimensional profile is square. Preferably, the three-dimensional profile is hemispherical, and preferably each hemisphere has a dimpled surface.

• wherein the feature is formed by one or more subsurface voids or inclusions.

Feature size

• wherein at least one dimension of the feature is less than the wavelength of the wave.

• wherein the width of a sub-feature of the feature is less than the wavelength of the wave.

• wherein the width of a sub-feature of the feature is less than 200 microns, or less than 100 microns.

Determining Wear

• wherein wear is determined by time domain analysis of the reflected wave.

• wherein the time domain analysis measures an amplitude, and preferably wherein the time domain analysis measures an amplitude relative to another amplitude within the reflected wave. • wherein the time domain analysis measures a time position of one or more points in the reflected wave, and preferably wherein the time domain analysis measures a time position of one or more points in the reflected wave relative to another point in the reflected wave.

• wherein wear is determined by spectral analysis of the reflected wave, preferably using a Fast Fourier Transform (FFT).

• wherein the spectral analysis measures an amplitude at a specific frequency in the amplitude- frequency spectrum, preferably wherein the spectral analysis measures a first amplitude at a first frequency and a second amplitude at a second frequency in the amplitude- frequency spectrum and compares the first amplitude and the second amplitude.

• wherein the spectral analysis measures a frequency value at which the amplitude dips in the amplitude-frequency spectrum. Preferably, wherein the spectral analysis measures a first frequency value at which the amplitude dips with a second frequency at which the amplitude also dips in the amplitude-frequency spectrum and compares the first frequency and the second frequency.

• wherein the spectral analysis measures a phase value in the phase-frequency spectrum. Preferably, wherein the spectral analysis measures a phase value relative to another phase value in the phase-frequency spectrum and compares the two phase values. Preferably, wherein the spectral analysis measures a first phase at a first frequency and a second phase at a second frequency in the phase-frequency spectrum and compares the first phase and the second phase.

• wherein the spectral analysis measures a frequency value at which the phase dips in the phase- frequency spectrum. Preferably, wherein the spectral analysis measures a first frequency value at which the phase dips with a second frequency at which the phase also dips in the amplitude-frequency spectrum and compares the first phase and the second phase.

• wherein wear is determined by removing phase information from the reflected wave to create an idealised reflected wave. Preferably, wherein the phase information is removed by performing a Fast Fourier Transform (FFT) on the reflected wave, discarding the phase information and performing an inverse FFT on the amplitude- frequency information. Preferably, wherein wear is determined by measuring the arc length of idealised reflected wave. Preferably, wherein the arc length is compared with a reference arc length.

Boundary Surface

• wherein the boundary surface is an outer surface of a component.

• wherein the component is in one of a sliding contact arrangement, an abrasive contact arrangement or a pivoting contact arrangement.

• wherein the boundary surface is coated, preferably wherein the boundary surface is coated with a layer of between around 0.01 microns to 100 microns. Preferably, wherein the boundary surface is coated with a ceramic material.

Wave

• wherein the wave is a mechanical wave.

• wherein the wave is an ultrasonic wave.

• wherein the wave is an ultrasonic wave with a frequency in the range of 0.2 to 50 MHz, preferably around 10 MHz.

• wherein the wave is a sinusoidal wave.

Use of a feature having a non-planar impedance discontinuity in the path of a wave arranged to induce a characteristic into the wave for the purpose of wear measurement.

Here, we define wear measurement as the measurement of how much material has been removed from a surface.

Brief Description of the Drawings

For a better understanding of the invention, and to show how example embodiments may be carried into effect, reference will now be made to the accompanying drawings in which:

Fig.l is an schematic overview of a wear measurement system in accordance with the invention;

Fig.2 is a perspective view of an example component with example feature on a top surface; Fig 3 is a side view of the example component of Fig.2;

Fig.4 is a graph plotting amplitude versus time of a captured waveform reflected from a wearing surface of a component; Fig.5 is a graph showing a close up of the waveform 'B' as shown in Fig.4;

Figs.6-8 are graphs plotting amplitude versus time of a modelled reflected waveform from a boundary surface of the component of Fig.2 with 0 μιη (0%), 500 μιη (50%) and 1000 um (100%)) of wear respectively; Figs.9- 11 are graphs plotting FFT amplitude versus frequency of the modelled reflected waveforms of Figs.6-8 respectively (i.e. 0 μιη (0%>), 500 μιη (50%>) and 1000 μιη (100%) of wear);

