促进大规模新能源消纳的电力市场管理研究

The large-scale new energy consumption is an important problem for the new energy development in China. It has important significance in promoting the sustainable development of new energy. In order to meet the needs of China's new energy development and electricity market reform, a research topic on electricity market management to promote large-scale new energy consumption is put forward. According to the logic, aiming at the problem of low prediction accuracy of new energy generation, an intelligent forecasting model is established for wind power and photovoltaic power generation based on spatial correlation and deep learning research; Aiming at the unreasonable market mechanism of peaking ancillary service, this project focuses on the compensation mechanism and sets up a electricity market management system for peak ancillary service to promote large-scale new energy consumption; Aiming at the defect that the original electricity market transaction didn’t consider new energy, this project builds long-term trading mechanism and spot trading mechanism for electricity market that promotes large-scale new energy consumption. This project explores the operational change rules and consumptive market mechanism of large-scale new energy consumption under the influence of uncertainty and establishes a set of electricity market management theories and systems with Chinese characteristics to promote large-scale new energy consumption to obtain breakthroughs progress for the fusion research of large-scale new energy accommodation and electricity market reform, better play the decisive role of market mechanism in the optimal allocation of resources and contribute to the development of green power, resource conservation and environmental protection in China.

大规模新能源消纳是当前我国新能源发展面临的重要难题，对促进新能源可持续发展具有重要意义。为了应对我国新能源发展和电力市场改革的需求，提出促进大规模新能源消纳的电力市场管理研究课题。按照前后逻辑关系，针对新能源发电预测精度低的问题，通过空间相关性和深度学习研究，建立含风电和光伏发电的新能源发电功率智能预测模型；针对调峰辅助服务市场机制的不合理，研究补偿机制并构建促进大规模新能源消纳的调峰辅助服务电力市场管理体系；针对原有电力市场交易中未考虑新能源的缺陷，建立促进大规模新能源消纳的电力市场中长期交易机制和现货交易机制。探索不确定影响下大规模新能源消纳的运行变化规律和消纳市场机理，建立一套中国特色的促进大规模新能源消纳的电力市场管理理论和体系，使大规模的新能源消纳与电力市场改革的融合研究取得突破性进展，更好地发挥市场机制在资源优化配置中的决定作用，为发展我国绿色电力、节约资源、保护环境做出贡献。

本项目从新能源参与对电力市场影响、新能源发电功率预测、计及新能源的中长期合约与现货日前市场的衔接与出清机制、电力现货市场各阶段市场的衔接与出清机制等多个方面展开研究。基于电力现货市场价格信号的复杂性，构建由三个模块构成的新能源对电力现货市场影响分析模型，结果证实了新能源发电对于电价影响高于常规历史数据、占比分类错误率较高等结论。构建基于和声搜索算法优化的核极限学习机的混合新能源发电功率预测模型，首先通过皮尔森相关系数筛选模型输入数据，减少数据冗余；而后，采用CEEMD-SE的组合数据预处理策略，对发电功率时间序列进行分解和重构，消除数据噪声；其次，采用HS-KELM模型进行建模预测，集成处理后得到最终的新能源发电功率预测值具有更高的预测精度。提出计及中长期合约电量分解与新能源参与的日前电力市场交易优化模型，首先构建中长期合约电力分解模型，将分解得到的每日中长期合约电量作为约束引入日前市场的优化模型中，保证中长期合约电量物理执行；构建新能源参与的日前市场多目标出清优化模型，利用模糊优选方法对多目标进行转换，较好地平衡经济性与节能减排目标；最后采用基于GA-PSO组合优化模型对构建模型进行求解。结果表明，本项目构建的多目标优化函数能够在保证系统运行经济效益的基础上，实现环境效益最大化，达到节能减排的效果；同时随着新能源渗透率的增加，系统不确定性增加，常规机组的成交电量有所下降。构建了基于拉丁超立方采样进行场景集生成法与改进谱聚类分析的场景集削减策略，能够选择出最具代表性的场景集。基于电力现货市场出清流程，将含有新能源较多的系统将引入日内市场，以减小实时市场的功率偏差，提高系统运行的经济性和稳定性；构建日前市场和模拟日内市场联合出清优化模型，在各个日内市场考虑对应实时市场新能源偏差功率的不确定性、电价不确定性，建立各日内市场和模拟实时市场联合优化模型。