摘要
提出组合预测及智能算法滤波应用于风电功率预报曲线。对BP神经网络、小波神经网络、支持向量机3种风电预测技术进行研究,借助仿真软件进行预测误差比较分析;利用小波包分解平滑技术,对上述3种预测结果进行波动平抑,对比分析平抑前后波动率。综合比较分析预测误差及平滑效果,确定风电场功率平滑兼准确预报的系统优化方案,为后续参数控制选择提供依据。
The combination forecasting and intelligent algorithm filtering were proposed to apply to the wind power forecast curve.Three wind prediction techniques including BP neural network,wavelet neural network and support vector machines were studied,and comparative analysis was given by simulation.Based on wavelet packet decomposition smoothing technology,prediction results based on these techniques were stabilized respectively further,fluctuation analysis was done.Optimization scheme was obtained by analyzing and comparing the prediction and smoothing effects,providing a selection reference for the related follow-up control parameter of engineering application.
作者
朱建红
黄琼
孟棒棒
ZHU Jian-hong;HUANG Qiong;MENG Bang-bang(School of Electrical Engineering,Nantong University,Nantong 226019,China)
出处
《计算机工程与设计》
北大核心
2018年第10期3284-3289,共6页
Computer Engineering and Design
基金
国家自然科学基金面上基金项目(51377047)
江苏省高校自然科学研究面上基金项目(5KJB470014)
关键词
风电并网调度预报
功率预测
功率平滑
小波包分解
最优预报
wind power grid scheduling forecast
power prediction
power smooth
wavelet package decomposition
optimal forecast