Fig.12 is a graph plotting frequencies of dips (visible in Figs. 9 and 10) against wear rate for three samples that were tested; Fig.13 is a graph plotting frequencies of dips against wear rate comparing the results of Fig.12 with a computational simulation of the test;

Fig.14 is a graph plotting frequencies of dips against wear rate comparing the results of Fig.12 with a lubricated test;

Figs.15-19 are graphs plotting modelled frequencies of dips against wear rate for a boundary surface similar to that of Fig.2 having the following gap and tooth widths: 600 μιη, 500 μιη, 400 μιη, 300 μιη, and 200 μιη;

Figs.20-23 are graphs plotting a modelled amplitude versus time response for a boundary surface similar to that of Fig.2 having a 100 μιη gap width and 100 μιη tooth width and a tooth height of 1000 μιη that has experienced 200 μιη, 600 μιη, 800 μιη, and 900 μιη of wear respectively;

Fig.24 is a graph plotting superimposed modelled amplitude versus time responses for a boundary surface similar to that of Fig.2 having a 100 μιη gap width and 100 μιη tooth width and a tooth height of 1000 μιη that has experienced 100 steps of 10 μιη of wear, and the graph shows the waveform envelope of the surface as wear goes from none to complete wear; Fig.25 is a graph plotting the superimposed FFT amplitudes of the waveforms shown in Fig.24;

Fig.26 is a graph plotting the superimposed FFT phases of selected waveforms shown in Fig.24 (10 steps of 100 μιη wear);

Fig.27 is a graph plotting an idealised reference pulse obtained by FFT of a modelled pulse of the arrangement of Fig.24 when there is no wear on the boundary surface; Fig.28 is a graph plotting an arc length of the idealised reference pulse of Fig.27 versus wear;

Fig.29 is a graph plotting an arc length of the reflected wave including phase information versus wear;

Fig.30 is a graph plotting the ratio of arc length variation calculated by dividing the measured or modelled arc length by the reference arc length, versus wear;

Fig.31 is a schematic diagram of a sound wave reflecting from a stepped boundary surface;

Figs.32, 34, 36, 38 are graphs plotting amplitude versus time of a modelled waveform reflected from a 1.5 mm, 1 mm, 0.8 mm, 0.6 mm step respectively;

Figs.33, 35,37, 39 are graphs plotting FFT amplitude of the waveforms of Figs.32, 34,36, 38 respectively;

The following graphs refer to the testing done with a triangular tooth- shaped pattern in the surface.

Fig.40 is a graph plotting amplitude versus time for reflected waves showing the effect of wear from 100 to 200 μιη on a component; Fig.41 is a graph plotting amplitude versus time for reflected waves showing the effect of wear from 400 to 500 μιη on a component;

Fig.42 is a graph plotting amplitudes versus time for reflected waves showing three positions where there are various degrees of spread of the zero crossing across wear depths of 420 to 500 μιη; Fig.43 is a graph plotting FFT amplitudes versus frequency of the pulses of Fig.40 showing the effect of wear from 100 to 200 μιη, plotted in 10 μιη wear intervals;

Fig.44 is a graph plotting a ratio of amplitude at 10.2 MHz divided by amplitude at 6.5 MHz for wear in the range 450 to 500 μιη;

Fig.45 is a graph plotting FFT amplitudes versus frequency of the pulses of Fig.40 showing the effect of wear from 200 to 300 μιη, plotted in 10 μιη wear intervals;

Fig.46 is a close up of Fig.45 corresponding to the encircled part;

Figs.47 and 48 repeat the graphs of Figs.45 and 46 but for wear going from 400 to 500 μιη and showing in close up the minimum amplitudes of dip, which occur at around 6 MHz; Fig.49 is a plot of wear with frequency of the minimum dip in the frequency spectrum for the FFTs of Fig.48;

Fig.50 is a flow chart showing a method of performing wear measurement; and

Fig.51 is a flow chart showing a method of determining wear. Detailed Description of the Example Embodiments

Fig.l is an schematic overview of a wear measurement system 10 in accordance with one embodiment illustrating the invention.

