Introduction
Ultrasonic NDT by the surface acoustic waves (SAW) is essentially urgent in investigations of screw-like and other rough surfaces. The quantitative evaluation of defects may be fulfilled by measurements of amplitudes, velocities and spectra of reflected signals. The theoretical models assume that the surface under investigation is perfectly flat and that surface Rayleigh waves propagate without losses. The different situation is in reality. SAW die away with covered distance. The attenuation factor depends not only on the frequency of the wave but also on mechanical characteristics of the surface.
The purpose of our work was to investigate the dependence of attenuation of SAW on the roughness of surface using measurement of amplitudes, velocities and spectra of penetrated pulse signals. At the same time the influence of liquid layers of various thickness upon the surface was evaluated. We tried to find possibilities to reduce attenuation of SAW that propagate on rough surfaces.
The equations for evaluation of scalar and vector potentials of surface waves may be presented in form [1]:
| (1) |
where q2= k2 - kl2 , s2=k2 - kl2 , B and D are constants, k is wavenumber of wave under investigation, kl, kt are wave numbers of longitudinal and transversal waves, z is coordinate to the depth of solid body from surface and x is coordinate along the wave transmission direction on the surface. For the Rayleigh wave these potentials are in the form:
| (2) |
These equations show that wave decreases with the depth according to 
for the longitudinal component and for the transversal component. The average energy density of the wave also decreases strongly with depth into the solid body till »lR. As was shown in [1, 2], we can expect in this way that conditions of the surface layer can influence transmission of Rayleigh waves.
The transmission of these waves depends also on the conditions upon the surface under investigation. We can also expect changes in transmission if layer of liquid covers the surface. The deepness of localization of surface waves depends on thickness of liquid layer [1].
Experimental technique
The digital defectoscope for SAW was used in investigation. Its diagram is shown in Fig.1. It contains pulse generator, exiting angular SAW transducer, gate pulse generator, electronic gate, SAW receiver connected to amplifier the output signal of which we observed on the screen of computerized spectrum analyzer PCS64i. Signal shape, time delay and spectrum information were observed on the PC monitor and stored in the PC memory.
Fig 1: Schematic of measuring arrangement for the SAW transmission experiment: S - specimen, SR - surface relief, T - SAW transmitter, R - SAW receiver,b - variable incidence angle
|
In SAW velocity evaluations using two identical angular transducers the measurements were carried out in two distances lTR1 and lTR2 with replacement lTR by D
lTR . The SAW signal delay time D
t was registered. The SAW velocity was then calculated by c=D
lTR/D t. Attenuation of SAW was defined using ratio of signals amplitudes when the distance lTR was substituted by lTR + D
lTR :
| (3) |
or attenuation factor
| (4) |
Fig 2: Models of surface relief structure: a - flat surface, b, c - triangular relief, d - rectangular relief
|
Investigations of roughness influence were carried out on the periodic surface structure of aluminum specimen shown in Fig.2. Measurements on the surface relief are affected by problem that we cannot ensure stability of excitation conditions (stability of SAW amplitudes) by changing position of transmitter. Therefore attenuation and velocity of SAW on the surface relief was evaluated by results of three measurements. First we measured on the flat surface amplitude and time delay when lTR1=0 and lTR2= DlTR = lsr (Fig.2). Then the third measurement on the surface relief by lTR2 was fulfilled.
Experimental results
Influence of liquid layer thickness on the propagation of ultrasonic SAW was investigated in cases of flat surface and on rough surfaces having different structures. Fig. 3 shows pulses and spectra of Rayleigh waves transmitted in aluminum specimen. We evaluated velocity of waves by measuring time delay Dt of SAW pulses by increasing covered distances D lTR. The attenuation was calculated from the magnitude changes of spectra maximum.
The results of measurements presented in the table 1 show that velocity and essentially attenuation of SAW on the surface structures in comparison with flat surface increases. This means that the change of velocity is not connected with absorption of SAW but it depends on scattering of waves and transformation into inhomogeneous (quasi transversal) waves. The energy is transmitted in these waves in the layers under structures and the surface influences weaker waves' propagation. These waves become similar to bulk transversal waves.
| cR, m/s
| aR, 1/cm
|
| Flat surface
| 2890
| 0.007
|
| Small triangular structure (Fig.2b)
| 3180
| 0.35
|
| Large triangular structure(Fig.2c)
| 3420
| 0.45
|
| Rectangular structure (Fig.2d)
| 3080
| 0.69
|
| Table 1: Velocities and attenuation factor of 1.8 MHz Rayleigh waves measured on aluminum |
| Fig 3: Pulses and their spectra of Rayleigh waves transmitted in aluminum specimen by 1.8 MHz transducer: a - flat surface, b - surface relief of case d) in Fig.2
| | |
Fig. 4 and 5 show dependencies of SAW pulses normalized amplitudes on various surface structures versus thickness of liquid layer.
Fig 4: Dependence of measured normalized amplitudes versus relative thickness of liquid layer d/l on aluminum surface in case of triangular relief: 1 - h0/l; 2 - h1/l; 3 - h2/l
|
Fig 5: Dependence of SAW pulse normalized amplitude versus relative liquid layer thickness on the aluminum surface: 1 - flat surface; 2 -structure d) in Fig. 2
|
The models of rough surfaces with different roughness were used in investigations. The characteristic parameter was the depth of scratches h. The ratio of scratches depth and SAW wavelength l was of several values: h0/l=0, h1/l= 0.28, h2/l= 0.66. Accordingly we obtained different amplitudes of penetrated SAW on the free aluminum surface: A0 : A1 : A3 = 1 : 0.33 : 0.28.
In Fig. 6 and 7 results of similar measurements with various frequencies are presented.
Fig 6: Dependence of SAW pulses normalized amplitude versus liquid layer thickness on the flat aluminum surface by various frequencies: 1 - f=1.8 MHz; 2 - f=2.5 MHz; 3 - f=5.0 MHz
|
Fig 7: Dependence of SAW pulses normalized amplitude versus liquid layer thickness on the aluminum with surface relief (case d) in Fig.2) by various frequencies: 1 - f=1.8 MHz; 2 - f=2.5 MHz; 3 - f=5.0 MHz
|
These curves show that the thin layers in the region 0...0.1 d/l influence transmission of SAW in the strongest way. The dependencies of amplitudes of SAW have an interference nature taking maxima and minima in certain points. Interference in the liquid layer is not stable because of uneven lower limit of layer. Such interference was predicted theoretically [1] and was named transformation of Rayleigh waves to normal surface waves.
Conclusions
The liquid layer on the surface under investigation influences SAW attenuation. SAW are transmitted in the structure solid surface - liquid layer like in waveguide and dependence of the signal amplitude versus liquid layer thickness has an interference nature. The SAW attenuation in the case of thin liquid layer (d <l
) is smaller then attenuation in immersion case (d>>l
). This phenomenon may increase sensitivity in investigations of industrial products having rough surfaces (screws, hulls of ships).
References:
- Viktorov I.A. Surface sound waves in solids.- Moscow: Nauka, 1981. 287 p. (in Russian).
- Sajauskas S., Zvanorius V. Investigation of acoustic waves in layered structures // Ultrasonics (Proceedings of Higher Schools of Lithuania). 1993. N.25. P.23-33.
