Data used in simulation is researched, derived, calculated, estimated, measured, or simply guessed at for the majority of pipeline physical parameters. In pipeline simulation, first-time users often find themselves in a quagmire of data gathering exercises. They spend many valuable hours researching •as built• documentation and performing laborious calculations with the expectation of constructing a highly accurate pipeline model. This paper is intended to help the new, and perhaps, the experienced modeler cut through what is important and what is superfluous.
What is frequently not clear is how much of the collected and collated data is actually necessary and what level of accuracy. Additionally, how data quality will affect the simulation is often not recognized early in the process. At the initial stage of data gathering, other important decisions must also be made; these decisions can affect, or be affected by, the practice of simulation within an organization. Examples are the selection of the equation of state, the friction factor equation, or the reference conditions.
This paper discusses these issues by describing an approach that allows model sensitivity to be performed at an early enough stage in the modeling process to ensure that the key parameters are well understood. This process is demonstrated through examples of both gas and liquid pipeline simulations. Various "universal" parameters are presented.
Data required for pipeline simulation can be categorized in general types.
The physical geometry of the pipeline, such as leg lengths, diameters and elevation profile.
Fluids properties, for example gas compositions or liquid physical properties such as density and viscosity
Material properties of the pipeline and its environment, such as thermal coefficients and yield stresses.
The data required for each of these types may be further differentiated. In some cases small data variation leads to large discrepancies, whilst in other large differences yield smaller variations. During this tutorial generic forms of pipeline simulation equations will be presented, Data variations are examined and discussed. In some cases filtering data with the aim of simplifying the output may be appropriate. During this tutorial we will primarily be looking at data requirements for offline simulation. For real time simulation where detailed model hydraulics may be necessary, particularly for leak location applications, a different and more detailed set of data is required.
Pipeline simulation software normally uses a node/equipment description to configure the model of the pipeline. A node is simply the connection point between pieces of equipment. Nodes have no physical dimension, although set-points or constraints may be configured at a node.