The system 10 comprises a structure 20 which is subject to wear measurement, an ultrasonic sensor 30, and processing circuitry 40. The structure 20 is a steel cylinder test sample which represents a part or component normally subject to wear, for example a component in an engine or bearing. The structure 20 has a diameter of 12 mm and a length of 30 mm. The structure is 316 stainless steel.

The ultrasonic sensor 30 is a commercially available piezo ceramic disk (0.2 mm thick with a 7 mm diameter) sensor and is arranged both to transmit and receive ultrasonic waves. In this example, the ultrasonic sensor 30 is arranged to transmit an ultrasonic wave centred at 10 MHz. In the steel cylinder 20, a 10 MHz sound wave travels at approximately 5880 m/s, and has a wavelength of approximately 0.588 mm. The ultrasonic sensor 30 is also arranged to receive a reflection of the wave. The sensor is bonded onto the test sample using a cyanoacrylate adhesive. A coaxial cable is used to connect the sensor to the ultrasonic hardware. The sensor is pulsed with a negative square wave of approximately 20V.

The processing circuitry 40 comprises an input/output module 42, a digitizer 44, a processor and memory 46, and a wear determining module 48. The input/output module 42 is connected to the ultrasonic sensor 30 and is arranged to provide and receive signals to and from the ultrasonic sensor 30. In this example embodiment, the processing circuitry 40 is an FMS-100 ultrasonic system manufactured by Tribosonics Ltd. The FMS-100 system 40 comprises a PCI based 8 channel ultrasonic pulser receiver 42 and 100 MSPS digitizer 44. National Instruments Lab VIEW (RTM) software is used as the wear determining module 48 to control the hardware for pulsing, receiving, digitising and signal processing. The structure 20 has a wear surface 22. The ultrasonic sensor 30 is mounted on the structure 20 at a sensing end 24 which is opposite the wear surface 22. The ultrasonic sensor 30 is arranged to transmit a wave toward the wear surface 22, and to receive a reflected wave from the wear surface 22.

Fig.2 is a perspective view of the example component or structure 20 in more detail. Fig 3 is a side view of the example component or structure of 20 of Fig.2. The wear surface or boundary surface 22 comprises a feature 26 having a plurality of acoustic impedance discontinuities 26a, 26b for indicating wear, feature 26 being arranged to induce a characteristic into the wave.

The wear determining module is configured to receive a waveform indicative of the wave reflected from the feature 26 and to measure the characteristic and to determine wear on the component 20 from the characteristic.

Here, the feature 26 is formed by creating parallel channels 26a across the wear surface 20, the channels being defined by walls 26b. The feature 26, when shown in profile as in Fig.3, is crenelated, and the channels 26a and walls 26b are 0.5 mm wide. The channels are 1 mm deep. Recalling that the frequency of the ultrasonic sensor 30 is 10 MHz resulting in a wavelength of 0.588 mm, the widths of the channels 26a and walls 26b are sub-wavelength. Other methods using ultrasonic waves require features that are substantially larger than a wavelength and are normally four or five wavelengths.

Referring back to Fig.1, the system 10 is configured so that the processing circuitry 40 sends a pulse to the ultrasonic sensor 30 via the input/output module 42. The ultrasonic sensor 30 then creates an ultrasonic wave. The processing circuitry 40 then waits to receive a pulse back from the ultrasonic sensor 30 via the input/output module 42. The received pulse is amplified by an amplifier (not shown) and passed to a digitizer (not shown) for signal processing.

Fig.4 is a graph plotting amplitude versus time of a captured waveform reflected from the wearing surface 22 of component 20. Waveform part 'A' shows an amplifier recovery artefact associated with exciting the sensor 30 and associated swamping of the amplifier.

Waveform part 'B' shows the reflected wave from the wearing surface 22. It is the 'B' part of the waveform that is of interest in detecting the amount of wear on the wearing surface 22.

However, it should be noted that the waveform in Fig.4 was captured from a component 20 and wearing surface 22 having a profile different to that previously described with reference to Figs.2 and 3. Here, for this illustrative measurement, the profile of the feature on the wearing surface 22 is triangular (18 equilateral triangles, with a height of 0.6 mm) and the diameter of the component 20 is 12.7 mm. The feature was created using a precision wire EDM process.

Fig.5 is a graph showing a close up of the measured reflected wave shown in Fig.4. Now, for the wearing surface described with reference to Figs.2 and 3, modelled results are shown and described.

Figs.6-8 are graphs plotting amplitude versus time of a modelled reflected waveform from a boundary surface of the component 22 of Figs.2 and 3 with 0 μιη, 500 μιη and 1000 μιη of wear respectively. Here, wear has been simulated by reducing the depth of the channels 26a. As can be seen, the shape of the reflected wave (part 'B' of the waveform) changes as the channels 26a become shallower. In particular, more energy is shifted earlier in the response as wear progresses.

Figs.9- 11 are graphs plotting FFT amplitude versus frequency of the modelled reflected waveforms of Figs.6-8 respectively (i.e. 0 μιη, 500 μιη and 1000 μιη of wear). In Figs.9 and 10, dips in the frequency spectrum are clearly visible and the number of dips reduces and the amplitude of the remaining dips increases as wear progresses from 0 μιη to 500 μιη. When fully worn as shown in Fig.11 (i.e. the feature 26 no longer exists on the wear surface 22) there are no clearly visible dips in the frequency spectrum.

Fig.12 is a graph plotting frequencies of dips (visible in Figs. 9 and 10) against wear rate for three samples that were tested. The samples tested were those shown in Figs.2 and 3. Fig.13 is a graph plotting frequencies of dips against wear rate comparing the results of Fig.12 with computational simulations of the three tests. There is a reasonably good correlation between the measured data and simulated data.

Fig.14 is a graph plotting frequencies of dips against wear rate comparing the results of Fig.12 with a lubricated test. Here, the lubricated test means that the component is lubricated to simulate a component which is lubricated during use.

As can be seen from Figs.12-14, the frequencies of dips move as the wear increases. For example, a dip in the FFT amplitude occurring at just below 10 MHz with no wear increases in frequency to 13 MHz after around 400 μιη of wear. Other dips occur only after a certain amount of wear and last only for a certain amount of further wear. These graphs show that the techniques work in lubricated and unlubricated, gaseous or liquid, contaminated or clean surfaces. This is a significant improvement on other techniques that can be distorted by effects on or beyond the surface. There is strong correlation between the simulated data and the test data indicating that features can be designed and modelled with high certainty that they will operate as expected in the real world.

Figs.15- 19 are graphs plotting modelled frequencies of dips against wear rate for a boundary surface similar to that of Fig.2 having the following gap and tooth widths: 600 μιη, 500 μιη, 400 μιη, 300 μιη, and 200 μιη, respectively.

The graphs of Figs.15-19 demonstrate that each configuration of feature 26 on the wearing surface 22 produces a different dip characteristic, which can be used to identify the amount of wear on a component 20 in use. Different tooth geometries can be used to recognise wear, at or below sub-wavelength scale.

However, for the feature 26 shown in Figs.2 and 3 having a gap width of 100 μιη and a tooth width of 100 μιη, there are no significant dips in the FFT amplitude- frequency spectrum. This indicates that it may not be possible to use the above-described dip-location technique to detect wear in a surface having feature dimensions of around 100 μιη, at least with the ultrasonic sensor 30 described herein. Therefore, another method of detecting wear measurement would be useful.

The time-domain response of the reflected wave was measured for different wear increments and is shown in Figs.20-23.

In particular, Figs.20-23 are graphs plotting a modelled amplitude versus time response for a boundary surface similar to that of Fig.2 having a 100 μιη gap width and 100 μιη tooth width and a tooth height of 1000 μιη that has experienced 200 μιη, 600 μιη, 800 μιη, and 900 μιη of wear respectively.

As can be seen, the energy in the reflected wave moves towards the start of the waveform as wear increases. Fig.24 is a graph plotting superimposed modelled amplitude versus time responses for the feature 26 of Fig.2 having a 100 μιη gap width and 100 μιη tooth width and a tooth height of 1000 μιη that has experienced 100 steps of 10 μιη of wear. Fig.24 shows the waveform envelope of the surface as wear goes from none to complete wear.

