Page 79 - Reliability and Maintainability of In service Pipelines

68 Reliability and Maintainability of In-Service Pipelines

found that changes in pipes, from clean condition to having tar levels of just
0.25 mm, are detectable and also results in a decrease of received signal ampli-
tude. However, this method is yet to be explored in terms of its potential for dif-
ferent type of pipe applications.
A challenging field based test was undertaken by Lewis and Fisk (2005),
where they used the sonic/ultrasonic method to measure a 152 cm concrete
pipe wall which was thinning due to a number of factors. For this purpose test
pits were excavated along the pipe to allow four sonic sensors based on a
curve to be placed on the pipe outside the coating. Unfortunately, no informa-
tion on the sensor type and processing technique was provided. However, the
results showed that recorded resonate frequencies of 16 to 18 kHz corre-
sponded to pipe wall thicknesses of 11.2 11.7 cm and frequencies as high as
21 29 kHz indicated that the pipe wall thickness was reduced. Unfortunately
no assessment of method accuracy, comparison, and further application was
suggested.
The above research was carried on further and Fisk and Marshall (2010) pub-
lished extended field study results, where 81 prestressed concrete cylinder pipes
(PCCP) all of 122 cm diameter were assessed either from inside, having the pipes
fully drained, or from outside by digging up test pits. The sonic/ultrasonic sensor
identified 26 pipes with wall thickness loss, as the resonate frequency shifted. For
122 cm pipes with a thickness of 10 cm, the normal resonate frequency was
20 24 kHz; however, for thinner walls measurements of 29 30 kHz were
recorded. To validate the findings, the 26 pipes were excavated in those places
where the sonic/ultrasonic measurements had detected wall thinning; all findings
were confirmed. However, no quantitative measures were provided.
Horoshenkov et al. (2004, 2011) developed a quick and low-cost acoustic
method of pipe inspection using a sinusoidal chirp sound transmitted in the
range 100 Hz 20 kHz and applying methods of signal reflectometry, sound
attenuation and Fourier Transform standard deconvolution. The acoustic sensor
consists of a speaker and a microphone array which are attached to a pole that
is stationed in a manhole below the pipe soffit. The cable from the sensor is
attached to an electronic box leading to a portable computer from where the sen-
sor is activated; the survey results are instantly displayed and stored. Undertaken
sets of laboratory and field trials on clay, concrete, PVC, and Perspex pipes of
various diameters, and including a number of defects, conclude that the acoustic
sensor is able to detect large defects such as blockages, roots, missing pipe
segments, cracks, fat build-up, and sediment, as well as small pipe wall perme-
ability contrasts such as microcracks, encrustation, and pipe wall roughness (as
corrosion). It was reported that the software defect recognition “signature” data-
base can be built up and trained; however no physical measure of the condition
scale is reported.