考虑锅炉效率不确定的蒸汽动力系统鲁棒优化设计

Robust Optimal Design of Steam Power System Considering the Uncertainty of Boiler Efficiency

摘要: 锅炉效率在蒸汽动力系统运行过程中存在波动，为保证系统的高效安全运行，本文研究锅炉效率不确定情况下的蒸汽动力系统鲁棒优化问题。本研究基于一种数据驱动的自适应鲁棒优化方法，利用工业已有数据样本，通过鲁棒核密度估计构建锅炉效率不确定集合，降低了鲁棒优化问题的保守性，再结合混合整数线性规划(MILP)模型，建立自适应鲁棒混合整数线性规划模型。优化模型以蒸汽动力系统的总运行成本最小为目标函数，决策变量包括描述蒸汽轮机选型的整数变量和表征系统蒸汽分布情况的连续变量。应用实例表明，与确定性方法相比，鲁棒优化得到的最优操作策略，鲁棒优化在理想效率下总成本增加约24.00%，在低效率下总成本降低约12.17%，对于系统的鲁棒性上提高了15.14%，鲁棒生产超高压蒸汽，可以随需求转换为其他等级蒸汽，系统工作状态受到外界因素影响小，在锅炉效率存在波动的情况下，结果更加具有经济性。

Abstract: The boiler efficiency fluctuates during the operation of the steam power system. In order to ensure the efficient and safe operation of the system, this paper studies the robust optimization of the steam power system when the boiler efficiency is uncertain. This research is based on a data-driven adaptive robust optimization method, using existing industrial data samples to construct a boiler efficiency uncertainty set through robust kernel density estimation, reducing the conservativeness of the robust optimization problem, and combining it with mixed integer linearity Planning (MILP) model to establish an adaptive robust mixed integer linear programming model. The optimization model takes the minimum total operating cost of the steam power system as the objective function. The decision variables include integer variables that describe the selection of steam turbines and continuous variables that characterize the steam distribution of the system. Application examples show that, compared with the deterministic method, the optimal operation strategy obtained by robust optimization, the total cost of robust optimization increases by about 24.00% under ideal efficiency, and the total cost is reduced by approximately 12.17% under low efficiency, which is very effective for the robustness of the system. The robustness has been increased by 15.14%. The robust production of ultra-high-pressure steam can be converted to other grades of steam as required. The working state of the system is less affected by external factors. In the case of fluctuations in boiler efficiency, the result is more economical.