| 引用本文: | 班潇璇,李少林,李庆,秦世耀,李春彦.基于秒级风速预测的风电频率主动支撑能力动态评估[J].电力自动化设备,2025,45(7):54-60,70 |
| BAN Xiaoxuan,LI Shaolin,LI Qing,QIN Shiyao,LI Chunyan.Dynamic evaluation of wind power active frequency support capability based on second-level wind speed forecast[J].Electric Power Automation Equipment,2025,45(7):54-60,70 |
|
| 摘要: |
| 考虑风电出力具有随机波动性，为准确估计风电频率主动支撑能力，提出了一种基于预测误差模态分解的改进长短期记忆网络秒级风速预测方法。通过提取误差中的有效信息，实时修正预测风速，提升了未来数十秒风速的预测精度；进而建立了综合考虑风速与转速变化的风电有功调频能力聚合评估模型，实现了风电频率主动支撑能力及变化趋势的动态估计。在MATLAB/Simulink中搭建了风电频率主动支撑能力动态评估模型，基于风电实际运行数据，验证了秒级风速预测与频率主动支撑能力动态估计方法的准确性与有效性。 |
| 关键词: 风力发电 频率主动支撑 风速预测 长短期记忆网络 能力评估 |
| DOI：10.16081/j.epae.202503008 |
| 分类号: |
| 基金项目:国家电网公司总部科技项目(4000?202315093A-1-1-ZN) |
|
| Dynamic evaluation of wind power active frequency support capability based on second-level wind speed forecast |
|
BAN Xiaoxuan, LI Shaolin, LI Qing, QIN Shiyao, LI Chunyan
|
|
State Key Laboratory of Renewable Energy Grid-Integration, China Electric Power Research Institute, Beijing 100192, China
|
| Abstract: |
| Considering the stochastic fluctuation of wind power output, in order to accurately estimate the frequency active support capability of wind power, an improved long short-term memory network-based second-level wind speed forecast method based on the forecast error modal decomposition is proposed. By extracting the effective information from the errors and correcting the forecasted wind speed in real time, the forecast accuracy of wind speed in the next few tens of seconds is enhanced. Furthermore, an aggregated evaluation model of wind power active power frequency regulation capability is established with the comprehensive consideration of the changes for wind speed and rotational speed, and the dynamic estimation of wind power frequency active support capability and variation trend is realized. A dynamic evaluation model for the wind power frequency active support capability is developed in MATLAB/Simulink. The accuracy and effectiveness of the second-level wind speed forecast and dynamic evaluation method for frequency active support capability are verified based on actual operational data of wind power. |
| Key words: wind power frequency active support wind speed forecast long short-term memory network capability evaluation |
