Abstract

This paper presents several studies designed to explore low-cost methods of rupture detection on a natural gas transmission system. In each method, Planning ran transient models with scheduled ruptures and analyzed the resulting pressure and/or flow data. The results reveal that reliability and sensitivity of rupture detection packages increase with the amount of capital spent on them. The more poignant revelation is that we should not rely on these methods alone to detect and accurately locate all ruptures. Continued investigation of inexpensive software-based rupture detection methods will likely reveal solutions that are more robust. 3 Introduction If there were an ability to monitor gas pipeline conditions to predict ruptures, then fewer ruptures caused by internal pipeline conditions would result. Until that technology becomes more readily available, the focus remains on detecting ruptures as quickly as possible, thereby minimizing linepack loss. Pipeline planning engineers have been exploring methods of rupture and leak detection for more than twenty years. Our purpose is to explore low-cost methods of rupture detection. The three methods we studied primarily utilized software-based rupture detection. The first method evaluates how the quantity of pressure measurement and auto-closure devices affect the time elapsed in detecting a rupture. Our second method assesses the effectiveness of rules established to detect flow rate of change and pressure drop in response to a rupture. Our third method analyzes a rupture detection package utilized to examine the advantage of flow measurement at the discharge header of each station.

Assumptions

We made some assumptions in these studies to facilitate our analysis, some of which are unlikely to occur in typical day-to-day operations. Some of the assumptions varied in each study and are described in each study's methodology.

Description of Model Pipeline

We modeled each intra-station segment with adjacent pipeline. Each simulation model consisted of the components shown in Figure 1 with solid lines. We focused primarily on the behavior of the pipeline following a simulated rupture occurring somewhere between the discharge of Station B and the suction of Station C. Our models, however, extended from the discharge of the station upstream of Station B to the suction of the station downstream of Station C. The inclusion of this additional pipeline gives a better approximation of the actual pipeline behavior. We used this model for all three methods.

4 Figure 1. Graphical Representation of our Simulation ModelReceipts and Deliveries

We held gas receipts and deliveries constant Ruptures For the ruptures, we simulated a complete line break. In essence, this is equivalent to two lines discharging to atmosphere, see Figure 2. We placed at least one rupture on each loop line near the discharge, middle, and suction of the section between Station B and Station C. This allows for various distances from pressure measurement.

Procedure, Results and Analysis Model Building

For each line segment, we built a transient simulation model similar to Figure 1, containing a full physical description of the pipeline components (e.g., the length, diameter, roughness, and elevation of each span of pipe).

This content is only available via PDF.
You can access this article if you purchase or spend a download.