THESIS
2018
xiii, 99 pages, 2 unnumbered pages : illustrations (some color) ; 30 cm
Abstract
Water conveyance are the lifeline of 3 billion people globally; however, these vital systems
are aging and fraught with defects, such as leakages and blockages, due to physical and/or
chemical processes. These defects result in wastage of energy and financial resources, reduction
in carrying capacity, and increased potential for contamination. In order to conduct
accurate leak detection, people should first understand the particular change induced by the
leak inside out and then detect the leak using this particular information. In this dissertation,
both of the above study are conducted. The understanding of the impact induced by the
leak is called forward problem, while the detection of leak using the particular information
induced by the leak is called inverse problem. The fo...[
Read more ]
Water conveyance are the lifeline of 3 billion people globally; however, these vital systems
are aging and fraught with defects, such as leakages and blockages, due to physical and/or
chemical processes. These defects result in wastage of energy and financial resources, reduction
in carrying capacity, and increased potential for contamination. In order to conduct
accurate leak detection, people should first understand the particular change induced by the
leak inside out and then detect the leak using this particular information. In this dissertation,
both of the above study are conducted. The understanding of the impact induced by the
leak is called forward problem, while the detection of leak using the particular information
induced by the leak is called inverse problem. The forward problem focuses on the impact
of leak on the resonant frequencies of the pipeline. The inverse problem promotes an leak
detection method, which is called arrival time imaging.
This analytical and numerical research investigates from a basic physics perspective of
the leaks and the controversy regarding their effect on the system response function of the
pipeline.
This research explores two distinct aspects: (i) it clarifies an complementary knowledge
about the application of leak detection method based on system response function in the frequency
domain; (ii) proposes a new defect detection approach based on the system response
function in the time domain.
For the complementary knowledge on the application of the leakage detection based
on system response function in frequency domain, the impact of a leak on the resonant
frequencies is investigated. Trajectories of resonant frequencies in the wavenumber complex
plane are studied with varying leak size. The variation in the system response function in
the frequency domain and at system resonant frequencies is discussed. The key parameter
representing the leak size and controlling trajectories of the resonant frequencies is the ratio
of pipe impedance to leak impedance. It is found that, as the impedance ratio becomes
greater than zero (i.e., no leak), each normal resonant mode shifts towards the upper half
of the complex plane in response to a leak, where the imaginary part stands for the leak-induced
damping of the wave. When the impedance ratio is less than the order of one, the leak-induced shift in the normal mode is negligible, which supports the theory put forward
by proponents of the hypothesis that leak does not induce additional peaks to the frequency
response function function (FRF). When the impedance ratio is of order one or larger, not
only the shift in FRF’s peak is significant, but new peaks also appear, which supports the
theory raised by proponents of the hypothesis that leak induce additional peaks to the FRF.
Moreover, as the impedance ratio tends to infinity, the leak behaves as a reservoir where the
pipe is fully open to the atmosphere. The analysis presented in this thesis provides insights
into the signature of a leak on normal resonant modes of the pipeline ranging from small
leak size to extremely large leak size, which clarify and resolve the previous controversy
regarding the leak’s impact on the frequency response of the pipeline.
For the development of the new defect detection approach based on the system response
function in time domain, an arrival time imaging (ATI) method is developed and presented
that utilizes the reflection information of the transient wave, is proposed. This method
translates the time-domain impulse response signal into a space domain objective function
along the pipe, which is able to combine the information obtained from multiple sensors
and helps visualize the defects more directly and accurately. A boundary reflection erasure
technique is carried out, which helps obtain a clearer objective function from the influence of
reflections at the pipe boundary. Numerical simulation shows that the proposed ATI method
can estimate leakage/blockage in a noisy environment for signal-to-noise ratios (SNR) as
low as -4 dB. It also shows that the ATI method is able to detect leakages and blockages
(multiple defects) simultaneously without the need to model leakage or blockage. This
method does not require the use of an accurate simulation model or a leak-free benchmark.
Knowledge of the pipe topology, flow and roughness values, or the role of unsteady friction
on the transient event is unnecessary.
The investigation of a leak’s impact on the normal resonant modes of the pipeline provides
improved understanding of the leak mechanism in frequency domain and raises the
potential for further improvements in frequency domain leak detection methods. The new
method of multiple defect detection significantly improves the present time domain reflectometry
method and is proved to increase the accuracy of defect detection predictions.
Post a Comment