Sub-bottom pipeline and cable detection
using automated contact recognition
Author: J. KWEE
ABSTRACT
As more and more wind energy parks are build
offshore these days, the need for detecting and
mapping the existing infrastructure below the water
bottom becomes increasingly vital. Relatively small
and easy to deploy sub-bottom profi ling systems
provide high resolution 2-D single channel seismics.
In this high resolution seismic data, pipelines and
cables are identifi ed by the display of hyperbolas.
The top of the hyperbola indicates the top of the cable,
which is used to determine the exact position and depth
of burial. These contacts can be determined manually,
taking into account human error and inconsistency.
A new development within the geophysical software
package Silas is the automated contact recognition
technique for the detection of these hyperbolas.
The contact recognition consists of the automated
determination of two parameters: 1) the semblance
of the apex of the hyperbola and 2) the accumulated
power of the hyperbola. These parameters are quantifi
able fi gures which enables a more objective classifi
cation of detected contacts. Other datasets, as
multibeam, sidescan sonar and magnetometer data,
can be combined with the sub-bottom data within the
software for cross correlation and clear imaging of
the sea bed and below.
INTRODUCTION
With the presence of more and more pipelines and
cables in the coastal waters, the need to determine
the exact location increases. This extended abstract
will focus on sub-bottom detections of cables and
pipelines using single channel shallow seismic
data and describes the automated contact detection
of Silas. Vital information with regards to cables
is the depth of burial of cables and pipelines.
A changing bathymetry can create free span of
cables, which could lead to potential dangers of an
interrupted power or communication grid. If the
location and depth is known, risks can be mitigated
before mayor problems arise. Methods using an
active signal on pipeline or the cable may detect this
properly, but result in a shutdown of the cable for
the time of monitoring. A passive method to detect
these cables and pipelines is sub bottom profi ling
and will be discussed in this extended abstract.
Sub-bottom profi ling -
Single channel 2-D seismic data
Most of the sub-bottom marine infrastructure is
situated in the top part of the seabed. Sub-bottom
profi ling uses transducers both sending and
receiving the signal resulting in 2-D, single channel
seismic profi les. These small systems can easily
be mounted, also on smaller vessels, allowing easy
surveys in coastal and even inland waters.
The lower the frequency of a signal, the deeper the
penetration, but at the cost of a lower resolution. This
tradeoff should always be kept in mind, when picking
a frequency for object detection. Important questions
are: at what (burial) depth do we expect the cable and
what is the size of the object that should be detected.
Especially for the detection of cables, a high resolution
is needed in the top part of you sub-surface.
The beam width and footprint of a sub bottom profi ler
are bigger than each of the individual beams of
www.stema-systems.nl
survey methods that only image the seafl oor itself,
as a multibeam. This inherently decreases the resolution
of detection. Using multiple transducers in an
array increases the signal/noise ratio in the overlapping
ellipse.
Nevertheless, the bigger beam width can be used
as an advantage for the detection of cables and
pipelines. These features are visible in sub bottom
data as hyperbolas. As can be seen in Figure 1, part
of the signal refl ects on top of an object before the
vessel sails right above it, due to this larger beam
width. It is, however, plotted as if it is detected right
below the vessel, but at greater depth. The closer the
vessel sails to the pipeline or cable, the shallower
it shows up in the sub-bottom data. The shallowest
point is reached right on top of the cable. A reverse
trend occurs moving away from the linear object,
resulting in a hyperbola.
Acquisition and processing
The acquisition of sub bottom data for the detection of
cables and pipelines involves the sailing of crosslines
on top of the linear object. The more pings refl ect on
top of the linear object, the more clear the hyperbola
can occur in the data, so a slow survey speed and a
high pingrate are recommended.
As described in the fi rst section, the recorded data
is a sub-bottom acoustic profi le. As always the case
with sub-bottom data, the raw signal is processed,
but interpretation to identify the seafl oor, geological
layers and objects has to be made. Results are
less straight forward hydrographical methods like
multibeam surveys, that give direct results of the
depth of the seafl oor.
Other than multichannel data, single channel
sub-bottom profi les can be viewed directly as
recorded without any processing needed. However,
no fi lters or heave reduction are (fully) applied to
the online data. This makes it less clear if objects
are detected directly. For data quality purposes data
should be processed quickly and verifi ed. An easy
option to apply a quick review of the data is to use
an automated batch processing as is available in the
Silas Processing software suite.
Figure 1: Object detection in acoustic data. Above the different path
lengths to the object are displayed depending on the location. Below
the corresponding acoustic traces, resulting in a hyperbole, highlighted
in red.
Another advantage that sub-bottom profi ling (and
seismic data in general) have above other detection
methods is that, besides the detection of the objects
themselves, the complete subsurface is imaged.
This allows also to detect bottom features related to
the constructions of pipes and cables as trenches,
initially dredged to lay them in. The extra information
helps interpretation and could give an indication,
even if the pipelines and cables are not detected
themselves.
Interpretation of the seafl oor is relatively easy and is
mostly done by auto-tracing algorithms. For layers
in the subsurface, interpretation need to be a bit
more manual, especially if layers are not as distinct.
The interpretation of objects can even be more complicated.
Several diffi culties that can occur during
this interpretation, as is also stated by (Wunderlich
et al., 2005) , are:
• Objects can be masked by refl ections of nearby
layers (as the seafl oor or other layer boundaries)
and other structures.
• Weak echo strength due to acoustic attenuation
in the sediment.
• Small refl ection coeffi cients due to small
acoustic difference (density, sound velocity) to
surrounding material.
• Small dimensions of objects and unknown or
imprecise know positions.
The picking of cables, pipelines and objects in sub-bottom
data is mainly interpretation work that requires
experience of the personnel and these man made
interpretations change between different individuals.
25
