Spectral Analysis of Surface Waves (Sasw) : Prantik Maity
Spectral Analysis of Surface Waves (Sasw) : Prantik Maity
Spectral Analysis of Surface Waves (Sasw) : Prantik Maity
Prantik Maity
BCE IV 2017-2018
SPECTRAL ANALYSIS OF SURFACE
WAVES (SASW)
PREPARED BY:
PRANTIK MAITY
ROLL NO: 001410401112
BCE IV 1st SEMESTER
SECTION: - B1
i|Page
CERTIFICATE OF APPROVAL
This is to certify that Prantik Maity with roll no. 001410401112 has prepared
the seminar paper entitled “SPECTRAL ANALYSIS OF SURFACE WAVES
(SASW)” under my supervision as a part of his final year curriculum of
Department of Civil Engineering, Jadavpur University.
Date:
DR. NARAYAN ROY
Department of Civil
Engineering
Jadavpur University
Kolkata: 700032
ii | P a g e
ACKNOWLEDGEMENT
I would like to place on record my deep sense of gratitude to, DR. NARAYAN
ROY, Department of Civil Engineering, Jadavpur University for his generous
guidance, help and useful suggestions rendered to me to collect all relative theory,
data related to my seminar report and for his active support and encouragement
all along the seminar session.
Your sincerely,
iii | P a g e
TABLE OF CONTENTS
4. COLCLUSION ……………. 17
iv | P a g e
1. INTRODUCTION
The Spectral Analysis of Surface Waves (SASW) method is widely used to determine the
dynamic shear modulus and the material damping ratio of soils. It is based on an in situ
experiment where waves are generated by means of an impact hammer, a falling weight, or a
hydraulic shaker. The resulting wave field is recorded by a number of sensors at the surface of
soil and used to determine dispersion and attenuation curves. An inverse problem is solved to
identify the shear wave velocity and the material damping ratio profiles. The theoretical
dispersion and attenuation curves are calculated for a given soil profile and compared to the
corresponding experimental curves derived from the surface wave test. The soil profile is
subsequently adjusted in order to minimize the distance between the experimental and the
theoretical curves.
With the advent of spectral analysis and portable computers in late 1970s, the traditional
surface wave technique has revolutionised to the Spectral Analysis of Surface Waves (SASW)
method. Over the past decade, the SASW method has attracted many engineers and has been
utilised in different applications. These application areas include characterisation of
foundation, non-destructive evaluation and characterisation of pavement systems evaluation of
concrete structures, and in situ determination of ground stiffness.
1|Page
2. THEORITICAL BACKGROUND
Only longitudinal and transverse waves can propagate in a homogeneous, isotropic and
unlimited medium. If the medium is bounded, another type of waves, surface waves, can be
guided along the surface of the medium. These waves usually form the principal phase of
seismograms. There are two types of surface elastic waves:
1) Rayleigh waves: These waves are elliptically polarised in the plane which is
determined by the normal to the surface and by the direction of propagation. Near the
2|Page
surface of a homogeneous half-space, the particle motion is a retrograde vertical ellipse
(anticlockwise for a wave travelling to the right).
2) Love waves: The particle motion in these waves is transverse and parallel to the
surface. As opposed to Rayleigh waves, Love waves cannot propagate in a
homogeneous half-space. Love waves can propagate only if the S-wave velocity
generally increases with the distance from the surface of the medium.
Fig. 2.1. Motion induced by a typical plane wave that propagates in the x direction. Wave
motion does not vary in the y-direction
𝜕𝛷 𝜕𝛹
u= + (2.2.a)
𝜕𝑥 𝜕𝑧
𝜕𝛷 𝜕𝛹
w= - (2.2.b)
𝜕𝑧 𝜕𝑥
(2.2.c)
The rotation in the x-z plane is given by,
(2.2.d)
3|Page
Use of the potential functions allows separation of the effects of dilation and rotation i.e.,
(equation 2.2.c and 2.2.d) indicate that Φ and Ψ are associated with dilation and rotation,
respectively. Therefore, Rayleigh waves can be thought of as combinations of p- and s-waves
(SV waves for this case, since the x-z plane is vertical) that satisfy certain boundary conditions.
Substitution of the expressions for u and w into the equations of motion,
(2.2.e)
(2.2.f)
Solving equations 2.2.e and 2.2.f we get,
(2.2.g)
If the wave is harmonic with frequency ω and wave number kR, so that it propagates with
Rayleigh wave velocity VR = ω/ kR, the potential functions can be expressed as,
Were F and G are functions that describe the manner in which the amplitude of the dilatational
and rotational components of the Rayleigh wave vary with depth. Substituting these
expressions for Ψ and Φ into (equation 2.2.g) gives
4|Page
The potential functions can finally be written as,
(2.2.h)
Since neither shear nor normal stresses can exist at the free surface of the half-space, σxz = 0
and σzz = 0 when z=0. Therefore,
Using the potential function definitions of u and w (equations 2.2.a & 2.2.b) and the solutions
for the potential functions (equation 2.2.h) the free surface boundary conditions can be written
as,
(2.2.i)
(2.2.j)
With these results, the velocities and displacement pattern of Rayleigh waves can be
determined.
