A Non-Sequential Dynamic Programming Approach for Natural Gas Network Optimization

Abstract: The fuel cost minimization problem on steady-state natural gas pipeline networks system is addressed. From the optimization perspective, this problem is modeled as a (non-convex) nonlinear program (NLP), where we consider two types of decision continuous variables: mass flow rate through each arc and pressure value at each node. The proposed method consists of two phases. In phase 1, a set of feasible flows is found by a reduction technique, which makes use of a pre-processing procedure. Then, in phase 2, an optimal set of pressures is found (for the given flow) by applying a non-sequential dynamic programming technique. This method avoids the many numerical difficulties inherent to this very complex while treated with classical nonlinear programming techniques. We work with several different types of topologies, many of those being cyclic structures. A computational study reveals the effectiveness of the proposed procedure when tested over a wide variety of problem instances.