However, as shown in Fig.25 (a graph plotting the superimposed FFT amplitudes of the waveforms shown in Fig.24), when the reflected waves are Fast Fourier Transformed there is virtually no variation in the amplitude with frequency. Fig.26 is a graph plotting the superimposed FFT phases of selected waveforms shown in Fig.24 (10 steps of 100 μιη wear). As can be seen, the phase- frequency response changes as wear increases. In particular the unwrapped phase in radians increases negatively as wear increases and is more pronounced at higher frequencies (i.e. over 10 MHz). In real world testing, however, temperature effects can cause significant problems for a phase-based measurement due to expansion of the material with temperature. An increase or decrease in the temperature will result in a change in the phase of the reflected pulse. In order to get around this problem a method of calculating the change in the signal is proposed. As the amplitude with frequency stays constant over the range of wear values as shown in Fig.25, it is possible to construct a theoretical reflected wave from a fully worn surface, for example from a measurement pulse. The theoretical wave is to remain constant with the amount of wear on the component 20.

The theoretical reflected wave is constructed by performing a FFT on the measured reflected wave from the ultrasonic sensor 30, discarding the phase information and performing an inverse FFT on the amplitude against frequency information. This yields an idealised time domain response that is purely a function of the FFT amplitude against frequency information without any phase information.

Fig.27 shows the idealised reference pulse.

While we now have an idealised reference pulse, we cannot yet calculate the amount of wear from the surface 20. However, we now have a reference pulse and a reflected pulse which can be compared with each other.

It can be seen from Fig.24 that there is variation in the time domain that can be described as a shortening of the length of the reflected wave, or measured pulse, from the surface with increasing amounts of wear. Hence by measuring the arc length (the total distance that a particle would have to travel if it took the trajectory shown in one of the waveforms) it is possible to get a measurement of the 'shortening' evident in Figs.20-24. However, it will be seen that if the amplitude of the reflected pulse varies then the arc length will similarly vary. With changes in temperature the output of the ultrasonic sensor 30 can change, and that will result in changes in the measured arc length of the measured pulse. This is where the arc length of the reference pulse is important; it tells us how much the sensor has changed with temperature. If the arc length of the pulse is computed for a sensor at constant temperature, there is minimal variation over a wide range of wear conditions. Fig.28 is a graph plotting a reference arc length of the idealised reference pulse of Fig.27 versus wear.

If, similarly, the arc length is computed for the reflected wave containing the phase information, then there is considerable variation. Fig.29 is a graph plotting an arc length of the reflected wave including phase information versus wear, showing variation of the arc length.

By dividing the measured arc length by the reference arc length then the ratio of arc length variation can be calculated.

Fig.30 is a graph plotting the ratio of arc length variation. An area of the graph of Fig.30 has been highlighted by a dashed line rectangle. In this region there is considerable linearity and a significant change in the arc length ratio with small amounts of wear. By creating a pattern in the surface 22 that is 200 μιη deep with grooves 100 μιη wide spaced 100 μιη apart there is potential to easily measure up to 100 μιη of wear using the arc length ratio technique. This technique will be largely independent of temperature as the effect of material thickness changes due to temperature are largely removed through the arc length referencing technique.

Fig.31 is a schematic diagram of a sound wave reflecting from a stepped boundary surface, or wear surface 220 of a component 200. A sensor 300 is attached to the component 200 on a mounting surface 240 opposite to the wear surface 220. The stepped boundary surface 220 includes a first part 260a and a second part 260b, separated by a step 260c. The first and second parts 260a and 260b are surfaces which are parallel to one another and the mounting surface 240. The step 260c is normal to the mounting surface 240. Fig.31 shows a wave 280 emitted from the sensor 300 which reflects from first and second parts 260a and 260b. The wave is broken into first and second wave parts 280a and 280b which are reflected from each of the first and second parts 260a and 260b respectively.

Fig.31 helps to explain how a characteristic is induced into a single wave by a feature (the stepped surface 260) in order to measure wear on a component. This is done by changing the height of the step 260c.

Figs.32,34,36, 38 are graphs plotting amplitude versus time of a modelled waveform reflected from a 1.5 mm, 1 mm, 0.8 mm, 0.6 mm step 260c respectively. Figs.33, 35,37, 39 are graphs plotting FFT amplitude of the waveforms of Figs.32, 34,36, 38 respectively.

As can be seen from Fig.32, when the step is 1.5 mm (2.55 times the wavelength) there are two distinct pulses in the reflected wave. In the corresponding FFT (Fig.33) there are 6 distinct frequency dips occurring at approximately 6, 7.75, 9.5, 11.25, 13 and 14.75 MHz.

