摘要
基于狄利克雷过程混合模型和变分推断算法建立了风电出力-光伏出力-电负荷-热负荷高维数据驱动不确定模糊集,在此基础上提出了考虑高维不确定性的热电联产虚拟电厂(CHP-VPP)两阶段随机鲁棒优化调度策略。两阶段中第1和第2阶段分别以日前和实时市场收益最大为目标函数,考虑各机组运行、功率平衡、市场交易和网络结构等多种约束,开发了加速列与约束生成(AC&CG)算法来对此优化调度问题进行求解。结果表明:所提出的随机鲁棒优化方法实现了CHP-VPP经济性和鲁棒性的均衡;模糊集中不确定边界值与CHP-VPP总收益呈负相关,随着不确定边界值的增大,系统的总收益降低,鲁棒性增加。
Based on the Dirichlet process mixture model and the variational inference algorithm,a high-dimensional data-driven uncertain fuzzy set of wind power,solar power,electric load and heat load was established.On this basis,a two-stage stochastic robust optimization scheduling strategy for combined heat and power-virtual power plant(CHP-VPP) was proposed considering high-dimensional uncertainty.In the first and second stage,taking the maximum daily and real-time market revenue as the objective function,considering various constraints such as unit operation,power balance,market transactions and network structure,the accelerated column and constraint generation(AC&CG) algorithm was developed to solve the optimal scheduling problem.Results show that the proposed stochastic robust optimization method achieves the balance between the economy and robustness of CHP-VPP.The uncertain boundary value of the fuzzy set is negatively correlated with total revenue of the system.With the increase of uncertain boundary value,total revenue of the system decreases and the robustness increases.
作者
元志伟
于松源
房方
刘吉臻
YUAN Zhiwei;YU Songyuan;FANG Fang;LIU Jizhen(School of Control and Computer Engineering,North China Electric Power University,Beijing 102206,China;State Key Laboratory of Alternate Electrical Power System with Renewable Energy Sources,North China Electric Power University,Beijing 102206,China)
出处
《动力工程学报》
CAS
CSCD
北大核心
2023年第2期194-204,共11页
Journal of Chinese Society of Power Engineering
基金
国家自然科学基金资助项目(52176005)
国家重点研发计划资助项目(2018YFE0106600)。
关键词
热电联产
虚拟电厂
数据驱动
不确定性
随机鲁棒优化
combined heat and power
virtual power plant
data-driven
uncertainty
stochastic robust optimization
