AUTHOR=Luo Haifeng , Dou Xun , Sun Rong , Wu Shengjun 

TITLE=A Multi-Step Prediction Method for Wind Power Based on Improved TCN to Correct Cumulative Error

JOURNAL=Frontiers in Energy Research

VOLUME=Volume 9 - 2021

YEAR=2021

URL=https://www.frontiersin.org/journals/energy-research/articles/10.3389/fenrg.2021.723319

DOI=10.3389/fenrg.2021.723319

ISSN=2296-598X

ABSTRACT=Wind power generation is likely to hinder the safe and stable operations of power systems for its irregularity, intermittency and non-smoothness. Since wind power is continuously connected to power systems, the step length required for predicting wind power is increasingly extended, thereby causing an increasing cumulative error. How to correct the cumulative error to predict wind power in multi-step should be urgently solved. In this study, a multi-step wind power prediction method was proposed by exploiting improved TCN to correct the cumulative error. First, using multi-scale convolution (MSC) and self-attentiveness (SA), the problem that a single-scale convolution kernel of TCN is difficult to extract temporal and spatial features at different scales of the input sequence was optimized, and the MSC-SA-TCN model was built to recognize and extract different features exhibited by the input sequence and improve the single-step prediction of wind power to be more accurate and stable. On that basis, the multi-channel time convolutional network with multiple inputs and multiple output codec technologies was adopted to build the nonlinear mapping between the output and input of the TCN multi-step prediction, improve the problem that a single TCN is difficult to tap the different nonlinear relationships between the multi-step prediction output and the fixed input, and build the MMED-TCN multi-step wind power prediction model for separating linearity and nonlinearity between input and output and reducing the multi-step prediction error. The MMED-TCN multi-step wind power prediction model was developed to separate linearity and nonlinearity between input and output and reduce the cumulative error in the multi-step prediction. An experimental comparative analysis was conducted based on the measured data from two wind farms in Shuangzitai, Liaoning and Keqi, Inner Mongolia. As revealed from the results, MAE and RMSE of the MMED-TCN-based multi-step prediction model achieved the cumulative mean values of 0.0737 and 0.1018, respectively, and the MAE and RMSE metrics outperformed those of the VMD-AMS-TCN and MSC-SA-TCN models.