Turning to Fig.34, when the step is decreased to 1 mm (1.7 times the wavelength) the two pulses are beginning to merge but are still distinct in the reflected wave. In the corresponding FFT (Fig.35) there are now 4 distinct frequency dips occurring at approximately 6.5, 9.25, 12 and 14.25 MHz. Turning to Fig.36, when the step is decreased to 0.8 mm (1.36 times the wavelength) the two pulses are merged further in the reflected wave. In the corresponding FFT (Fig.37) there are now 2 distinct frequency dips occurring at approximately 8.25 and 11.75 MHz.

Finally, turning to Fig.38, when the step is decreased to 0.6 mm (~1 wavelength) the two pulses are merged further to be indistinguishable in the reflected wave. In the corresponding FFT (Fig.39) there are now 2 distinct frequency dips occurring at approximately 7 and 11.5 MHz.

The above description illustrates how reflecting an ultrasound wave off a feature or features in the surface of a component or other structure can be used to calculate wear.

There are other techniques which can be used, and these techniques are now described with reference to previous Figs.1 and 5 and newly mentioned Figs. 40-49.

In order to simulate wear the component 20 described in relation to Fig.5 was mounted vertically in a horizontal grinding machine. The grinding machine was set to grind away five μιη steps as required. The grinding machine made several passes across the component 20 in order to achieve each 5 um step. No cooling fluid was used. Fig.40 is a graph plotting amplitude versus time for reflected waves showing the effect of wear from 100 to 200 μιη on the component 20. The pre-wear reflected wave is also shown.

It can be seen that there is a clear decrease in the reflected wave peak-to-peak amplitude with increasing wear from 100 to 200 μιη.

Fig.41 is a graph plotting amplitude versus time for reflected waves showing the effect of wear from 400 to 500 um on the component 20. It can be seen that there is a clear change in the time position with increasing wear from 400 to 500 μιη. In particular, the time position decreases.

Fig.42 is a graph plotting amplitudes versus time for reflected waves showing three positions where there are various degrees of spread of the zero crossing across wear depths of 420 to 500 μιη.

It can be seen that there is a clear change in the zero crossing position with increasing wear.

Fig.43 is a graph plotting FFT amplitudes versus frequency of the pulses of Fig.40 showing the effect of wear from 100 to 200 μιη, plotted in 10 um wear intervals.

It can be seen that the FFT amplitude decreases with increasing wear from 100 to 200 μιη. Fig.44 is a graph plotting a ratio of amplitude at 10.2 MHz divided by amplitude at 6.5 MHz for wear in the range 450 to 500 μιη.

As can be seen, the ratio of 10.2 MHz divided by 6.5 MHz increases with increasing wear. A polynomial trendline can be derived and in this case is given by Equation 1 below:

Equation 1 : y = 0.0003x² - 0.2503x + 53.849 Fig.45 is a graph plotting FFT amplitudes versus frequency of the pulses of Fig.40 showing the effect of wear from 200 to 300 μιη, plotted in 10 μιη wear intervals.

It can be seen that the FFT amplitude varies with increasing wear. There are also dips in the FFT amplitudes, with minimum dips occurring at around 12 MHz as shown encircled in Fig.45. Fig.46 is a close up of Fig.45 corresponding to the encircled part. Locations of minimum amplitudes in the dips are plotted. The trajectory of the minimum amplitude of each dip is shown when moving from 200 to 300 μιη of wear in 10 μιη intervals. The frequency and/or value of the minimum amplitude can be used to determine the wear on the component 20.

Figs.47 and 48 repeat the graphs of Figs.45 and 46 but for wear going from 400 to 500 μιη and showing in close up the minimum amplitudes of dip, which occur at around 6 MHz.

Fig.49 is a plot of wear with frequency of the minimum dip in the frequency spectrum for the FFTs of Fig.48. A straight line approximation is also plotted which could be used for an analysis of wear.

Fig.50 is a flow chart showing a method of performing wear measurement. Here, the basic steps are shown of: S500: Emitting a wave toward a feature having a non-planar impedance discontinuity.

S501 : Inducing a characteristic into the wave as a result of the feature.

S502: Receiving the wave having the characteristic.

