融合小波分解的Lasso光伏净负荷预测方法

Lasso Photovoltaic Net Load Forecasting Method Based on Wavelet Decomposition

为实现碳达峰、碳中和的目标,有必要对光伏净负荷进行预测,辅助电网根据不同用户的用电需求进行智能分配用电。文章提出了一种小波分解结合Lasso回归模型的预测模型,其中小波分解将时间序列数据的时频域进行对调,聚焦到数据的细节,更适合描述光伏净负荷的内在特性,而在Lasso回归模型中引入该方法将原始数据映射到合适的高维特征空间,使得Lasso回归模型应用于非线性的光伏净负荷数据。在实验验证中首先根据已有的10个台区的光伏净负荷数据进行分析,然后通过这些光伏净负荷数据使用预测模型进行训练和预测,实验结果表明该预测模型具有较高的准确性。

In response to the demand of carbon peak and carbon neutralization,it is necessary to predict the photovoltaic net load,and assist the grid to intelligently distribute power according to the power demand of different users.A forecasting model based on wavelet decomposition combined with Lasso regression model is proposed.The wavelet decomposition adjusts the time-frequency domain of time series data and focuses on the details of the data,which is more suitable to describe the internal characteristics of photovoltaic net load.In lasso regression model,kernel method is introduced to map the original data to an appropriate high-dimensional feature space,so that Lasso regression model can be applied to nonlinear photovoltaic net load data.In the experiment,firstly,the photovoltaic net load data of 10 transformer areas are analyzed,and then the forecasting model is trained and predicted through these photovoltaic net load data.The experimental results show that the forecasting model has high accuracy.