光伏出力的随机性和负荷用电的波动性对微电网的优化调度影响显著，为此提出了预测-调节-决策一体化的策略框架。基于高斯过程回归(Gaussian process regression，GPR)将光伏出力和负荷用电典型日历史数据自适应生成的置信区间与鲁棒优化中不确定集的构建相结合，建立了基于区间概率不确定集的自适应鲁棒优化调度模型。首先，通过GPR生成自适应鲁棒优化调度模型中不确定集的固定项，然后调节决策环节所考虑的风险水平以确定不确定集中的波动项，进而确定衡量不同调度保守度下的不确定集边界;接着采用预测区间质量评测指标来考核各个不确定集所对应的区间优劣。最后，通过改进的IEEE-37节点微电网系统验证了所提模型在有效抵御光伏出力和负荷用电波动的同时保持较低的运行成本。
Two-stage Adaptive Robust Optimal Scheduling Based on the Interval Probability Uncertainty Set for Microgrids
The randomness of photovoltaic output and the fluctuation of load consumption have a significant impact on the optimal scheduling of microgrids.To this end,a strategic framework of integrated prediction-regulation-decision-making is proposed.Based on Gaussian process regression(GPR),an adaptive robust optimal scheduling model based on the interval probabilistic uncertain sets is established by combining the confidence interval generated adaptively from the historical data of typical days of photovoltaic output and load power consumption with the construction of uncertain sets in robust optimization.First,the fixed term of the uncertain set in the adaptive robust optimal scheduling model is generated by GPR.Then,the fluctuation term in the set is determined by adjusting the risk level considered by the decision-making link,thereby determining the boundary of the uncertain set under different scheduling conservatism.Next,the quality evaluation index of the prediction interval is used to assess the quality of each interval corresponding to each uncertain set.Finally,the test results on the modified IEEE 37-bus microgrid system verifies that the proposed model can effectively resist the fluctuation of photovoltaic output and load power consumption while maintaining low operating costs.
国家科技期刊平台
NSTL
万方数据
