当前位置：首页 > 过刊浏览->2026年第52卷第4期
基于混合整数非线性规划的LNG接收站罐容优化模型
Tank capacity optimization model of LNG receiving terminals based on mixed-integer nonlinear programming
浏览(12) 下载(1)
- DOI:
- 作者:
- 董一直，杨晓芳，徐海龙
Dong Yizhi, Yang Xiaofang, Xu Hailong
- 作者单位:
- 1.中海石油气电集团有限责任公司，北京 100028 ；2.国家管网集团油气调控中心，北京 100013
- 关键词:
- 混合整数非线性规划；液化天然气；动态罐容优化
mixed-integer nonlinear programming; liquefied natural gas; dynamic storage capacity optimization
- 摘要:
- 液化天然气（LNG）储罐在天然气储气调峰体系中发挥着重要作用，科学合理地计算和规划储罐罐容
是LNG接收站前期研究阶段的重要任务之一。目前，工程项目通常采用静态经验公式与动态调整船期相结合的
方法来估算LNG储罐罐容，该方法依赖人工经验，未考虑用气市场的波动性、调峰需求、卸船时间和卸船流量
对动态罐容的影响，导致计算结果普遍大于LNG接收站实际需求。引入混合整数非线性规划（MINLP）方法，
将全年储罐最大动态储量最小作为目标函数，以LNG船到港分配方案作为决策变量，建立离散调度决策与库存
动态平衡联立求解的罐容优化模型。以某拟建LNG接收站为算例，采用Gurobi求解器求解，并与传统常规方
法计算结果进行对比。结果表明：静态法、动态调整船期法与MINLP法计算所得罐容分别为188.1×104、
55.4×104、53.1×104 m3，优化模型在满足安全储备与调峰约束的基础上，既可确定经济合理的LNG运输船
来船日期、LNG接收站最小罐容，又可输出LNG储罐库存动态曲线，模型具备良好的可行性与有效性。
MINLP算法克服了传统常规计算方法的局限性，可推广应用于LNG接收站罐容优化设计实际工作中。
Liquefied natural gas (LNG) plays a vital role in natural gas storage and peak-shaving. The scientific and rational calculation and planning of storage tank capacity is a critical task in the preliminary research phase of LNG receiving terminals. Current engineering practice generally employs static empirical formulas combined with dynamically adjusted shipping schedules to estimate LNG storage tank capacity. Such approaches rely heavily on empirical heuristics and fail to account for the impact of gas market volatility, peak shaving demands, unloading time, and unloading flow rate on dynamic tank capacity. Consequently, the calculated results are overestimated compared with the actual demands of LNG receiving terminals. A mixed-integer nonlinear programming (MINLP) approach was introduced. The minimum dynamic storage tank capacity was taken as the objective function. The ship arrival schedules were selected as decision variables. A tank capacity optimization model was thus developed to simultaneously solve discrete scheduling decisions and dynamic inventory balance. With a proposed LNG receiving terminal as an example, the model was solved using the Gurobi solver, and the results were compared against those obtained from conventional methods. The results indicate that the tank capacity derived from the static method, the dynamically adjusted shipping schedules, and the MINLP method is 188.1×104, 55.4×104, and 53.1×104 m3, respectively. Under the safety reserve and peak-shaving constraints, the optimization model not only determines economically justified ship arrival dates and the minimum tank capacity of LNG receiving terminals, but also provides a complete dynamic profile of the inventory curve variations. The model proves to be both feasible and effective. The MINLP method overcomes the inherent deficiencies of conventional calculation approaches and demonstrates strong potential for broader application in the tank capacity optimization design of LNG receiving terminals.