WO2003081228A2

WO2003081228A2 - Acoustic emission

Info

Publication number: WO2003081228A2
Application number: PCT/GB2003/001174
Authority: WO
Inventors: John Little
Original assignee: Imes Limited
Priority date: 2002-03-19
Filing date: 2003-03-19
Publication date: 2003-10-02
Also published as: GB2403540A; AU2003214430A1; GB2403540B; GB0206460D0; GB0423219D0; WO2003081228A3

Abstract

Apparatus for testing a number of pressure vessels (10) comprises an accumulator (11) to deliver fluid pressure via hydraulic hoses and a pressure gauge (12). Detectors (14) are mounted to the vessels to sense acoustic emission (AE) when the vessels are subjected to cyclical pressurisation and de-pressurisation. AE signals from the detectors are transmitted to a central processing unit (13) for analysis and display on a display unit (15). A prediction of the working lifespan of the vessels is made by analysing AE signals occurring when the pressure in the vessels is on the increase.

Description

Acoustic emission

This invention relates to acoustic emission and more particularly, to the use of acoustic emission techniques to monitor and test structures such as pressure vessels.

There is a need for a reliable way of monitoring pressure vessels such as composite cylinders and non-destructively testing them in order to quantify their performance capabilities, for example to predict their expected working life.

Composite cylinders are structures that are made f om two or more different materials. They are typically made from an inner liner of metal, such as spun aluminium, around which an outer layer is formed from a fibre/resin matrix using materials such as Kevlar or carbon fibre. Such composite cylinders are known in the art and one of the leading manufacturers of them is SCI Inc. in California. Composite cylinders are increasingly being used in applications where previously steel cylinders were used.

Composite cylinders have applications in many different fields. In use, however, they will typically be required to be filled and re-filled with pressurised fluid many times over. This cyclical pressurisation and de- pressurisation puts stress on the vessels, which leads to their deterioration from effects such as microcracking, de-lamination and fibre breaks, and ultimately causes their failure.

It is known that when a pressure vessel is pressurised, the stresses and strains thereby induced in the material of which it is made produce small amounts of high frequency vibrations. Such vibrations are known as acoustic emissions (AE) and these can be detected by ultrasonic transducers applied to the surface of the vessel. The effects of AE testing in metal vessels such as steel cylinders is well documented. However, there appears to be no documentation about AE testing of composite cylinders for lifespan prediction, and AE testing work on metal vessels would not obviously have any direct application to AE testing of composite cylinders, because of the very different nature of their respective materials and compositions.

The present invention is concerned with a system in which one or more composite cylinders are arranged to be subjected to a series of cyclical pressure tests. Each cylinder will have at least one, and preferably two or more AE sensors, eg piezoelectric transducers, attached to its outer surface. Any AE vibrations generated during the tests will be picked up by these sensors, which are in turn arranged to transmit signals to a central processing unit for the data to be logged and processed. By analysing the AE data obtained in this way, a prediction can be made as to the health of each cylinder and its expected lifespan.

The correlation between AE signals obtained in pressure tests on metal vessels and residual life expectancy is known qualitatively, and practical destructive testing of composite cylinders has shown that a similar correlation exists for these too. Essentially, the correlation is that absolute AE energy is inversely proportional to cylinder lifespan. What the applicant has found is that a reliable and more accurate prediction is achieved if the analysis is based only on AE data acquired (a) whilst the fluid pressure in the cylinder is on the increase and (b) preferably also, only if the rate of increase of the fluid pressure at the time is below a certain threshold value.

This refinement is significant, because in actual practice, the pressurisation of cylinders in a test rig is in general not likely to be perfectly smooth. On the contrary, if the test cycle is typically for the cylinders to be taken up to pressure and down again several times a minute, there is likely to be some fluctuation in the pressure applied. This would be likely to manifest itself in the form of a pulsing or fluttering effect, most probably due to the nature of the pressure fluid used and/or possibly to some extent to the configuration of the test rig itself. Of course, one way of possibly eUminating such pressure variations might be to ensure that pressurisation of the vessels takes place at a sufficiently slow and steady rate. As a practical matter, however, this would not be a useful way of testing cylinders, because it would take far too long.

According to the invention there is provided a method of testing a pressure vessel comprising the steps of alternately pressurising and de-pressurising the vessel, monitoring acoustic emission (AE) vibrations in the vessel during said pressurisation and de-pressurisation, selectively filtering the AE signal data by excluding AE signals occurring whilst the vessel pressure is static or falling, and using the selectively filtered data to predict the lifespan of the vessel.

The invention also provides apparatus for testing a pressure vessel, in which the vessel is subjected alternately to pressurisation and de-pressurisation, and where acoustic emission (AE) vibrations in the vessel during said pressurisation and de-pressurisation are monitored, said apparatus comprising means for selectively filtering said AE signal data to exclude AE signals occurring whilst the vessel pressure is static or falling, and using said selectively filtered data to predict the lifespan of the vessel.

By way of example, embodiments of the invention will now be described with reference to the accompanying drawings, in which:

Figure 1 illustrates a typical form of apparatus according to the invention,

Figure 2 illustrates a typical form of cyclical pressurisation test,

Figure 3 is a plot similar to Figure 2 but in greater detail, and Figure 4 is a plot similar to Figure 3 but with AE signal data superimposed.

