Magazine

A Novel Neural Model-Based Approach to Leak Detection and Localization in Oil Pipelines for Environmental Protection

Description
Monitoring oil transporting pipelines is an important task for economical and safe operation, loss prevention, and environmental protection from crude oil emission. The leak detection of oil pipelines, therefore, plays a key role in the overall
Categories
Published
of 5
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
Share
Transcript
  A Novel Neural Model-Based Approach to Leak Detection and Localization in Oil Pipelines for Environmental Protection ALIREZA PAIVAR Department of Instrumentation and Automation M.S. - Petroleum University of Technology Tehran-Iran KARIM SALAHSHOOR Department of Instrumentation and Automation PhD- Associate Professor.-Petroleum University of Technology Tehran-Iran Farzad Hourfar Department of Instrumentation and Automation M.S. - Petroleum University of Technology Tehran-Iran  Abstract  : - Monitoring oil transporting pipelines is an important task for economical and safe operation, loss prevention, and environmental protection from crude oil emission. The leak detection of oil pipelines, therefore, plays a key role in the overall integrity monitoring of a pipeline system. This paper proposes a neural decision-making approach to oil pipeline leak localization. The one main methods, model based fault detection is used (to find leaks quantity and location-making) to form a novel fault diagnosis scheme. This scheme can improve the precision of localization. An application example, 600m long oil pipeline leak detection and localization are illustrated, and the effectiveness of the proposed approach is demonstrated using practical results. Key-Words: -   Environmental and Safety Systems, Fault and Uncertainty Modeling in Dynamical systems, Process Supervision, Neural Nets 1 Introduction In recent years, there is an increasing demand on the safe operation and leak detection of oil transporting pipelines. , a leak detection system is an indispensable condition to allow its construction. In order to satisfy this demand, more and more methods are developed. In general, this is a question of fault diagnosis. There are two main kinds of available methods for pipeline leak detection and localization presented in the literature. The first is called the observer method [1],[2] , namely the pressure gradient method. It is   sensible to weak leak but not applicable to serious leak. It has distinct characteristics that this method has to install many sensors along the pipeline [3]. The second is called negative pressure wave method [4]. It is sensible to serious leak but not applicable to weak leak. It has a distinct advantage that each of the sending and receiving terminals in the pipeline only needs to install one sensor.   We propose a new localization method for the leak of oil pipeline in this paper, which we call diagnosing and locating approach based on neural model (NMLD). This method only need to install two sensors at each terminal of Proceedings of the 5th WSEAS Int. Conf. on CIRCUITS, SYSTEMS, ELECTRONICS, CONTROL & SIGNAL PROCESSING, Dallas, USA, November 1-3, 2006 291  pipeline, and can improve greatly the precision of localization. 2 Pipeline Model Research has been undertaken into mathematical equations governing the pressure and flow of liquid in pipelines. Some popular models and mathematical equations are given in [1],[2],[3].   The mathematical model of pipeline is as bellow that is derived from momentum and continuity equations. (1) Where P=pressure V=discharge velocity D=diameter of pipe A=cross sectional area of pipe G=gravitiaonal acceleration f=Darcy-Weisbach friction factor z=distance t=time α  =upward angle between the pipe and horizontal a=wave speed of liquid system  ρ   =fluid mass density If the flow were steady, the pressure loss due to fluid friction would be modeled by the familiar 22 2  DAQ f  p steady  f   ρ  =  (2) that give us the pressure loss along the pipe. 3 NMLD Leak Monitoring Concept Our leak monitoring requires a pipeline modeled by neural net in the form of a pipeline observer to compute the pipeline states assuming no leak. Further discussion will be focused on a discrete-time data processing scheme. The difference between the measured and estimated flow at inlet and outlet . ),(),()( ),(),()( ^^ k  Lvk  Lvk  y k ovk ovk  x −=−=  (3) where v(0,k) and v^(0, k) are the measured flow and the flow of the observer at the inlet of the pipeline, while v(L,k) and v^(L,k)) are the measured flow and the flow of the observer at the outlet of the pipeline, will be referred to as residuals. The leak flow and position can be estimated by  pleak leak   Lk  yk  x k  yk  xk  yk  xv ×−−=−= )()( )()()()(  (4) Where denotes the length of pipeline [4] This leak detection and localization scheme requires a pipeline model to compute the states of a pipeline without a leak. The basic mathematical model of a pipeline is a nonlinear distributed parameter model. It describes the one-dimensional compressible