绿色电力证书交易影响下计及输电阻塞影响的综合型能源发售电商竞价策略研究

Power Price Marketing Strategy of Comprehensive Energy-based Electricity Sales Company Participating in Electricity Market Competition Under Ubiquitous Environment of Internet of Things

近年来,我国在分布式能源的开发与利用上发展迅猛,但与此同时带来的政府补贴资金短缺和"弃风弃光"问题也日趋严重。随着电力市场改革的逐步深入,考虑可再生能源配额制和绿色电力证书交易政策的驱使作用,含火电和风电机组的综合型能源发售电商如何制定竞价策略使得其既符合政策要求又满足利益最大化显得尤为重要。首先,在对配额制和绿证政策分析的基础上,提出通过"绿证交易市场"处理风电机组的出力偏差,其价格则由绿证用户投标竞价决定;其次,建立发售电商参与竞价的日前市场与绿证交易市场联合优化模型,针对风电实际出力不足和过剩2种情况,发售电商在绿证交易市场中进行购电竞价和售电竞价,并采用随机森林回归算法和多场景缩减技术将风电出力预测结果与误差场景相结合,模拟实际风电出力多场景;然后考虑网络中输电阻塞对发售电商报价的影响,构造了基于节点电价法的输电阻塞成本最小化模型,并采用利用蒙特卡洛模拟和遗传算法进行求解;最后通过算例分析验证文中模型和方法的准确性和合理性。

In recent years, China has developed rapidly in exploitation and utilization of distributed energy, but at the same time, shortage of government subsidy funds and the problem of "curtailment of wind and solar energies" become increasingly serious. With promotion of electricity market-oriented reform and the influence of renewable energy quota system and green power certificate trading policy, it is particularly important for integrated energy generators, including thermal and wind generators, to formulate bidding strategies, so that they can meet policy requirements and maximize benefits. Firstly, based on analysis of quota system and green certificate policy, this paper proposes a method to deal with the output deviation of wind turbines through "green certificate trading market", whose price is determined by the bidding of green certificate users. Secondly, a joint optimization model of day-ahead market and green certificate trading market for generating companies participating in bidding is established. By considering the situations of insufficient and excessive outputs of wind power, and generation companies bid for electricity purchasing and selling in green certificate trading market, stochastic forest regression algorithm and multi-scenario reduction technology are used to combine the prediction results of wind power output with error scenarios to simulate actual wind power output. Finally, an example is given to verify accuracy and rationality of the proposed model and method.