Optimization techniques can be used to generate, in advance, control strategies for efficiently operating pipelines through periods of fluctuating loads. A control strategy is simply a specific schedule for changing compressor station set point values during a particular scenario period. The methods are particularly valuable when used to determine how to position line-pack resources to meet upcoming stressful conditions, such as the onset of electric power plant loads, or even loads temporarily in excess of sustained pipeline capacity.[1][2] We present extensions to these methods that address the question of future uncertainty in load patterns. Rather than consider a single possible future load scenario, the methods allow inclusion of several scenarios simultaneously. For instance, one scenario might consider the case where no power plants come online, a second might consider plants coming online for 6 hours on one part of the system, and a third might consider a different power plant coming on in a remote part of the system but with an uncertain onset time. Optimal pack management strategies for each particular expected load considered individually may differ widely from the optimal strategy needed to prepare for meeting any one of several possible loads. Therefore, in our new approach, rather than committing at the outset to one possible future in generating a control plan, we generate one that positions line-pack defensively before the possible loads so as to be able to meet any of the included future load pattern scenarios efficiently. The concept is analogous to "hedging" in the financial world. The computed set point schedule will, at each point in time, position line-pack in the proper strategic distribution to allow operators to dependably and efficiently meet whichever load pattern actually occurs. We present examples of this hedging with several scenarios.
The goal of a transmission pipeline is to move gas from suppliers to end customers in contracted amounts, dependably and with minimum complications. Two of the major complications in operations and planning are fluctuating pipeline conditions, and uncertainty about the future. Fluctuating pipeline conditions are easy to conceptualize, although sometimes difficult to manage. Tomorrow's customer loads may be distributed differently from today's loads due to changing weather and market conditions. Equipment availability may be different tomorrow due to maintenance or outages. Ambient temperature (and hence compression capacity) may be different tomorrow. All of these examples imply that one may need a different operational plan for tomorrow than was used today in order to operate efficiently, profitably, and safely. Tomorrow's operational plan is thus often generated using steady state optimization tools loaded with tomorrow's expected (average) conditions. Having a different plan for tomorrow also implies that you need to be able to transition the system from today's conditions to tomorrow's conditions, dependably and with minimum complications. This is complicated by the fact that some customer loads do not typically ramp smoothly from "today's" state to "tomorrow's" state but may fluctuate widely from hour to hour.
