Hourly solar irradiance forecasting based on encoder-decoder model using series decomposition and dynamic error compensation

被引:25
作者
Tong, Junlong
Xie, Liping [1 ,2 ]
Fang, Shixiong
Yang, Wankou
Zhang, Kanjian
机构
[1] Southeast Univ, Sch Automat, Nanjing 210096, Peoples R China
[2] Minist Educ, Key Lab Measurement & Control Complex Syst Engn, Nanjing 210096, Peoples R China
关键词
Solar irradiance forecasting; Deep learning; Temporal convolutional network; Encoder-decoder; Long short term memory; Error compensation; PREDICTION;
D O I
10.1016/j.enconman.2022.116049
中图分类号
O414.1 [热力学];
学科分类号
摘要
Accurate solar irradiance prediction is crucial for harnessing solar energy resources. However, the pattern of irradiance sequence is intricate due to its nonlinear and non-stationary characteristics. In this paper, a deep hybrid model based on encoder-decoder is proposed to cope with the complex pattern for hourly irradiance forecasting. The hybrid deep model integrates complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN), encoder-decoder module, and dynamic error compensation (DEC) architecture. The CEEMDAN is implemented to reduce the nonlinear and non-stationarity of the irradiance sequence. The encoder-decoder integrates temporal convolutional networks (TCN), long short-term memory networks (LSTM), and multi-layer perceptron (MLP) for temporal features extraction and multi-step prediction. The DEC architecture dynamically updates the model based on adjacent error information to mine the predictable components of error information. Furthermore, a new loss function is further proposed for multi-objective optimization to balance the performance of multi-step forecasting. In the hourly irradiance forecasting experiments on the three public datasets, the root mean square error (RMSE), mean absolute error (MAE), and correlation coefficient (R) of the proposed model are observed to be in a range of 30.693-34.433 W/m2, 19.398-22.900 W/m2, and 0.9872-0.9902, respectively. Compared to the benchmark models (including MLP, LSTM, and TCN), the RMSE and MAE reduce by 10.76%-22.00% and 5.47%-20.40%, respectively. The experimental results indicate that the proposed model shows accurate and robust forecasting performance and is a reliable alternative to hourly irradiance forecasting.
引用
收藏
页数:18
相关论文
共 51 条
  • [1] A novel approach based on integration of convolutional neural networks and deep feature selection for short-term solar radiation forecasting
    Acikgoz, Hakan
    [J]. APPLIED ENERGY, 2022, 305
  • [2] [Anonymous], MEASUREMENT INSTRUME
  • [3] Bai SJ, 2018, Arxiv, DOI [arXiv:1803.01271, DOI 10.48550/ARXIV.1803.01271]
  • [4] Neural network based method for conversion of solar radiation data
    Celik, Ali N.
    Muneer, Tariq
    [J]. ENERGY CONVERSION AND MANAGEMENT, 2013, 67 : 117 - 124
  • [5] Cho KYHY, 2014, Arxiv, DOI arXiv:1406.1078
  • [6] Probabilistic forecasting of the solar irradiance with recursive ARMA and GARCH models
    David, M.
    Ramahatana, F.
    Trombe, P. J.
    Lauret, P.
    [J]. SOLAR ENERGY, 2016, 133 : 55 - 72
  • [7] Nonlinear combination method of forecasters applied to PM time series
    de Mattos Neto, Paulo S. G.
    Cavalcanti, George D. C.
    Madeiro, Francisco
    [J]. PATTERN RECOGNITION LETTERS, 2017, 95 : 65 - 72
  • [8] A wavelet-coupled support vector machine model for forecasting global incident solar radiation using limited meteorological dataset
    Deo, Ravinesh C.
    Wen, Xiaohu
    Qi, Feng
    [J]. APPLIED ENERGY, 2016, 168 : 568 - 593
  • [9] Review of solar irradiance forecasting methods and a proposition for small-scale insular grids
    Diagne, Maimouna
    David, Mathieu
    Lauret, Philippe
    Boland, John
    Schmutz, Nicolas
    [J]. RENEWABLE & SUSTAINABLE ENERGY REVIEWS, 2013, 27 : 65 - 76
  • [10] DKASC, AL SPRING SYST