Research on operation optimization technology for underground natural gas storage

To maximize economic benefits while ensuring peak shaving and supply security, this paper proposes a practical energy consumption optimization method under the dual constraints of "supply guarantee and cost control" during the gas injection period. A business-finance integration framework-comprising standard costing, energy consumption prediction, and parameter optimization-was developed alongside a standard cost system for underground gas storage clusters to define key equipment cost components. Principal Component Analysis(PCA)was utilized to identify critical characteristic parameters, including inlet / outlet pressure, daily injection volume, and operation time. Finally, a business-finance model was established using optimized Support Vector Regression(SVR)combined with Particle Swarm Optimization(PSO)machine learning method to integrate equipment energy consumption, production data, and cost elements. This model enables real-time prediction of energy consumption costs across different injection-production schemes and determines optimal compressor parameter mix for various tasks. By providing a replicable parameter tuning scheme for gas storage operation optimization without altering existing technological processes, this method facilitates refined cost control and efficiency enhancement.