The test rig in Figure 1 is set up for monitoring three composite cylinders 10. An accumulator 11 is connected by hydraulic hoses to each cylinder 10 to supply each with hydraulic fluid under pressure. A pressure transducer 12 in the line senses the fluid pressure going to the cylinders 10 and signals this to a central processing unit 13. A number of piezoelectric detectors 14 are attached to the outer surface of each cylinder 10. In this case, there are two such detectors 14 per cylinder 10. The detectors 14 sense any AE vibrations which occur in the cylinders 10 and they are all connected up to transmit data signals to the central processing unit 13. The detectors 14 are preferably arranged on the cylinders 10 so as to sense AE vibrations in different orthogonal directions: After analysis of the input data by the unit 13, the results are transmitted to a display unit 15.

The system is preferably arranged to operate in real time, so that the results can be interpreted by the operator as the test takes place.

Figure 2 is a plot of fluid pressure against time and illustrates a typical cyclical pressurisation test on the cylinders 10. As will be seen, the cylinders are first rapidly taken up to their proof test pressure, where the pressure is maintained for a time before being released. The same cycle is then repeated a number of times, although in subsequent cycles the cylinders are taken to their working pressure, which is less than their proof test pressure.

Figure 3 is a plot of a typical pressurisation phase shown in much greater scale than in Figure 2. As will be seen, pressurisation of the cylinders does not in fact occur in a perfectly smooth, linear fashion. On the contrary, there are fluctuations in the pressure as it rises. This is due to the fact that the pressurisation is occurring over a relatively short space of time (the cycle time may typically be between 10 and 30 seconds) and is probably caused by the nature of the pressure fluid used in the system (which will typically be standard hydraulic fluid) and to a certain extent, the arrangement of the test rig itself. In any event, it will be seen that at any given instant, the pressure in the cylinders may actually be static or falling or rising, to a greater or lesser degree.

Figure 4 is a plot similar to Figure 3, ie showing variations of fluid pressure in the cylinders over time during a typical pressurisation phase. Here, however, there has additionally been superimposed a plot of AE signals generated by the vibrations that the detectors 14 from one cylinder have picked up during pressurisation. As will be seen, some of these AE signals (such as point a) occur at an instant when the fluid pressure in the particular cylinder is actually static; some (such as point b) when the pressure is momentarily falling and others (such as point c) when it is on the rise. According to the applicant's invention, the only AE signals that are to be used for the purposes of analysis are those that occur when the fluid pressure is truly on the rise. In terms of the plot shown in Figure 4, this means that the only AE points that are taken for analysis are those where the underlying pressure plot shows a line which slopes upwards from left to right.

It is also preferable to exclude AE signals from consideration if they occur at a time when the rate of increase of fluid pressure is above a certain value. This is because, at higher rates of pressure increase, the fluid may experience turbulence, giving rise to unwanted AE signals ("flow noise") as a result of vortices, for example, or resonance. The cut-off point for filtering the AE signals will vary from one application to another depending upon a number of factors, but here it might typically be set at a value of 5 bar per second rate of increase of pressure. In practice, such a filtering of the AE signals is achieved by processing all of the data using a specially designed algorithm in order to remove any AE signals other than those which meet the above criteria.

It is envisaged that in practice it should be possible to obtain sufficient AE data to obtain a reliable indication of expected cylinder lifespan with tests over as few as 20 cycles. This would have the advantage of enabling each cylinder to be tested individually and graded before leaving the factory, rather than being graded by batch according to a sample destructive test. Also, by arranging for the cylinders to be initially pressurised up to their proof test pressure, the cylinders can be given the necessary certification and graded in the same single test.

The amount of AE activity in pressure vessels is known to increase over the working life of the vessel. Therefore, the system described above can be used also to re-qualify cylinders that are in service in order to determine what stage they have reached in their expected lifespan.

Claims

1. A method of testing a pressure vessel comprising the steps of alternately pressurising and de-pressurising the vessel, monitoring acoustic emission (AE) vibrations in the vessel during said pressurisation and de- pressurisation, selectively filtering the AE signal data by excluding AE signals occurring whilst the vessel pressure is static or falling, and using the selectively filtered data to predict the lifespan of the vessel.

2. A method as claimed in claim 1 wherein the filtering step further excludes AE signals occurring whilst the rate of increase of vessel pressure is above a certain value.

3. A method as claimed in claim 1 or claim 2 wherein the pressurisation and de-pressurisation step is carried out over a number of regularly repeated cycles.

4. A method as claimed in any preceding claim wherein during a typical pressurisation cycle of the vessel, the vessel is taken up to its . working pressure.

5. A method as claimed in claim 4 wherein for the initial pressurisation cycle of the vessel, the vessel is taken up to its proof test pressure.

6. A method as claimed in any preceding claim wherein said testing occurs in real time.

7. A method as claimed in any preceding claim wherein the step of mordtoring AE vibrations is by means of at least one AE sensor mounted to the vessel.

8. A method as claimed in claim 7 wherein two or more AE sensors are mounted to the vessel at spaced apart locations.

9. A method as claimed in claim 8 wherein said two or more AE sensors are arranged to monitor AE vibrations in different orthogonal directions.

10. A method as claimed in any preceding claim in which two or more vessels are arranged to be tested at the same time.

11. Apparatus for testing a pressure vessel, in which the vessel is subjected alternately to pressurisation and de-pressurisation, and where acoustic emission (AE) vibrations in the vessel during said pressurisation and de- pressurisation are monitored, said apparatus comprising means for selectively filtering said AE signal data to exclude AE signals occurring whilst the vessel pressure is static or falling, and using said selectively filtered data to predict the lifespan of the vessel.

12. Apparatus as claimed in claim 11 wherein said selective filtering means further excludes AE signals occurring whilst the rate of increase of the vessel pressure is above a certain value.