考虑日周期性影响的光伏功率爬坡事件非精确概率预测

光伏功率爬坡事件的可靠预测对电力系统运行决策至关重要。针对现有光伏发电功率爬坡事件预测存在误报与漏报的问题，提出了一种考虑日周期性影响的光伏功率爬坡事件非精确概率预测方法。首先，定义了新的光伏爬坡特征量，以有效剔除光伏发电功率中的日趋势性变化。进而，为了避免光伏爬坡样本数据有限可能引发的预测误差，通过结构学习构建了最优信度网络，对光伏功率爬坡事件进行非精确概率预测；其中，信度网络节点关联的非精确条件概率由多状态随机变量的非精确狄利克雷模型统计得到。最后，根据给定气象条件，推理计算各爬坡状态发生的概率区间。基于某光伏电站数据的算例仿真验证了所述方法的有效性，表明所提方法可有效捕捉光伏发电功率变动中由气象条件引发的突变事件。

Imprecise Probabilistic Prediction of Photovoltaic Power Ramp Event Considering Daily Periodic Effect

Reliable prediction of photovoltaic power ramp event is very important for power system operation. To avoid false and missing report, an imprecise probabilistic prediction method of photovoltaic power ramp event considering daily periodic effect is proposed. Firstly, new ramp characteristic is defined, which can effectively eliminate the daily trend change of photovoltaic power. Then, the optimal credal network(CN)is constructed by structure learning to implement the imprecise probabilistic prediction of solar power ramp events, so that the prediction error caused by the limited sample is avoided. Specially, imprecise Dirichlet model(IDM)with multi-state random variables is used to evaluate the imprecise conditional probability associated with nodes of CN. Finally, according to the given meteorological conditions, the probability interval of ramp events is inferred. The effectiveness of the proposed method is verified by the data of a photovoltaic plant. The results show that the proposed method can effectively capture the mutation events caused by meteorological conditions of change of photovoltaic power.