5|Page
Which can expanded and rearranged into the equation,
This equation is cubic in kRs2, and real solutions for kRs can be found for various values of
Poisson’s ratio. These allow evaluation of the ratios of the Rayleigh wave velocity to both s-
and p-wave velocities as functions of v. the solution shown in (figure 2.2.) shows that Rayleigh
waves travel slightly than s-waves for all values of Poisson’s ratio except 0.5.
Figure 2.2. Variation of Rayleigh wave and body wave propagation velocities with
Poisson’s ratio
6|Page
2.4. RAYLIEGH WAVE DISPLACEMENT AMPLITUDE
Substituting the solutions for the potential functions Φ and Ψ (equation 2.2.h) into the
expression for u and w (equation 2.2.a & 2.2.b) and carrying out the necessary partial
differentiations yields
(2.2.k)
Where the terms in parentheses described the variation of the amplitudes of u and w with depth.
These horizontal and vertical displacement amplitudes are illustrated for several values of
Poisson’s ratio in (figure 2.3.). Examination of (equations 2.2.k) indicates that the horizontal
and vertical displacements are out of phase by .Hence the horizontal displacement will be zero
when the vertical displacement reaches its maximum (or minimum), and vice versa.
Figure 2.3. Horizontal and vertical motion of Rayleigh waves. A negative amplitude ratio
indicates that the displacement is the opposite direction of the surface displacement
The Rayleigh waves produced by earthquakes were once thought to appear only at very large
epicentral distance (several hundred km). It is not recognized, however, that they can be
7|Page
significant at much shorter distance (a few tens of kilometres). The ratio of minimum epicentral
distance, R, to focal depth, h, at which Rayleigh waves first appear in a homogeneous medium,
is given by
Where VP and VR are the wave propagation velocities of p-waves and Rayleigh waves,
respectively.
8|Page
3. SPECTRAL ANALYSIS OF SURFACE WAVES (SASW)
METHOD
Fig. 3.1. Variation of vertical particle motion for Rayleigh waves with different wavelengths
3.2. EQUIPMENTS
9|Page
c. Automated Software - USB Connector: The user enters the required test frequencies
and the software runs the complete test automatically. Data can be exported directly
into Microsoft Excel.
d. Sledge Hammer
e. Battery or Generator
10 | P a g e
frequency sources were utilized. For sampling deeper material, low frequency receivers were
used with long receiver spacing and low frequency sources.
Two profiles, a forward profile and a reverse profile, are typically obtained in SASW
measurements where the accessible surface is struck by a hammer on two opposite sides of the
receivers. A signal analyser is used to collect and transform the receiver outputs to the
frequency domain. Two functions in the frequency domain are of great importance in SASW
tests:
a) The cross power spectrum between the two receivers (used in the preparation of the
experimental dispersion curve).
b) The coherence function (used to ensure that high signal to-noise ratio data is being collected).
Inversion of Dispersion Curve- The process of determining the actual propagation velocities
at different depth (velocity profile) from the experimental dispersion curve is known as the
inversion (back-calculation) of the Rayleigh wave dispersion curve. There are basically two
inversion processes, a simple and a refined one.
a) Simple Inversion- The simple inversion process is based on an earlier inversion of
the steady-state vibration technique. The inversion is done by re-scaling the phase
velocity axis to get the shear wave velocity, and the wavelength axis to get the
depth. This procedure was used in the preliminary investigation of the SASW
method. The approximate inversion technique is performed by assigning the shear
wave velocity equals to 1.1 times the phase velocity and the depth is 0. 33 times the
11 | P a g e
wavelength. It provides satisfactory results for geotechnical sites where the shear
modulus smoothly increases with depth. However, in cases where there is a large
contrast in shear wave velocities, this simple method can lead to erroneous results.
This is especially true in a pavement system in which the stiffness of the materials
can differ greatly.
b) Refined Inversion- A more reliable in version procedure uses stress-wave
propagation theory. The propagation theory models a theoretical dispersion curve,
which is compared with experimental dispersion curve. There are several modelling
approaches available for SASW applications. Among them are: (l) the transfer
matrix method; (2) the dynamic stiffness matrix method; and (3) the finite
difference method. The transfer and stiffness matrix methods provide exact
formulation as compared to the other methods. Although the approach of the
inversion methods can be different, however, all the methods assume that the profile
consists of a set of homogeneous layers extending to infinity in the horizontal
direction. The last layer is usually considered as a homogeneous half-space.
