Review of the analysis and methods of natural gas pipeline network operation optimization

The expansion of natural gas consumption and pipeline construction makes integrating artificial intelligence into pipeline network operations increasingly essential. This review summarizes progress in operation optimization, gas transmission capacity evaluation, and solution algorithms. The review systematically summarizes objective functions, hydraulic/thermal and compressor constraints, and decision variables, all framed by operator objectives such as transmission capacity, economic benefits, and supply reliability. It highlights gas transmission capacity optimization and extended models, including those for hydrogen-blended and renewable energy-coupled scenarios. The review also analyzes applications of deterministic and stochastic intelligent algorithms, alongside deep learning and hyper-heuristic methods. Key findings indicate that: (1) Traditional models often lack safety, reliability, and low-carbon indicators; (2) Deterministic algorithms struggle with high dimensionality, while heuristic algorithms are prone to premature convergence; (3) Hydrogen blending and new energy integration necessitate revised constraints; and (4) Existing online dynamic optimization methods are insufficient. Finally, current shortcomings are identified, and future directions, such as advanced online dynamic optimization and cross-domain intelligence, are proposed. In conclusion, while artificial intelligence is crucial for natural gas pipeline network operations, significant limitations persist. Future research must prioritize addressing these gaps to advance the industry's intelligent, low-carbon, and reliable development.