S503 : Determining wear from the characteristic. Details of the various basic steps will be apparent from the above detailed discussion.

Fig.51 is a flow chart showing a method of determining wear. Here, the basic steps are shown of:

S510: Performing a FFT or equivalent on the received wave. S511 : Removing phase information. S512 Performing inverse-FFT or equivalent to obtain an idealised wave.

S513 : Comparing the idealised wave with a measured wave to determine wear. Final Statements

The examples described use a component 20, 200 made from 316 stainless steel. However, other materials could be used, whether metallic or non-metallic. Some other example materials are cast iron, Inconel, ceramics, polymers, steels, non-ferrous materials etc.

The components 20 and 200 are cylinders. However, the geometry of the component in not thought to be important, and other geometries could be used, extending the usefulness of the wear measurement technique.

The measurement technique described uses ultrasound. However, it may be that other frequencies can be used, and it may be that non-mechanical waves such as electro-magnetic waves could be used to measure wear in the same or in a similar way.

In the examples described, the same sensor 30, 300 is used both to transmit and receive a wave or pulse. It may be that other arrangements could be used. For example, separate sensors could be employed, and the receiving sensor could be arranged to receive a wave reflected from a feature 26, 260 or transmitted through a feature 26, 260.

The feature 26, 260 described is open in the sense that a roughness would be apparent on the wear surface 22, 220 of the component 20, 200 where the feature 26, 260 is located. The measurement technique relies on acoustic or other impedance discontinuities to induce a characteristic into a reflected or transmitted wave. Therefore, back-filled features could be used to create a flush wear surface 22, 220. In this respect, for example, materials could be used having similar hardness to the component material but a different density.

Other types of ultrasonic sensor 30, 300 could be used, having different dimensions, signal strengths, input resolutions and centre frequencies, for example. EMAT and laser sensors are two examples.

The features 26, 260 can be constructed in any suitable way, for example by eroding, engraving, micromachining, etching, stamping, casting, cutting, carving and laser ablation.

The technique described in the document provides a new, useful method and apparatus for measuring wear. The technique is non-invasive, can be carried out without stopping the equipment of which the component 20, 200 is part, can measure sub-wavelength features and wear, and is largely temperature independent. For example, for a +- 50 degC swing in temperature, the technique is capable of delivering a +- 0.15 μιη error in wear measurement. Also the technique does not rely on very accurate timing circuits reducing cost and complexity.

Also, various features could be used to induce the characteristic. For example, regular or irregular profiles or patterns could be used. Some examples include the following:

• regular 60 degree equispaced triangular profiles;

• variations on the above with flat peaks and troughs;

• combinations of two or more profiles;

• variations of angles;

• non-symmetrical grooves;

• combinations of various depths, widths, angles and horizontal spacing;

• concave or convex or combined features;

• three-dimensional surface profiles, with regular, irregular and dimpled profiles or patterns.

Some surfaces may require features to be placed/machined/ground/etched/turned/lasered in them in order to measure wear using this technique. Other surfaces may have a surface that is suitable for measurement without further modification. For example in internal combustion engines the cylinder bore or liner is often honed with a cross hatch pattern that has grooved that protrude into the surface. Another example is where hydrodynamic bearings have a dimpled surface to improve performance. In both these cases it may be possible to use the proposed technique to measure any wear without further modification.

Although a few preferred embodiments have been shown and described, it will be appreciated by those skilled in the art that various changes and modifications might be made without departing from the scope of the invention, as defined in the appended claims.

Claims

Claims:

1. A method of measuring wear, the method comprising: emitting a wave toward a boundary surface, the boundary surface comprising a feature having a non-planar impedance discontinuity for indicating wear, the feature being arranged to induce a characteristic into the wave so that the wave contains the characteristic; receiving the wave, measuring the characteristic from the wave and determining wear on the component from the characteristic.

2. An apparatus for measuring wear, the apparatus comprising: a receiver for receiving a wave from a boundary surface comprising a feature having a non-planar impedance discontinuity for indicating wear, the feature being arranged to induce a characteristic into the wave; a wear determining module configured to measure the characteristic and to determine wear on the boundary surface from the characteristic.

3. A system for measuring wear, the system comprising: a surface boundary comprising a feature having a non-planar impedance discontinuity for indicating wear; a transmitter arranged to emit a wave toward the boundary surface so that the wave comprises a characteristic representative of the feature; a receiver for receiving the wave from the boundary surface; and a wear determining module configured to measure the characteristic and to determine wear on the boundary surface from the characteristic.

