Steady-State Thermohydraulic Optimization of Natural Gas Transmission Networks with MILP-Embedded Compressor Surrogates and Adaptive-Neighborhood Prediction-Correction
Industrial & Engineering Chemistry Research · 05.08.2026 · 65 (30) · 16173-16195
DOI: 10.1021/acs.iecr.6c01626 ↗Kurz gefasst
Fachartikel in der Zeitschrift Industrial & Engineering Chemistry Research, online erschienen am 26.07.2026. Verfasst von Jun Zhou und 5 weiteren Autor:innen. Die Arbeit nennt 54 Literaturangaben. Gefördert von: National Natural Science Foundation of China.
Abstract (Originalsprache)
Steady-state optimization of natural gas transmission networks requires a balance between the computational tractability and thermohydraulic fidelity. Existing tractable formulations often simplify either network-side physics or compressor unit behavior, which may lead to schedules that are difficult to verify under detailed thermodynamic and hydraulic calculations. This study develops a selective hybrid thermohydraulic optimization framework for large-scale steady-state gas transmission networks. Pipeline heat transfer, pressure-dependent compressibility, implicit friction, nodal temperature mixing, and valve hydraulic/thermal relations are retained in mechanistic form, whereas compressor-unit speed, power, discharge temperature, and surge/choke-envelope relations are represented by mixed-integer linear programming (MILP)-embedded rectified linear unit (ReLU) surrogates. The formulation combines the surrogate-embedded compressor model with a pressure-squared hydraulic relaxation and a pressure-dissipation regularization term, and the tightness of the relaxed pressure-drop relations is evaluated through post-solution residual checks. To solve the resulting mixed-integer model, an adaptive-neighborhood prediction-correction framework (ANPCF) is developed as a computational heuristic that coordinates hydraulic initialization, fixed-parameter scheduling, and physical-state recovery through adaptive damping and adaptive neighborhood updates. Tests on 45-, 135-, and 582-node networks show that the proposed model reduces the Case 1 nodal-pressure mean absolute percentage error from 0.72% to 0.33% relative to a simplified baseline and solves the nominal 582-node case in 102.41 s under the default initialization setting. The results indicate that selective compressor surrogate embedding and prediction-correction coordination can improve thermohydraulic fidelity while maintaining practical computational performance for large steady-state studies.
Förderung
- National Natural Science Foundation of China
Zitieren
APA
Zhou, J., Qin, C., Liang, G., He, Y., Liu, S., & Peng, J. (2026). Steady-State Thermohydraulic Optimization of Natural Gas Transmission Networks with MILP-Embedded Compressor Surrogates and Adaptive-Neighborhood Prediction-Correction. Industrial & Engineering Chemistry Research, 65(30), 16173-16195. https://doi.org/10.1021/acs.iecr.6c01626
Vancouver
Zhou J, Qin C, Liang G, He Y, Liu S, Peng J. Steady-State Thermohydraulic Optimization of Natural Gas Transmission Networks with MILP-Embedded Compressor Surrogates and Adaptive-Neighborhood Prediction-Correction. Industrial & Engineering Chemistry Research. 2026;65(30):16173-16195. doi:10.1021/acs.iecr.6c01626
BibTeX
@article{Zhou2026,
author = {Zhou, Jun and Qin, Can and Liang, Guangchuan and He, Ying and Liu, Shitao and Peng, Jinghong},
title = {Steady-State Thermohydraulic Optimization of Natural Gas Transmission Networks with MILP-Embedded Compressor Surrogates and Adaptive-Neighborhood Prediction-Correction},
journal = {Industrial & Engineering Chemistry Research},
year = {2026},
volume = {65},
number = {30},
pages = {16173--16195},
doi = {10.1021/acs.iecr.6c01626}
}
EUVIMED