Summary
An adaptive Luenberger-type estimator for the purpose of monitoring flow
conditions and locating and quantifying leakages in petroleum pipelines is
presented. The estimator only needs measurements of velocity, pressure, and
temperature at the inlet and velocity and pressure at the outlet to function.
The measurements are used to form a special set of boundary conditions for the
estimator that ensures fast convergence of the estimation error. Depending only
on measurements from inlet and outlet makes it possible to use OLGA, which is a
state-of-the-art computational fluid dynamics simulator, to govern the
one-phase fluid flow of the estimator. The estimator is tested with both a
straight, horizontal pipeline and an actual, long pipeline with inclinations,
and both simulations with oil and gas are carried out. In order to cope with
modeling errors and biased measurements, estimation of roughness in the
monitored pipeline is introduced.
Introduction
The benefits of a leak-detection system capable of locating the position of
the leak are obviously of an environmental kind. But the economical aspect of
it is also important. Leak detection based on dynamic modeling is a propitious
approach in the special field of leak quantification and location. There have
been numerous studies on model-based leak detection. We mention here the most
relevant ones with regard to our work. Based on a discretized pipe flow model,
Billmann and Isermann (1987) designed an estimator with friction adaptation. In
the event of a leak, the outputs from the estimator differ from the
measurements, and this is exploited in a correlation technique that detects,
quantifies and locates the leak. Verde (2001) used a bank of estimators,
computed by the method for fault detection and isolation developed by Hou and
Müller (1994). The underlying model is a linearized, discretized pipe flow
model on a grid of N nodes. The estimators are designed in such a way that all
but one will react to a leak. Which one of the N estimators that does not react
to the leak depends on the position of the leak, and this is the mechanism by
which the leak is located. The outputs of the remaining estimators are used for
quantifying the leak. The bank of estimators are computed using the recursive
numerical procedure suggested by Hou and Müller (1994); however, it was shown
in Salvesen (2005) that because of the simple structure of the discretized
model, the estimators may be written explicitly. This is important, because it
removes the need for recomputing the bank of estimators when the operating
point of the pipeline is changed. Verde (2004) also proposed a nonlinear
version, using an extremely coarse discretization grid. Several companies offer
commercial solutions to pipeline monitoring with leak detection. Fantoft (2005)
uses a transient model approach in conjunction with the commercial pipeline
simulator OLGA2000, while EFA Technologies (1987, 1990, 1991) uses an event
detection method that looks for signatures of no-leak to leak transitions in
the measurements. The detection method of Verde (2001) using a bank of
estimators can potentially detect multiple leaks. However, multiple
simultaneous leaks is an unlikely event, so the complex structure of a bank of
N estimators seems unnecessary. Aamo et al. (2006) instead employed ideas from
adaptive control, treating the mass rate and location of a single point leak as
constant unknown parameters in an adaptive Luenberger-type estimator based on a
set of two coupled 1D first-order nonlinear hyperbolic partial differential
equations. Heuristic update laws for adaptation of the friction coefficient,
mass rate of the leak, and position of the leak were suggested. The method was
developed further by Hauge et al. (2007) who remodeled the leak as a pressure-
and density-dependent function, thereby improving the leak-detection capability
during transient flow such as for instance pipeline shut-down.
In this paper, we continue the development by employing the state-of-the-art
multiphase flow simulator OLGA as the underlying flow model, enabling more
accurate flow predictions for complex pipeline conditions and thereby further
improving the leak detection capability of our approach. The simulator is
manipulated through MATLAB to work as an adaptive Luenberger-type estimator
using measurements from the supervised pipeline. Because the only measurements
fed into the estimator are velocity and pressure from the outlet and velocity,
pressure, and temperature from the inlet, which in most cases already are
available, the method can be used with most existing pipelines without
additional instrumentation. Also, the OLGA simulator, which is the backbone of
the estimator, is widely used in the petroleum industry. Our approach allows
for easily turning an existing OLGA model of a real pipeline into a monitoring
system, including leak-detection capabilities.
© 2009. Society of Petroleum Engineers
View full textPDF
(
1,271 KB
)
History
- Original manuscript received:
31 October 2007
- Revised manuscript received:
20 February 2009
- Manuscript approved:
7 March 2009
- Published online:
17 September 2009
- Version of record:
21 September 2009
View Cart
