LNG储罐内部压力监测与控制模型设计

Design of Internal Pressure Monitoring and Control Model for LNG Storage Tanks

摘要:

针对LNG储罐在复杂气液耦合工况下面临的压力监测滞后与控制精度不足的问题，本研究基于VOF多相流模型，开展了一系列监测与控制方法的深入研究。通过耦合VOF气液界面动态追踪机制、MPC的滚动优化策略以及DDPG算法的扰动补偿能力，构建了一套“监测–优化–补偿”一体化智能控制架构。采用VOF-WTD-WPE多尺度融合方法，实现了机理模型与实时监测数据的协同感知；结合MPC-DDPG协同优化框架，有效处理多约束条件与非线性扰动。实验测试结果表明，所提模型在压力监测误差均值方面降低85.2%，稳态波动范围缩小53.3%，异常检测灵敏度提升至98.2%，误报率降至1.5%；系统比能耗下降30.4%，计算资源占用减少68.7%。该研究成果显著提升了LNG储罐压力监测的实时性与控制精度，为LNG储运系统的智能安全运行提供了可靠的技术支持。

Abstract:

To address the issues of pressure monitoring lag and control inaccuracy in LNG storage tanks, this study investigates a VOF-based multiphase flow monitoring and control method. By integrating the gas-liquid interface tracking mechanism, the rolling optimization strategy of MPC, and the disturbance compensation capability of DDPG, an integrated "monitoring–optimization–compensation" control architecture is constructed. The VOF-WTD-WPE multi-scale fusion method is adopted to achieve mechanism-data collaborative perception, while the MPC-DDPG cooperative optimization framework handles multiple constraints and nonlinear disturbances. Experimental results show that the model reduces the mean pressure monitoring error by 85.2%, narrows the steady-state fluctuation range by 53.3%, achieves an anomaly detection sensitivity of 98.2%, and lowers the false alarm rate to 1.5%. Additionally, the specific energy consumption of the system is reduced by 30.4%, and computational resource usage is decreased by 68.7%. The results demonstrate that the proposed model significantly enhances the real-time performance and control accuracy of pressure monitoring, providing effective support for the intelligent and safe operation of LNG storage and transportation systems.