Dynamic programming (DP) has a fifty-year history of optimizing pipeline operations. While this was originally limited to deciding pump lineups and pump and valve setpoints in single steady states, the increasing power of modern computers makes it possible to use this technique to optimize more aspects of liquid pipeline performance.
This paper builds on the traditional use of DP - to determine pump lineups and pressure setpoints - first expanding it to optimize for flow rates considering different peak/off-peak energy costs. The author then develops algorithms that use this pressure-and-flow-rate solver to optimize for complex power contracts including demand terms, and to optimize drag-reducing additive (DRA) usage with full consideration of the DRA’s exact effects as it moves through the pipeline and degrades.
