Search

Reliable and accurate prediction models effectively deal with nonlinear and non-stationary characteristics of reservoir inflow and associated variables. Therefore, a Hybrid Deep Learning Inflow Prediction-Rolling Window (HDeepLIP-RW) framework is proposed in this study for efficient inflow prediction. The HDeepLIP-RW framework utilizes hybrid alternatives of four different RNN architectures bidirectional long short-term memory (BiLSTM), gated recurrent unit (GRU), long short-term memory (LSTM), and simple RNN (sRNN) by employing RW technique. Firstly, the conventional split data (CSD) technique is applied to divide the multivariate hydro-meteorological data of Ermenek Dam, Turkey, into training, validation, and testing periods. Afterward, the processed data is employed as input by standalone and hybrid models for inflow prediction. Thirdly, the RW technique is used to divide the data to train and validate the model. Finally, different standalone and HDeepLIP-RW prediction models utilizing the RW technique are also developed to compare the performance of the proposed model with counterparts. The prediction results in terms of statistical metrics and different plots manifest the applicability of models utilizing RW for inflow prediction in contrast to the CSD-built models. Furthermore, the HDeepLIP-RW based BiLSTM-GRU model outperforms all the counterpart models by demonstrating the highest performance and accuracy. The RW-based BiLSTM-GRU model reduces RMSE contrast to the CSD-based best performing by 0.300 for GRU-LSTM during the testing period. Likewise, the prediction accuracy according to NS was increased by 23.37% compared to standalone models and 19.26% according to hybrid methods by splitting the data with the RW method. The overall results are enlightening and substantiate the viability of the BiLSTM-GRU model utilizing the HDeepLIP-RW framework for inflow prediction with an equal opportunity for similar applications. [ABSTRACT FROM AUTHOR]

Accurate prediction models enable the efficacious utilization and integration of solar energy into the power system. Therefore, this study aims to develop novel hybrid prediction models by employing correlation analysis (CA), decomposition techniques, sample entropy (SE), and spatio-temporal attention (STA) based sequence2sequence (S2S) algorithm for accurate prediction of global horizontal irradiance (GHI). The decomposition techniques include variational mode decomposition (VMD), improved complete ensemble empirical mode decomposition with additive noise (ICEEMDAN), and maximum overlap discrete wavelet transform (MODWT). The VMD-STA-S2S hybrid model surmounts the associated hybrid and standalone prediction models by revealing the highest prediction efficiency and the lowest error. Compared to the SARIMAX, SVR, ANN, XGB, GRU, LSTM, and S2S models, the VMD-STA-S2S model reduced RMSE during testing by 80.927 W/m2, 75.426 W/m2, 73.487 W/m2, 62.394 W/m2, 57.811 W/m2, 52.007 W/m2, and 41.836 W/m2, respectively. Similarly, the reductions in RMSE by VMD-STA-S2S model compared to SA-S2S, TA-S2S, STA-S2S, MODWT-STA-S2S, ICEEMDAN-STA-S2S, ICEEMDAN-SE-STA-S2S, and VMD-SE-STA-S2S models are 24.054 W/m2, 20.951 W/m2, 12.702 W/m2, 15.396 W/m2, 9.921 W/m2, 6.103 W/m2 and 0.484 W/m2, respectively, during testing. Furthermore, considering NSE during testing, the VMD-STA-S2S model is 8.66%, 7.72%, 7.41%, 5.71%, 5.07%, 4.31%, 3.11%, 1.43%, 1.19%, 0.64%, 0.81%, 0.47%, 0.35% and 0.07%, more efficient than the SARIMAX, SVR, ANN, XGB, GRU, LSTM, S2S, SA-S2S, TA-S2S, STA-S2S, MODWT-STA-S2S, ICEEMDAN-STA-S2S, ICEEMDAN-SE-STA-S2S, and VMD-SE-STA-S2S models, respectively. The superior performance of VMD-STA-S2S over its counterparts corroborates the integration of the VMD technique and STA-based S2S algorithm for GHI prediction. The multivariate meteorological data of this study is decomposed by VMD into subcomponents more effectively than the ICEEMDAN and MODWT techniques. VMD decomposed subcomponents are further fed to the STA-S2S to efficiently extract and learn the spatial and temporal features, resulting in the enhanced and superior prediction outcomes of the VMD-STA-S2S model compared to all the counterpart models. Besides GHI prediction, the proposed model is also appropriate for other time-series data, including renewable energy, electrical load, and environment monitoring. [Display omitted] [ABSTRACT FROM AUTHOR]

