数据驱动下风电–抽蓄联合参与日前–实时市场随机鲁棒竞价策略

风电出力不确定性和波动性导致其市场竞争力较弱,且电价的预测偏差可能进一步加大其市场风险。为减少不确定性因素对风电-抽蓄联合系统收益的不良影响,由对抗变分贝叶斯神经网络生成以风电为代表的可再生能源出力场景,基于数据驱动方法对实时电价进行模糊不确定性建模。通过随机鲁棒优化建立风电-抽蓄联合参与日前和实时电力市场的三阶段模型,相较于常规的两阶段随机优化,增加了第三阶段鲁棒改进过程,保证了竞价方案能够既有经济性又能够有效应对极端场景、既有鲁棒性又不过分保守。结果表明,所提方法比传统不确定性分析方法具有明显优势,能够在避免人为假设的前提下,高效且真实反映可再生能源出力场景以及表征电价不确定性,可有效减少风电实时出力波动和电价预测偏差带来的较高不平衡惩罚。

Optimal Hybrid Stochastic Robust Bidding Strategy of Wind and Hydraulic Pumped Storage Jointly Participating in Day-ahead and Real-time Market Using Data-driven Method

The uncertainty and volatility of wind power output may cause weak competitiveness in the market,and the forecast deviation of electricity price may further increase its market risks.In order to reduce the adverse impacts of the uncertainty factors on the wind power storage station revenue,the scenarios of renewable energy output such as wind power are generated by the adversarial variational Bayes neural network,and the ambiguous uncertainty model of real-time prices is established based on a data-driven method.A three-stage model of a wind-pump power station participating in the day-ahead and the real-time electricity market is built with the stochastic robust optimization.Compared with the conventional two-stage stochastic optimization,the third stage robust improvement process added ensures the bidding strategy to be economical and effective to deal with the extreme scenarios,robust but not too conservative.The results show that the proposed approach has obvious advantages over the traditional uncertainty analysis method.It is able to efficiently and truly reflect the renewable energy output scenarios and characterize the uncertainty of electricity prices without any artificial assumptions,further effectively reducing the high imbalanced penalties caused by the fluctuation of the wind power output and the deviation of electricity price prediction.