fluid flow through the pipeline and is represented by a set of nonlinear partial differential equations [1]. No general closed-form solution of these equations is known yet. Numerical approaches like the Method of characteristics must be used instead [1]. In Fig 1 all basic parts of the described approach to the leak detection and classification can be seen. Proceedings of the 5th WSEAS Int. Conf. on CIRCUITS, SYSTEMS, ELECTRONICS, CONTROL & SIGNAL PROCESSING, Dallas, USA, November 1-3, 2006 292    Fig1-Leak Detector Scheme as we can see the use of mathematical model is difficult as it is complicated mathematical equation. then neural nets appears best to model the pipe. we derived pipeline data from Hysis software. For modeling of the pipeline one RBF neural net was used to model the pipeline. The quantity of neurons about 30 appears good after class validations of data. The very important think is that after modeling we see that we model the pipe with differential pressure of pipeline by using equation (3).then our network appears to be 1*30*1 net. 4 Application to Real Pipeline We simulated a real pipeline located in Iran by Hysis software. Our pipe is around 600m. simulations. Our fluid was crude oil. The simulation appears successful. In below figure we can see the detection of false leaks. In different location in the pipe. Fig2-Leak Estimation for Varied Leak locates. The accuracy of the leakage estimates depend on the magnitude of the leakage. The minimum magnitude of leakage that may be realized through estimation depends on the resolution of measurement devices. The accuracy of the simulation model will also affect the size of leak that may be detected. In a real system, a model calibration would be done to ensure some level of correlation between measurement and simulation. The degree of fitness between measurement and simulation will give some indication as to the magnitude of the estimates that may be estimated.   Five simulations of leakage at 300 meters from inlet of the pipeline. The magnitude of the leakage was varied: 1, 2, 5, 10 and 20% leakage flows were simulated. Figures (3) and (4) display the results of these estimations.   Proceedings of the 5th WSEAS Int. Conf. on CIRCUITS, SYSTEMS, ELECTRONICS, CONTROL & SIGNAL PROCESSING, Dallas, USA, November 1-3, 2006 293    Fig3- Leak Estimation for Varied Leak Quantity. Fig4-Leak Estimation for Varied Leak Magnitudes 5 Conclusions The technique described in this paper is a tool that could be implemented by itself or in conjunction with other leak detection tools. Based on the results of this study, the following conclusions were drawn. It may be concluded that the methods described in this paper present a successful method for quickly identifying leakage. The simulations displayed promising results and were capable of detecting the location of small leakages. However, this method is limited in its ability since only one leak may be detected, and if multiple leakages occur they may be falsely diagnosed as a single leak. The methods developed are best suited for long transmission lines or sections of distribution systems. The very important thing is that this method only uses two pairs of instruments at the inlet and outlet of the pipe and is very cheap method. This method could be implemented in process control and monitoring equipments.    References [1] Anderson J.D. Basic Philosophy of CFD. In J.F.Wendt, editor, Computational Fluid Dynamics. Springer-Verlag, Berlin Heidel- berg, 2nd edition, 1996. [2] Billmann L. Methoden zur Leckueberwachung und Regelung von Gasfernleitungen . TH Darmstadt, 1985 . [3] Geiger G. Leak monitoring in pipelines - state of the art. Seminar Fernleitungen der Landeanstalt fuer  Arbeitschutz, Kaiserau 3- 5.11.1999, 1999. [4] Geiger G. Application of a model-based method for leak detection and localization. GMA-Kongress Me_- und  Automatisierung- stechnik 18./19.  Juni 1988,VDI Berichte  Nr.1397,, 1998. [5] Isermann R. Integration of fault detection and diagnosis methods.  IFAC Symposium SAFEPROCESS, Espoo, Finland, 1994. [6] Rao C.V. and K.Eswaran. On the Analysis of Pressure Transientsin Pipelines Carrying Compressible Fluids.  Int. J. Pres. Ves. and Piping, 56:107{129, 1993. Proceedings of the 5th WSEAS Int. Conf. on CIRCUITS, SYSTEMS, ELECTRONICS, CONTROL & SIGNAL PROCESSING, Dallas, USA, November 1-3, 2006 294  [7]Fasol K.H. and Pohl G.M.(1990) Simulation controller design field tests for hydropower plant-A case study –  Automatica Vol26 [8]Gertler J.(1998)Fault detection and diagnosis in engineering systems-New Y ork [9]Guy J.J(1967) Computation of unsteady gas flow in pipe networks Conference of chemical engineering. London ,No,23 [10]Gunawickrama K.(2001) Leak detection methods for transmission pipelines Faculty of electronics ,Technical university of Gdansk Proceedings of the 5th WSEAS Int. Conf. on CIRCUITS, SYSTEMS, ELECTRONICS, CONTROL & SIGNAL PROCESSING, Dallas, USA, November 1-3, 2006 295
Search
Similar documents
Tags
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks