· Table of Contents
Corrosion-related Acoustic Emission and Its IdentificationRong S Geng
Beijing Aeronautical Technology Research Centre, Beijing 100076, China
Studies have been made on the characteristics of acoustic emission(AE) related with corrosions of aviation aluminum alloy. A theoretical model was proposed for source mechanism of AE produced by the initiation of pitting corrosion. Order of magnitude of surface displacement due to a single pitting was estimated. Special attention was paid to the differentiation between background noise and pitting-related genuine AE signals.
Keywords: pitting corrosion, acoustic emission (AE), plate wave
Corrosion, especially pitting and exfoliation are always of a great concern because they often cause significant damages to metals and structures. The calendar life of an aircraft is always mainly decided by its corrosion states and in many cases, great economical losses, safety problems, and even loss of life are attributed to corrosion. The landing-gears, bilge, empennage, fastener holes and fuel tanks of aircraft are places susceptible to corrosion most frequently. Corrosion of fuel tank often leads to leakage of the tank and subsequently severe accident. Although pitting corrosion takes place in a few individual locations, it can develop in due course and becomes pits, even through holes, which are potentially the most dangerous structural faults. If we let pitting develop without proper counter-measures, the whole structure may collapse at anytime.
People wish to monitor and detect corrosion at its early stage, best on its initiation, and then are able to take proper measures to prevent its development and effectively protect structures from being corroded. This paper aims to present a new method of using acoustic emission (AE) technique to detect corrosion initiation. Emphasis will be put on the description of source mechanism of AE induced by corrosion process and on the identification of thus produced AE signals from background noises.
The source mechanism of acoustic emission from pure corrosion activity is much debatable, but it is generally accepted that the most plausible source is the nucleation of hydrogen gas bubble from solution, or the breakage of a local passivation film. It is also generally accepted that pitting process is caused by the breakage of passivation film. In a system of metal-solution medium, there exists a specific threshold value of anode polarization potential, above which pitting corrosion will take place and below which, on the contrary, no pitting shall occur. This special potential is called the breakdown voltage of passivation film, or the critical potential of pitting nucleation. Because the breakage of passivation film is a dynamic process and will apply a pulse force onto the metal surface, it is therefore expected that acoustic emission will be produced by this process. In return, we can hence use AE to monitor corrosion. This is the theoretical basis for using AE to monitor and evaluate pitting corrosion process.
Let us first investigate the breaking process of a film and explain why such a process will produce acoustic emission. Assume the surface tension coefficient of the solution of corrosion fluid is a, the radius of (a bubble-like) film is R, then the extra pressure exerted by the film onto the surface of the specimen is as follows:
If using a=7´
10-2N/m, R= 5´
10-3mm, this will give a value of P=2.8´
104N/m2. The average force thus produced is F = 2pR2P(considering top half sphere only), making:
The surface of the plate-like specimen will produce very complicated motions while a pulse force of FH(t) is acting on it. According to Knopoff, the perpendicular displacement at epicenter is given by the following expression:
where, w=a t/b, y=bt/b, a=b/a and b is velocity of shear wave, a velocity of longitudinal wave, m shear modules, b plate thickness, and H(t) the Heaviside step function. The first term inside the bracket is the contribution by longitudinal wave component, whereas the second one is by shear wave component. As an approximation of first order, the displacement can be simply expressed as:
where only the contribution by longitudinal wave is under consideration(w=1 Substituting value of the force F , modulus m (2.6´1010N/m2) and plate thickness b(2mm) into above formula, the perpendicular displacement can be obtained as 9´10-15m.
Considering the sensitivity S of AE sensor of type WD is +60dB [0 dB ref. to 1V/(m/s)], i.e. 1000V/(m/s), after a few steps of calculations, it can be obtained that the output voltage on the sensor's output , corresponding to above-mentioned force, shall be (at around 300 kHz)on the order of magnitude of 17mV, which is much higher than the background noise level of 2.5 mV RMS or 7.5 mV of maximum peak value. This result shows that there is no any difficulty in detecting acoustic emission signal produced by the breakage of passivation film. The displacement due to shear wave components is much larger than that due to longitudinal ones, it can be further concluded, therefore, that its detection should be much easier. In short, present system is certainly capable of detecting AE signals produced by pitting process.
Next, let us investigate the frequency range of AE signals produced by pitting. One of the necessary conditions for a film to be completely broken-out is that the sound wave must propagate through whole film circumference. Assuming the velocity of sound wave is 340m/s, the circumference of a bubble-like film is 2pR=0.03mm, then the time for a bubble to be completely broken-out is 0.1ms. Theoretically, the frequency range can therefore be extended up to 10MHz. Even if the bubble is ten times larger, say R= 5´10-2mm, upper frequency of AE signals can still reach 1MHz. Experimental results have shown this finding and that high frequency contents of AE signal produced by pitting corrosion are abundant. This analysis also agrees well with that of other researchers' findings for corrosion-related acoustic emission. In the paper by Dai and Zhang, it was claimed that AE during slow strain rate process was characteristic of continuous type with frequency being rather low, whereas AE from stress corrosion was of burst type with frequency being concentrated at very high ranges(400kHz - 800kHz).
If, as in most cases, the thickness of the specimen is much less than the wavelength, then the predominant modes excited in the plate shall be the lowest two modes, i.e. the lowest symmetric mode S0 and lowest asymmetric mode A0, and that whichever is dominant depends on the way how the force is applied[3,4].
Pitting and exfoliation are both producing out-of-plane force, i.e. they take place mostly in the direction perpendicular to specimen plane, and hence they will both produce AE signals with A0 mode - flexural wave being dominant and S0 mode -extensional component negligibly smaller[3,4]. That is to say, the two types of corrosions will produce plate waves with flexural wave being the dominant mode.
Specimens used in the test are aluminum alloy of LY12, cut from wing skin of an aircraft. Corrosion solution is made of K3Fe(CN)6 11.5g , K4Fe(CN)6.3H2O 13.47g, NaCl 175.32g, H2O 930ml, making 1000ml solution in total. Mistras-2001 (Fig.1), a digital AE instrument by PAC, is used to capture and process AE signal produced by pitting corrosion. In addition to giving various parameters of AE signals, the instrument can also capture AE waveform in real time and analyze it in frequency domain. The block diagram of the measuring system is shown in Fig.2
|Fig 1: Mistras-2001 AE instrument.|
|Fig 2: The block diagram of testing system.|
During the first 24 hours of test after the specimen was partly put into the solution, the system was very "quiet", even though the threshold value Vt of system was set to 28dB(0 dB ref. to 1mV on the input of amplifier). After 72 hours, due to the increment of received signals, the threshold value Vt had to be set to 34dB. In the later stage of the test, the threshold voltage was fixed to 34dB. This already proves that AE technique is capable of detecting initiation of pitting corrosion, and is more sensitive than other means.
It is easy to differentiate between corrosion-related AE signal and background noise by using the combination of parameters and real time waveforms. As discussed before, the breakage of a passivation film is equivalent to a perpendicular force acting onto specimen surface, and this force will result in two lowest modes of Lamb wave - extensional and flexural waves, respectively. Moreover, the flexural wave will be predominant because of the force direction being perpendicular to plate surface. Noises show no such characteristics. In Fig.3 is shown the waveform of a typical noise, where distinctive extensional and flexural waves are not seen and are also characteristic of much smaller amplitude as compared with that in Fig.4. Typical pitting corrosion-related AE signals are shown in Fig.4, where the extensional wave has significantly small amplitude and arrives at early time, whereas the later one of flexural wave has much larger amplitude (about 56dB) and is characteristic of frequency dispersion.
Present test system has the function of using synchronized trigger, which can set the same starting point of time for all channels. With the help of this function, it is easy to locate source position, and to further differentiate between noise and genuine AE signal. The AE is a definite event, which in general arrives at different channels at different times, whereas noise has no such characteristics.
|Fig 3: Typical background noise.|
Fig 4: Typical pitting corrosion-related AE.|
(the arrow is pointing to extensional part)
In some cases, we can also differentiate between noise and AE by using parameter analysis, for example, by AE amplitude distribution. In general, noise shows a distribution concentrated at below 40dB, whereas the corrosion related AE has a rather wide range of amplitude, extending up to say 60dB, see Figs. 5 and 6, respectively.
|Fig 5: noise characteristic of amplitude being below 40dB||Fig 6: corrosion-produced AE characteristic of amplitude being up-to 60dB|
Another important conclusion may also be drawn that the AE produced by corrosion is burst type, not continuous one. Recording of AE signals in real time has clearly shown the fact that most AE events of influence are being occurring at discrete times , which means that corrosion, at least more severe pitting, also occurs at discrete instances. One of our major aims in the future is to relate AE with corrosion damage quantitatively, and to use AE as an effective means in evaluating corrosion damage for aviation materials.
|© AINDT , created by NDT.net|||Home| |Top||