虚拟发电厂
能量(信号处理)
功率(物理)
工程类
计算机科学
汽车工程
环境科学
可再生能源
分布式发电
电气工程
数学
物理
统计
量子力学
作者
Zeyuan Dong,Z. Zhang,Minghui Huang,Shaorong Yang,Jun Zhu,Meng Zhang,Dongjiu Chen
出处
期刊:Energy
[Elsevier]
日期:2024-04-11
卷期号:297: 131235-131235
被引量:12
标识
DOI:10.1016/j.energy.2024.131235
摘要
Large-scale new energy access to the power grid poses significant challenges to its stable operation. Differentiated user-side power consumption patterns further widen peak-valley differences in power demand. This paper focuses on operation scheduling problems of virtual power plants with coordinated optimization of diverse flexible loads and new energy, through efficient aggregation and optimization control of new energy and demand-side resources, improving power supply and demand mismatch. Firstly, Long Short-Term Memory and Latin Hypercube Sampling were employed for predicting the day-ahead output of wind and photovoltaic power. Secondly, wind and photovoltaic power, batteries and a pumped storage plant were aggregated into a virtual power plant, and the day-ahead optimization scheduling model was constructed considering system operation costs, energy curtailment costs and demand response costs. Finally, a simulation analysis was conducted. The results show that when large-scale new energy accesses to the power grid, traditional "Generation varies with Load" regulation modes will cause massive energy waste, while the "Generation-load Interaction" regulation mode can achieve the linkage optimization between the generation side and the demand side, enhancing the system acceptance of new energy. Demand response can optimize users' power consumption behaviors, reducing the charging costs by 52.13% and heating costs by 0.84%.
科研通智能强力驱动
Strongly Powered by AbleSci AI