Decomposition strategy for natural gas production network design under uncertainty

The use of natural gas for power generation has been rising rapidly in the past two decades. To ensure the security of supply of gas to the market and meet strict specifications on gas quality (e.g., sulfur content), natural gas production network design must address uncertainty explicitly as well as tracking the quality of each gas flow in the entire system. This leads to the stochastic pooling problem, which is a (potentially large-scale) nonconvex mixed-integer nonlinear program (MINLP). This paper presents a rigorous, duality-based decomposition strategy to solve the stochastic pooling problem, which guarantees finding an ε-optimal solution of the problem with a finite number of iterations. A case study involving a gas production network demonstrates the dramatic computational advantages of the decomposition method over a state-of-the-art global optimization method. The proposed method can be extended to tackle more general nonconvex MINLP problems, which may occur in the design of integrated energy systems involving fuel production, power generation and electricity transmission.