基于动态规划和黄金分割法的环状天然气管网运行优化

目前中国天然气管道主干线已经相互联接,形成了环套环的大型复杂天然气管网,随着各压气站燃驱压缩机组的增加,如何实现管网的稳态优化运行已成为一个技术难题。为此,在同时考虑管网中环状结构和枝状结构对于管道流量分配的影响,以及燃驱机组耗气量对管道内流量影响的基础上,结合管网模型分割思路,将黄金分割算法与动态规划算法进行融合,形成了改进后的动态规划和黄金分割混合算法。以华北地区某环状管网为案例的计算结果表明:①管网运行优化数学模型比单线管道模型增加了管道内过流量、燃驱耗气量等变量,进一步加大了管网模型的求解难度;②混合算法能够有效地对运行燃驱和电驱压缩机机组的包含环状和枝状小型管网的稳态运行优化数学模型进行求解;③混合算法的优化方案能提高管网压气站的出口压力、提高机组的运行效率,降低系统的功率消耗、减少系统能耗,同时增加管网的管存量,提供了管网额外的应急资源。结论认为,混合算法有效地破解了复杂管网优化技术难题,可以为当前中国大型天然气管网稳态运行优化提供参考。

Operation optimization of a looped natural gas pipeline network based on dynamic programming and the golden section method

At present, China's natural gas trunk pipelines have been connected with each other to form a large-scale complex natural gas pipeline network of ring by ring. With the increase of gas-driven compressor sets in each compressor station, the steady-state operation optimization of the pipeline network is now a technical problem. In this regard, this paper considered both the effect of looped structure and dendritic structure in the pipeline network on the distribution of pipeline flow rate and the effect of the gas consumption rate of gas-driven compressors on the flow rate inside the pipeline. Then, according to the segmentation idea of the pipeline network model, the golden section algorithm and the dynamic programming algorithm were integrated to form an improved hybrid algorithm of dynamic programming and golden section. The following calculation results were obtained with one certain looped pipeline network in the area of North China as the example. First, different form the single-line pipeline model, the mathematical model of pipeline network operation optimization has additional variables, e.g. the flow rate inside the pipeline and the gas consumption rate of gas-driven compressor, which increase the solution difficulty of the pipeline network model further. Second, the hybrid algorithm can effectively solve the mathematical model of steady-state operation optimization of looped and dendritic small pipeline networks with gas- and electric-drive compressor sets. Third, the optimization scheme of hybrid algorithm can increase the outlet pressure of the compressor station, enhance the operation efficiency of the compressor set, and reduce system power and energy consumption while increasing the stock of pipeline network and providing the extra emergency resources of a pipeline network. In conclusion, the hybrid algorithm can effectively solve the technical difficulties of complex pipeline network optimization and can provide reference for the steady-state operation optimization of large-scale natural gas pipeline networks in China.