465693 Pipe-Flows Leak Detection, Size Estimation and Localization

Wednesday, November 16, 2016: 1:24 PM
Carmel II (Hotel Nikko San Francisco)
Xiaodong Xu and Stevan Dubljevic, Chemical and Materials Engineering, University of Alberta, Edmonton, AB, Canada

Monitoring of pipelines integrity is important for fluid transportation since it is necessary in

timely manner to detect malfunctioning, e.g. leaks. In particular, for the petroleum industry,

leaks from pipelines may potentially result in environmental damage and pollution, as well as to

the economic loss. These motivating factors with requirement from environmental authorities

leads to the development of ecient leak detection systems. However, some leak detection

methods are hardware-based, where the leak detection techniques rely on physical equipment

being installed along the pipeline. In this work, we focus on software-based approaches that

work for cases with limited instrumentation. In particular, the instrumentation is restricted

to pressure and ow measurements at the pipeline inlet and outlet.

Some software-based leak detection methods utilize statistical analysis on measurements

(black box), while others are on the basis of mathematic models from physical principles. The

approach in this work belongs to the latter category. For more details about leak detection

techniques, we refer to a recent survey [1]. In this work, a dynamic model of the pipe

described by a set of two coupled hyperbolic partial differential equations (PDEs) is utilized to

realize leak detection system. In [2] and [3], a bank of observers are used to potentially detect

multiple leaks. However, multiple simultaneous leaks may not likely happen. Therefore, we

employ the idea from internal model control and adaptive control, addressing the size and

location estimation of a single point leak based on a set of two coupled one dimensional first

order hyperbolic PDEs. Moreover, the backstepping transformation is applied to simplify the

design of the observer for the size estimation and the observer gains can be obtained in closed-

form. Finally, simulations are provided to show the ability of the proposed system to detect,

quantify and locate leaks.

 [1] P.-S. Murvay and I. Silea (2012) A survey on gas leak detection and local-

ization techniques, J. Loss Prevent. Process Ind., vol. 25, pp. 966--973.

 [2] C. Verde (2001) Multi-leak detection and isolation in fluid pipelines, Control

Eng. Pract., vol. 9, pp. 673--682.

 [3] C. Verde (2004) Minimal order nonlinear observer for leak detection, J.

Dynam. Syst., Measur., Control, vol. 126, pp. 467--472.

Extended Abstract: File Not Uploaded
See more of this Session: Process Monitoring and Fault Detection
See more of this Group/Topical: Computing and Systems Technology Division