基于滚动时域分解策略的原油混输调度模型

Scheduling Model Based on Rolling-Horizon Algorithm for Crude Oil Transportation

为了解决单一完整数学规划模型无法快速编制出包含复杂约束原油储运方案的问题,考虑长输管线入线、出线油种不同步,码头储罐部分原油需经转油线、中转罐后才能到达厂区等约束,提出了一个新的连续时间混合整数线性规划(Mixed Integer Liner Programming,MILP)模型。该模型为避免产生非线性约束,以固定原油调合配比为基础,限制混合原油的质量。采用滚动时域分解策略对模型进行分步求解,在简单分解的基础上添加辅助时间片段和安全保障约束,从而保证了分解模型的可行性和优化性。以中国某沿海企业的原油调度问题为例,评估模型的可行性、优化性和计算性能,结果表明,滚动时域模型在短时间内编制出了优化的储运方案,能够用于长周期、大规模原油调度问题的解决。

Establishing on operations research model is the mainstream method for automatic optimization scheme,but it is not possible to use a single undecomposed mathematical model to quickly obtain a good solution for complex crude oil scheduling issue.Therefore,a continuous-time mixing integer linear programming(MILP)model based on an iterative rolling-horizon algorithm was developed.This model incorporated many realistic operational features,such as the delayed transportation of crude oil in a long-distance pipeline,the transportation of crude oil from the storage tank farm to the transit tank farm.To avoid non-linear constraints in the model,property constraints were based on fixed crude oil blending ratio.To ensure feasibility and optimization of the scheduling model,a rolling-horizon algorithm characterized by auxiliary time period and security guarantee constraints were applied.The model was further tested with a real scheduling problem in a China east coast refinery.Test results showed that this model can provide a good schedule in a timely manner for the long and complicated crude oil transportation process.