A novel recurrent neural network approach in forecasting short term solar irradiance. (February 2022)

• Record Type: Journal Article
• Authors: Jaihuni, Mustafa; Basak, Jayanta Kumar; Khan, Fawad; Okyere, Frank Gyan; Sihalath, Thavisak; Bhujel, Anil; Park, Jihoon; Lee, Deog Hyun; Kim, Hyeon Tae
• Is Part Of: ISA transactions. Volume 121(2022)
• Page Start: 63
• Page End: 74
• Publication Date: 2022-02
• Subjects: Unidirectional -- Bidirectional -- LSTM -- GRU -- Solar irradiance -- Short term forecasting; Engineering instruments -- Periodicals; Engineering instruments; Periodicals; Electronic journals; 629.805
• DOI: 10.1016/j.isatra.2021.03.043
• Languages: English
• ISSNs: 0019-0578

Abstract

Abstract: Forecasting solar irradiance is of utmost importance in supplying renewable energy efficiently and timely. This paper aims to experiment five variants of recurrent neural networks (RNN), and develop effective and reliable 5-minute short term solar irradiance prediction models. The 5 RNN classes are long–short term memory (LSTM), gated recurrent unit (GRU), Simple RNN, bidirectional LSTM (Bi-LSTM), and bidirectional GRU (Bi-GRU); the first 3 classes are unidirectional and the last two are bidirectional RNN models. The 26 months data under consideration, exhibits extremely volatile weather conditions in Jinju city, South Korea. Therefore, after different experimental processes, 5 hyper-parameters were selected for each model cautiously. In each model, different levels of depth and width were tested; moreover, a 9-fold cross validation was applied to distinguish them against high variability in the seasonal time-series dataset. Generally the deeper architectures of the aforementioned models had significant outcomes; meanwhile, the Bi-LSTM and Bi-GRU provided more accurate predictions as compared to the unidirectional ones. The Bi-GRU model provided the lowest RMSE and highest R 2 values of 46.1 and 0.958; additionally, it required 5.25*10 −5 seconds per trainable parameter per epoch, the lowest incurred computational cost among the mentioned models. All 5 models performed differently over the four seasons in the 9-fold cross validation test. On average, the bidirectional RNNs and the simple RNN model showed high robustness with less data and high temporal data variability; although, the stronger architectures of the bidirectional models, deems their results more reliable. Highlights: Forecasting solar irradiance is important in efficiently supplying renewable energy. Five variants of RNN models were developed to predict short term solar irradiance. The RNN models performed differently in the 9-fold cross validation test. Bidirectional RNN models provided better predictions than the Unidirectional ones.