The accuracy and consistency of streamflow prediction play a significant role in several applications involving the management of hydrological resources, such as power generation, water supply, and flood mitigation. However, the nonlinear dynamics of the climatic factors jeopardize the development of efficient prediction models. Therefore, to enhance the reliability and accuracy of streamflow prediction, this paper developed a three-stage hybrid model, namely, IVL (ICEEMDAN-VMD-LSTM), which integrated improved complete ensemble empirical mode decomposition with additive noise (ICEEMDAN), variational mode decomposition (VMD), and long short-term memory (LSTM) neural network. Monthly data series of streamflow, temperature, and precipitation in the Swat River Watershed, Pakistan, from January 1971 to December 2015 was used as a case study. Firstly, the correlation analysis and the two-stage decomposition approach were employed to select suitable inputs for the proposed model. ICEEMDAN was employed as a first decomposition stage, to decompose the three data series into intrinsic mode functions (IMFs) and a residual component. In the second decomposition stage, the component of high frequency (IMF1) was decomposed by VMD, as the second decomposition. Afterward, all the components obtained through the correction analysis and the two-stage decomposition approach were predicted by using the LSTM network. Finally, the predicted results of all components were aggregated, to formulate an ensemble prediction for the original monthly streamflow series. The predicted results showed that the performance of the proposed model was superior to the other developed models, in respect of several evaluation benchmarks, demonstrating the applicability of the proposed IVL model for monthly streamflow prediction. [ABSTRACT FROM AUTHOR]

[Display omitted] • Streamflow forecasting is indispensable for optimal exploitation of water resources. • A hybrid-forecasting model, namely ISGS (ICEEMDAN-SE-GI-Seq2Seq) is proposed. • ICEEMDAN, Sample Entropy, Gini Index and Sequence to Sequence approaches were used. • CC, NS, and d values of ISGS approach were 0.974, 0.949, 0.987; respectively. Accurate and reliable streamflow forecasting is indispensable to deal with the dynamics of streamflow parameters and for optimal use of water resources, flood, and drought control. In this study, a hybrid-forecasting model integrating improved complete ensemble empirical mode decomposition with additive noise (ICEEMDAN)-Sample Entropy (SE)-Gini Index (GI)-Sequence to Sequence (Seq2Seq) for streamflow forecasting, namely ISGS (ICEEMDAN-SE-GI-Seq2Seq), is proposed. Firstly, ICEEMDAN is employed to decompose the streamflow time series into subcomponents that are further combined based on SE values. Afterward, GI is used to select the most appropriate features to decrease the computational burden and ameliorate the model's performance. Finally, Seq2Seq architecture based on the long short-term memory (LSTM) algorithm is utilized for predicting the multivariate hydro-meteorological data at three stations located in Turkey. Besides ISGS, the ICEEMDAN-SE-Seq2Seq, DWT-GI-Seq2Seq, DWT-Seq2Seq, and standalone Seq2Seq models are also constructed as counterpart comparison models. The simulation under different scenarios, several statistical metrics, and judicious plots; substantiate ISGS validity for streamflow forecasting. The correlation coefficient, Nash-Sutcliffe coefficient of efficiency, and Willmott's index values presented by the ISGS model are 0.974, 0.949, and 0.987 for Kirkgoze station, 0.928, 0.839, and 0.947 Kosk station, 0.888, 0.769, and 0.931 for Karagobek station. The performance revealed by the proposed ISGS model proves its viability for streamflow forecasting, along with equal applicability for the other time-series forecasting tasks. [ABSTRACT FROM AUTHOR]