4. The method, apparatus or system of any preceding claim, wherein the impedance discontinuity is an acoustic impedance discontinuity.

5. The method, apparatus or system of any preceding claim, wherein the feature comprises at least two impedance discontinuities arranged to interfere with the wave.

6. The method, apparatus or system of claim 5, wherein the at least two acoustic impedance discontinuities are displaced from each other with respect to the wave by one or more of a greater than, equal to and sub-wavelength dimension.

7. The method, apparatus or system of any preceding claim, wherein the feature comprises two materials having different acoustic properties and the boundary surface in the region of the feature is planar.

8. The method, apparatus or system of any preceding claim, wherein the feature comprises a step.

9. The method, apparatus or system of any preceding claim, wherein the feature repeats so that there are more than two acoustic impedance discontinuities.

10. The method, apparatus or system of any preceding claim, wherein at least one dimension of the feature is less than the wavelength of the wave.

11. The method, apparatus or system of any preceding claim, wherein the width of a sub-feature of the feature is less than the wavelength of the wave.

12. The method, apparatus or system of any preceding claim, wherein wear is determined by time domain analysis of the reflected wave.

13. The method, apparatus or system of claim 12, wherein the time domain analysis measures an amplitude.

14. The method, apparatus or system of claim 13, wherein the time domain analysis measures an amplitude relative to another amplitude within the reflected wave.

15. The method, apparatus or system of any preceding claim, wherein wear is determined by spectral analysis of the reflected wave.

16. The method, apparatus or system of claim 15, wherein the spectral analysis uses a Fast Fourier Transform (FFT).

17. The method, apparatus or system of claim 15 or 16, wherein the spectral analysis measures an amplitude at a specific frequency in the amplitude-frequency spectrum.

18. The method, apparatus or system of claim 17, wherein the spectral analysis measures a first amplitude at a first frequency and a second amplitude at a second frequency in the amplitude-frequency spectrum and compares the first amplitude and the second amplitude.

19. The method, apparatus or system of claims 15 to 18, wherein the spectral analysis measures a frequency value at which the amplitude dips in the amplitude-frequency spectrum.

20. The method, apparatus or system of claim 19, wherein the spectral analysis measures a first frequency value at which the amplitude dips with a second frequency at which the amplitude also dips in the amplitude-frequency spectrum and compares the first frequency and the second frequency.

21. The method, apparatus or system of claims 15 to 20, wherein the spectral analysis measures a phase value in the phase- frequency spectrum.

22. The method, apparatus or system of claim 21, wherein the spectral analysis measures a phase value relative to another phase value in the phase-frequency spectrum and compares the two phase values.

23. The method, apparatus or system of claims 15 to 22, wherein the spectral analysis measures a frequency value at which the phase dips in the phase-frequency spectrum.

24. The method, apparatus or system of claims 15 to 23, wherein the spectral analysis measures a first frequency value at which the phase dips with a second frequency at which the phase also dips in the amplitude- frequency spectrum and compares the first phase and the second phase.

25. The method, apparatus or system of any preceding claim, wherein wear is determined by removing phase information from the wave to create an idealised wave.

26. The method, apparatus or system of claim 25, wherein the phase information is removed by performing a Fast Fourier Transform (FFT) on the reflected wave, discarding the phase information and performing an inverse FFT on the amplitude-frequency information.

27. The method, apparatus or system of claim 25 or 26, wherein wear is determined by measuring the arc length of idealised reflected wave.

28. The method, apparatus or system of claim 27,wherein the arc length is compared with a reference arc length.

29. The method, apparatus or system of any preceding claim, wherein the boundary surface is an outer surface of a component.

30. The method, apparatus or system of any preceding claim, wherein the component is in one of a sliding contact arrangement, an abrasive contact arrangement or a pivoting contact arrangement.

31. The method, apparatus or system of any preceding claim, wherein the boundary surface is coated.

32. The method, apparatus or system of any preceding claim, wherein the wave is an ultrasonic wave.

33. Use of a feature having a non-planar impedance discontinuity in the path of a wave arranged to induce a characteristic into the wave for the purpose of wear measurement.