Multi-Task Autoencoders and Transfer Learning for Day-Ahead Wind and Photovoltaic Power Forecasts

被引:5
|
作者
Schreiber, Jens [1 ]
Sick, Bernhard [1 ]
机构
[1] Univ Kassel, Intelligent Embedded Syst, Wilhelmshoher Allee 71, D-34121 Kassel, Germany
关键词
transfer learning; wind power; photovolatic power; autoencoders; deep learning; time series; PREDICTION; NETWORK; REGRESSION;
D O I
10.3390/en15218062
中图分类号
TE [石油、天然气工业]; TK [能源与动力工程];
学科分类号
0807 ; 0820 ;
摘要
Integrating new renewable energy resources requires robust and reliable forecasts to ensure a stable electrical grid and avoid blackouts. Sophisticated representation learning techniques, such as autoencoders, play an essential role, as they allow for the extraction of latent features to forecast the expected generated wind and photovoltaic power for the next seconds up to days. Thereby, autoencoders reduce the required training time and the time spent in manual feature engineering and often improve the forecast error. However, most current renewable energy forecasting research on autoencoders focuses on smaller forecast horizons for the following seconds and hours based on meteorological measurements. At the same time, larger forecast horizons, such as day-ahead power forecasts based on numerical weather predictions, are crucial for planning loads and demands within the electrical grid to prevent power failures. There is little evidence on the ability of autoencoders and their respective forecasting models to improve through multi-task learning and time series autoencoders for day-ahead power forecasts. We can close these gaps by proposing a multi-task learning autoencoder based on the recently introduced temporal convolution network. This approach reduces the number of trainable parameters by 38 for photovoltaic data and 202 for wind data while having the best reconstruction error compared to nine other representation learning techniques. At the same time, this model decreases the day-ahead forecast error up to 18.3% for photovoltaic parks and 1.5% for wind parks. We round off these results by analyzing the influences of the latent size and the number of layers to fine-tune the encoder for wind and photovoltaic power forecasts.
引用
收藏
页数:20
相关论文
共 50 条
  • [1] A Day-Ahead Photovoltaic Power Prediction via Transfer Learning and Deep Neural Networks
    Miraftabzadeh, Seyed Mahdi
    Colombo, Cristian Giovanni
    Longo, Michela
    Foiadelli, Federica
    FORECASTING, 2023, 5 (01): : 213 - 228
  • [2] Combining forecasts of day-ahead solar power
    Dewangan, Chaman Lal
    Singh, S. N.
    Chakrabarti, S.
    ENERGY, 2020, 202
  • [3] Validation of ANN Training Approaches for Day-Ahead Photovoltaic Forecasts
    Nespoli, Alfredo
    Ogliari, Emanuele
    Dolara, Alberto
    Grimaccia, Francesco
    Leva, Sonia
    Mussetta, Marco
    2018 INTERNATIONAL JOINT CONFERENCE ON NEURAL NETWORKS (IJCNN), 2018,
  • [4] The effect of wind and solar power forecasts on day-ahead and intraday electricity prices in Germany
    Guertler, Marc
    Paulsen, Thomas
    ENERGY ECONOMICS, 2018, 75 : 150 - 162
  • [5] Prediction Intervals for Day-Ahead Photovoltaic Power Forecasts with Non-Parametric and Parametric Distributions
    Fonseca, Joao Gari da Silva, Jr.
    Ohtake, Hideaki
    Oozeki, Takashi
    Ogimoto, Kazuhiko
    JOURNAL OF ELECTRICAL ENGINEERING & TECHNOLOGY, 2018, 13 (04) : 1504 - 1514
  • [6] An adaptive deep learning framework for day-ahead forecasting of photovoltaic power generation
    Luo, Xing
    Zhang, Dongxiao
    SUSTAINABLE ENERGY TECHNOLOGIES AND ASSESSMENTS, 2022, 52
  • [7] An improved spatial upscaling method for producing day-ahead power forecasts for wind farm clusters
    Yang, Mao
    Yan, Qi
    Dai, Bozhi
    Chen, Xinxin
    Ma, Miaomiao
    Su, Xin
    IET GENERATION TRANSMISSION & DISTRIBUTION, 2022, 16 (19) : 3860 - 3873
  • [8] Incorporating Spatial and Temporal Correlations to Improve Aggregation of Decentralized Day-Ahead Wind Power Forecasts
    Mararakanye, Ndamulelo
    Dalton, Amaris
    Bekker, Bernard
    IEEE ACCESS, 2022, 10 : 116182 - 116195
  • [9] Wind power forecasting using ensemble learning for day-ahead energy trading
    Suarez-Cetrulo, Andres L.
    Burnham-King, Lauren
    Haughton, David
    Carbajo, Ricardo Simon
    RENEWABLE ENERGY, 2022, 191 : 685 - 698
  • [10] Distributed Reconciliation in Day-Ahead Wind Power Forecasting
    Bai, Li
    Pinson, Pierre
    ENERGIES, 2019, 12 (06)