Signal Processing and Dispersion Curve Construction- The particle motion velocity (or
acceleration) recorded at the two receivers is used to evaluate the Rayleigh wave phase velocity
as a function of frequency. This step involves the computation of the time delay associated to
the wave arrival at the two successive positions. Being this time delay frequency dependent,
an appropriate algorithm has to be used.
The two signals in time domain (y1 (t) and y2 (t)) (Figure 3.4) are firstly translated in frequency
domain using a Fast Fourier Transform, obtaining the related linear spectra (Y1 (ω) and Y2 (ω)).
Using spectral analysis techniques, it is then possible to get information about the quality of
the records and eventually the phase velocity as a function of frequency.
Figure 3.4. Example of SASW signals: (a) whole signals; (b) wave-train arrivals
12 | P a g e
The assessment of signal quality is made using the “coherence function”, namely a spectral
quantity obtained comparing different registrations, that is a measure of the degree by which
input and output signals are linearly correlated. A value close to unity is an index of good
correlation and hence the recorded signals can be considered genuine and unaffected by ambient
noise.
Phase velocity as a function of frequency can be obtained from the phase of the average Cross-
Power Spectrum. The quantities involved in the evaluation of the dispersion curve, in the
succession according to which they are evaluated, are the following:
Auto-power spectra (Figure 3.5. c, d):
The use of the cross-power spectrum phase for the evaluation of the frequency dependent time
delay is based on the hypothesis that the wave group under consideration is composed of a
single mode of propagation and hence the phase velocity is function only of frequency. Under
such assumption the signal that is associated to a wave propagating along the x direction can
be expressed as a superposition of harmonic waves of the same mathematical form of that
propagating in a homogenous half space:
13 | P a g e
Where y(x, ω) is the amplitude spectral density, k (ω) is the wavenumber and ϕ (ω) is a constant
phase term due to effects other than propagation. The Fourier transform of such signal is:
Hence considering the cross-power spectrum relative to two different detection of the same
wave along the x -axis at location x1 and x2:
14 | P a g e
Figure 3.5. Spectral quantities evaluated from the signals of Figure 3.4.: (a) Phase of the cross-
power spectrum; (b) Coherence function; (c) Auto-power spectrum (first receiver); (d) Auto-
power spectrum (second receiver).
The plot between wavelength and surface wave velocity for comparison between experimental
data and theoretical dispersion curve has been shown in Figure 3.6 for sample experiment.
10
100
Experimental Data
Theoretical Dispersion Curve
Figure 3.6 Wavelength vs. Surface wave velocity for experimental data and theoretical
dispersion curve.
15 | P a g e
After that shear wave velocity profile with depth is generated and the plot for sample
experiment is shown is Figure 3.7.
20
Depth, m
40
60
80
The SASW method offers significant advantages. In contrast to borehole measurements which
are point estimates, SASW testing is a global measurement, that is, a much larger volume of
the subsurface is sampled. The resulting profile is representative of the subsurface properties
averaged over distances of up to several hundred feet. The resolution in the near surface (top
25 ft.) is typically greater than with other methods. Because the SASW method is non-invasive
and non-destructive, it is relatively easy to obtain the necessary permits for testing. At sites that
are favourable for surface wave propagation, the SASW method allows appreciable cost and
time savings.
16 | P a g e
3.7. APPLICATIONS OF SASW
Condition assessment of concrete, including liners in tunnels, slabs, and other structural
concrete members.
Evaluation of alkali-silica, fire, freeze-thaw and other cracking damage. Surface-
opening crack depth measurement.
Determination of abutment depths of bridges.
Determination of pavement system profiles including the surface layer, base and
subgrade materials with optional WINSASW software.
4. CONCLUSION
The SASW method is a relatively new seismic technique and has been under continuous
development during the last few years, in particular the data inversion analysis. The method
avoids the problems associated with borehole based methods. The non-destructive and non-
invasive nature of the testing procedure avoids sampling disturbance and unrepresentative
sampling. The most important characteristic of SASW method is that it can be used for irregular
profiles, including a profile with a softer layer trapped between stiffer layers, a profile with a
stiffer layer sandwiched between two softer layers, and a profile with a softer layer at depths.
The SASW methods should be able to be extended to studies on geological formations that
have distinct layering that model the half-space, such as layered sediments and Meta sediments.
One weak aspect of the spectral analysis typically adopted in the SASW test is the unwrapping
of the cross power spectrum phase, a necessary step for the estimation of phase velocity.
Usually the presence of near field effects and ambient noise strongly deteriorates the cross
power spectrum estimate at low frequencies and hence the unwrapping becomes a very ticklish
matter. Phase unwrapping is the problem that essentially inhibits the automation of dispersion
curve estimation. With the adoption of a multichannel procedure, phase unwrapping is avoided
and the dispersion curve estimation is easily automated.
17 | P a g e