考虑光伏出力和负荷的不确定性,研究一种适合工程推广的综合能源系统经济调度求解方法。首先,建立综合能源系统混合整数非线性概率模型;其次,采用特殊序列集合(SOS-2)方法对非线性模型进行分段线性化;然后,采用对称采样策略的无迹变换方法进行确定性转换,以揭示概率性背后隐藏的信息;最后,对比分析混合整数线性规划(MILP)求解方法与粒子群算法的计算结果。采用 PuLp开源建模框架对某示范性综合能源项目进行建模、 求解,结果表明:一方面,对称采样策略的无迹变换方法求解效率较高,其概率密度分布结果可为运行人员揭示概率性背后隐藏的运行状况信息;另一方面,与粒子群算法相比,PuLp开源框架的 MILP求解方法效率可提高数倍,适于工程推广。
For the purpose of popularization in application scenarios, this paper presents an economic dispatch method for the integrated energy systems considering the uncertainty of photovoltaic output and power load. Firstly, a mixed-integer nonlinear probability model for integrated energy systems is established. Secondly, the nonlinear model is segmented and linearized by the special sequence set (SOS-2) method. Then, the symmetric sampling strategy-based traceless transformation method is used to perform deterministic transformation, and to reveal the underlying probabilistic information. Finally, the calculation results of MILP and particle swarm optimization are compared and analyzed. The PuLp open-source modeling framework is utilized to model and solve the potential problems in a demonstration integrated energy project. The results show that, on one hand, the traceless transformation method which is based on the symmetric sampling strategy is efficient, and the generated probability density distribution can reveal the underlying probabilistic density distribution so that the operation information can be available for the system operator. On the other hand, compared with the particle swarm algorithm, the efficiency of solving MILP problems via the PuLp open-source framework can increase several times, which is suitable for real applications.
陈晨,李端超,王海伟,苏寒,徐晶晶.考虑不确定性的综合能源系统日前经济调度[J].电力科学与技术学报,2021,(2):24-30. Chen Chen, Li Duanchao, Wang Haiwei, Su Han, Xu Jingjing. Study on day-ahead economic dispatch of integrated energy system considering uncertainty[J]. JOURNAL OF EIECTRIC POWER SCIENCE AND TECHNOLOGY,2021,(2):24-30.复制