摘要
根据光伏发电功率的变化是非稳定和随机的特点,为进一步提高光伏功率预测精度,该文提出一种基于VMD-SE-LSSVM和迭代误差修正的预测方法。该方法首先采用变分模态分解(VMD)技术将历史光伏发电功率分解成一系列有限带宽的子模态,避免模态混叠和噪声冲击的影响;然后用最小二乘支持向量机分别预测光伏发电功率和误差,将误差补偿后的功率值作为最终的预测结果。此外,该文还采用样本熵(SE)原理将天气类型量化作为特征向量输入支持向量机(SVM)参与预测,兼顾了天气因素和时间维度对预测值的影响。经过仿真和与传统方法的预测结果比对,该文所提出的方法在不同天气类型中均提高了准确性,在光伏电站功率预测中具有一定的理论与实用价值。
The change of photovoltaic power series is non-smooth and stochastic. In order to improve the prediction accuracy of photovoltaic power,a prediction method based on VMD-SE-LSSVM and iterative error correction is proposed.First,a variational modal decomposition(VMD)technique is used to decompose the photovoltaic power generation into a series of bandwidth-limited sub modal,which avoids the influence of modal aliasing and the impact of noise. Then,the least squares support vector machine is applied to predict the PV power and power error respectively;afterwards,the power value after error compensation is acted as the final prediction results. In addition,this paper also uses the sample entropy principle to quantify weather types as feature vectors input SVM to participate in prediction,with the impact of weather factors and the time dimension on the predicted value took into account. Compared with the prediction results of the traditional methods,the simulation results show that the accuracy of power forecasting in different weather conditions is improved through this proposed approach.
作者
余向阳
赵怡茗
杨宁宁
岳同耿日
高春阳
Yu Xiangyang;Zhao Yiming;Yang Ningning;Yue Tonggengri;Gao Chunyang(Instiute of Water Resources and Hydro-Eetric Enginering,Xi an Uninersity of Technology,Xi'an 710048,China)
出处
《太阳能学报》
EI
CAS
CSCD
北大核心
2020年第2期310-318,共9页
Acta Energiae Solaris Sinica
基金
青年科学基金(51507134):陕西省协同创新计划(2014XT-21)。
关键词
预测模型
支持向量机
模态分解
误差分析
熵
prediction model
support vector machine
modal decomposition
error analysis
entropy
