基于SWT-SDA与LightGBM的风电爬坡事件预测

刘云冰; 文佳乐; 何晓霞; 李春丽

doi:10.19783/j.cnki.pspc.260332

欢迎来到能源电力期刊网！

English Version 期刊登录

您当前的位置：

首页 >

文章列表页 >

基于SWT-SDA与LightGBM的风电爬坡事件预测

设计开发 | 更新时间：2026-10-15

- 基于SWT-SDA与LightGBM的风电爬坡事件预测
- Wind power ramp event prediction based on SWT-SDA and LightGBM
- 电力系统保护与控制 2026年54卷第20期
- 作者机构：
  
  武汉科技大学数学与系统科学学院,湖北,武汉,430065
- 作者简介：
- 基金信息：
- DOI：10.19783/j.cnki.pspc.260332
  中图分类号：
- 纸质出版：2026
- 稿件说明：
移动端阅览
刘云冰, 文佳乐, 何晓霞, 等. 基于SWT-SDA与LightGBM的风电爬坡事件预测[J]. 电力系统保护与控制, 2026,54(20).

LIU Yunbing, WEN Jiale, HE Xiaoxia, et al. Wind power ramp event prediction based on SWT-SDA and LightGBM[J]. Power System Protection and Control, 2026, 54(20).
刘云冰, 文佳乐, 何晓霞, 等. 基于SWT-SDA与LightGBM的风电爬坡事件预测[J]. 电力系统保护与控制, 2026,54(20). DOI： 10.19783/j.cnki.pspc.260332.

LIU Yunbing, WEN Jiale, HE Xiaoxia, et al. Wind power ramp event prediction based on SWT-SDA and LightGBM[J]. Power System Protection and Control, 2026, 54(20). DOI： 10.19783/j.cnki.pspc.260332.

摘要

风电爬坡事件(wind power ramp event

WPRE)易引发电网频率波动，对电力系统安全稳定运行造成重大影响。为提高其预测精度，覆盖极端风险场景，提出一种结合风险量化的WPRE预测方法。首先，采用同步挤压小波变换与摆动门算法(synchronous squeezing wavelet transform-swinging door algorithm

SWT-SDA)进行爬坡检测，即通过SWT滤除原始功率信号噪声，结合SDA算法，将条件风险价值引入爬坡阈值设定，实现原始数据爬坡识别与标签标注。然后，将爬坡标签作为输入，初步预测风电功率，由LightGBM构建误差修正模型对风电功率初步预测结果进行修正。最后，利用所提爬坡检测方法对修正后功率进行爬坡识别，实现WPRE预测。案例分析表明，所提方法显著提升了WPRE的预测精度。基于条件风险价值的阈值设定使风险等级界定更符合实际需求，提高了电力系统对爬坡风险的防控能力。

Abstract

Wind power ramp events (WPREs) can easily cause fluctuations in grid frequency

posing significant challenges to the safe and stable operation of power systems. To improve the prediction accuracy of WPREs while accounting for extreme risk scenarios

this paper proposes a WPRE prediction method integrated with risk quantification. First

the synchronous squeezing wavelet transform-swinging door algorithm (SWT-SDA) is employed for ramp detection. Specifically

SWT is used to filter noise from raw power signals

and the conditional value-at-risk is incorporated into the ramp threshold setting through the SDA algorithm to identify ramp events and generate labels from the raw data. Second

the ramp labels are used as inputs to achieve preliminary wind power prediction

and LightGBM is employed to build an error correction model for revising the preliminary wind power prediction results. Finally

the proposed ramp detection method is used to identify ramp events from the corrected power data

thereby achieving WPRE prediction. Case studies show that the proposed method significantly improves WPRE prediction accuracy. The threshold setting based on conditional value-at-risk enables risk level definition to better align with practical requirements

thereby enhancing the capability of power systems to prevent and mitigate ramp risks.

关键词

Keywords

references

HE L, SHUAI Z, WAN J, et al. Hybrid frequency direction feature-based protection scheme for inverter-dominated islanded microgrids[J]. Protection and Control of Modern Power Systems, 2026, 11(4): 177-195.

JAYALAKSHMI J, MARY L M. Piecewise linear approximation for identifying wind power ramp events[J]. Automatika, 2023, 64(4): 1306-1314.

QU Y, XU J, SUN Y, et al. A parameter and resolution adaptive algorithm for rapid detection of ramp events in different timescale databases of the power system[J]. International Journal of Electrical Power Energy Systems, 2019, 112: 393-403.

CUI M, ZHANG J, FLORITA A R, et al. An optimized swinging door algorithm for identifying wind ramping events[J]. IEEE Transactions on Sustainable Energy, 2015, 7(1): 150-162.

CUI Y, HE Y J, XIONG X, et al. Algorithm for identifying wind power ramp events via novel improved dynamic swinging door[J]. Renewable Energy, 2021, 171: 542-556.

LIAN L, HE K. Wind power prediction based on wavelet denoising and improved slime mold algorithm optimized support vector machine[J]. Wind Engineering, 2022, 46(3): 866-885.

DOU J, LIU C, WANG B. Short-term wind power forecasting based on convolutional neural networks[J]. IOP Conference Series: Earth and Environmental Science, 2018, 170(4).

MOHAMMED A A Q, AHMED A E, AHMED O A, et al. Wind power forecasting using optimized LSTM by attraction repulsion optimization algorithm[J]. Ain Shams Engineering Journal, 2024, 15(12).

XIONG B, MENG X, WANG R, et al. Combined model for short-term wind power prediction based on deep neural network and long short-term memory[J]. Journal of Physics: Conference Series, 2021, 1757(1).

DHIMAN H S, DEB D, GUERRERO J M, et al. On wavelet transform based convolutional neural network and twin support vector regression for wind power ramp event prediction[J]. Sustainable Computing: Informatics and Systems, 2022, 36.

ZHANG P, LI C, QIAN C, et al. A wind power forecasting method based on GWS-STNet spatio-temporal meteorological prediction[J]. Energy, 2026, 140.

CHINNAPPAN C C. Integrating data-driven and physics- based approaches for robust wind power prediction: a comprehensive ML-PINN-Simulink framework[J]. Scientific Reports, 2025, 15(1): 1-29.

HE Y, ZHU C, CAO C, et al. A wind power ramp prediction method based on value-at-risk[J]. Energy Conversion and Management, 2024, 315.

HE Y, ZHU C, AN X. A trend-based method for the prediction of offshore wind power ramp[J]. Renewable Energy, 2023, 209: 248-261.

PFLUG G C. Some remarks on the value-at-risk and the conditional value-at-risk[M] // Probabilistic Constrained Optimization: Methodology and Applications, Springer, Boston, MA, USA, 2000: 272-281.

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

计及风电不确定性与尾部风险的孤岛微电网可靠性快速评估

基于数据驱动的可解释性低压台区理论线损率计算方法

基于超参数优化和误差修正的STAGN超短期风电功率预测

发电商参与碳市场与电力中长期市场联合决策模型

考虑CVaR的“新能源+储能”电厂日前市场投标策略