摘要
成品油管网具有多批次顺序输送特性,其管道网络参数为连续时变变量,导致压力约束中存在多个连续时变变量相乘项,具有很强的非线性。为此,该文借鉴流体力学中描述流体运动的思路,首先,将建模视角从传统基于欧拉法的管网视角切换到基于拉格朗日法的批次视角对模型进行等价重构以降低非线性;然后,综合大M法、分段线性化和二维凸组合等多种线性化技术进一步消除非线性,将非线性模型转换为易于求解的混合整数线性规划模型;最后,进行算例分析,结果表明所提模型较现有模型更能反映管网的复杂工况,在确保运输安全的同时提高经济性,其中电力成本降低27.95%,平均近似误差从19.74%降低至0.61%。
Unlike water pipeline networks that transport a single medium,refined oil pipeline networks transport multiple media sequentially.As a result,the pipeline network parameters continuously vary over time,leading to multiple continuous time-varying variable multiplication terms in pressure constraints that exhibit strong nonlinearity.Existing research typically divides the transport of refined oil into two separate decision stages or simplifies the constraints to reduce model nonlinearity.However,most methods suffer from low safety or poor economics.To address these issues,this paper proposes a joint scheduling and operation optimization model for multi-batch refined oil pipeline network based on pressure constraint reconstruction.The model's accuracy is improved through model reconstruction and linear approximation,which accurately reflects the complex transportation conditions of refined oil pipeline networks.First,according to different approaches to describe fluid motion in fluid mechanics,the construction of refined oil transport models switches from the traditional Eulerian-based network perspective to the Lagrangian-based batch perspective.Without compromising model accuracy,continuous time-varying variables in pressure constraints are reconstructed equivalently as discrete variables to reduce model nonlinearity.Simultaneously,some difficult-to-quantify variables are equivalently reconstructed as new variables that are easy to express and have intuitive physical meanings.Various linearization techniques,such as the big-M method,piecewise linearization and two-dimensional convex combination,are further utilized to eliminate the nonlinearity of the reconstructed model.Within the acceptable range of practical engineering errors,the two-stage joint optimization model is transformed into a tractable mixed integer linear programming model.The proposed model is applied to a refined oil transportation pipeline network in the southwest region of China.Compared with the pump scheduling of existing models,the electricity cost of the proposed model is only 1.39%higher than the optimal solution.On the other hand,the electricity cost of the existing joint optimization model is 27.95%higher than the optimal solution.These results validate the power cost accuracy of the proposed model.Furthermore,pump station characteristics and pressure loss in the proposed model are closer to the actual values,with maximum approximation errors of only 0.41%and 1.80%,respectively.This validates the model's approximation accuracy.In contrast,the maximum approximation error of existing models reaches 19.74%,making it difficult to reflect the actual pressure conditions of refined oil pipeline networks.This can lead to incorrect pump allocation schemes,which worsens both economic efficiency and operational safety.Simulation results show that the pumping scheduling provided by the existing model could not be used in the implement field.The simulation analysis leads to the following conclusions:(1)The proposed model has superior approximation accuracy compared to existing two-stage joint optimization models,which reflects the complex conditions of refined oil pipeline networks more accurately.(2)The proposed model can determine more accurate pump allocation schemes compared to existing models,which ensures transportation safety and improves economic efficiency.(3)A larger number of segments can improve the approximation accuracy of the constraints,which further reduces electricity costs.However,this comes at the cost of model solution speed.Therefore,the number of segments should be selected based on the actual model size and computation time.
作者
刘晶冠
艾小猛
方家琨
崔世常
王盛世
Liu Jingguan;Ai Xiaomeng;Fang Jiakun;Cui Shichang;Wang Shengshi(State Key Laboratory of Advanced Electromagnetic Engineering and Technology Huazhong University of Science and Technology Wuhan 430074,China)
出处
《电工技术学报》
EI
CSCD
北大核心
2023年第S01期32-43,共12页
Transactions of China Electrotechnical Society
基金
国家自然科学基金项目(52177089)
国家管网科技攻关项目(GWHT20200001399)资助。
关键词
多批次输送
线性化技术
电力成本
输油节能
Multi-batch transportation
linearization technique
electricity costs
energy saving in oil transportation
