An accurate forecasting of solar global horizontal irradiance (GHI) provides the estimation of future output from photovoltaic (PV) cells, resulting in guaranteed sustainable functioning of solar integrated smart grids. Traditional forecasting algorithms find it relatively hard to obtain the intrinsic non-linearity of solar GHI, leading to poor forecasting accuracy. This work offers a deep learning (DL)-based hybrid model called RLMD BILSTM that combines robust local mean decomposition (RLMD) algorithm with bidirectional long short-term memory (BILSTM) predictor. The proposed RLMD BILSTM model has been developed and evaluated for two separate Indian locations: Hisar and Jaipur, by comparing with benchmark models: persistence, long short-term memory (LSTM), gated recurrent unit (GRU), BILSTM, and their RLMD-based hybrid variations (RLMD LSTM and RLMD GRU). The performance of the proposed model has been verified by the statistical and visual results of short-term annual and seasonal forecast of solar GHI. From the results, it has been observed that the proposed model achieved lower root mean square error (RMSE) (16.34 W/m2-35.07 W/m2) compared to contrast mforecastsodels. The higher annual R2 yielding a value of 0.977–0.995 for the proposed model compared to the contrast model proved its better fitness. Moreover, the study showed that the RLMD with BILSTM improved the RMSE (59.16%–88.88%) and mean absolute error (MAE) (66.45%–92.22%) over various contrast models. The study also conducted the sensitivity analysis using the forecasting efficacy and forecast skill to prove the significance of the proposed model.
11 March 2022 in Arabian Journal for Science and Engineering
An accurate forecasting of solar irradiation plays a vital role in grid balancing, scheduling, and maintenance. With this aim, a deep learning solar forecasting model based on dual decomposition with error correction strategy is proposed in this paper. In this model, historical time series of solar irradiance and error series are decomposed using complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) and variational mode decomposition (VMD), respectively. The bidirectional long short-term memory (BiLSTM) deep learning (DL) network is used with each series to extract the different long- and short-term characteristics of the data. To observe the forecasting accuracy of the proposed model, nine contrast models are built to estimate one to three steps ahead solar irradiance for two different locations. The experimental results demonstrate that the proposed model achieved lower RMSE (1.81 W/m2–28.46 W/m2) and MAPE (0.2–4.40%) compared to the considered models. It improves the RMSE (17.55–72.82%) and MAPE (2.65–74.66%) over a non-error correction strategy-based hybrid model. In addition, the improvement in RMSE (13.82–65.10%) and MAPE (11.46–68.33%) are observed over single decomposition-based error correction model. Moreover, to validate the forecasting performance of the proposed model Diebold–Mariano Hypothesis (DMH) test, percentage improvement and its behavior in different weather conditions are also presented.