Abstract:The accuracy of the wind speed prediction is of crucial significance for the operation and dispatch of the power grid system reasonably. However, wind speed is so random and intermittent that the accuracy of wind speed prediction always remains unsatisfactory. Moreover, the coupling relationship between other meteorological variables and wind speed in the time and frequency domains has rarely been studied. Subsequently, a hybrid wind speed prediction model based on weather research and forecasting (WRF) simulation is proposed according to a multivariate data decomposition method and deep learning algorithm optimized by an attention mechanism and a grid search algorithm. Firstly, the WRF simulation is utilized to obtain the predicted wind speed and other meteorological variables are also extracted from WRF different domains. Furthermore, the pearson correlation coefficient (PCC) method is adopted to select principal meteorological variables as the input series. Additionally, the complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) method decomposes input series and historical data into respective intrinsic mode functions (IMFs). Then, a new hybrid deep learning model, combining a convolutional neural network (CNN) and a bidirectional long short-term memory network (BLSTM) optimized via an attention mechanism (AM) and a grid search method (GS), is proposed to predict the error and correct the wind speed from WRF innermost domain. Finally, the validation case study is conducted to verify the effectiveness of the proposed model. The results indicate that the proposed model outperforms other comparative models in terms of single-step and multi-step wind speed prediction. Specifically, the values of the mean absolute error (MAE), the mean absolute percentage error (MAPE), and the root mean square error (RMSE) are 0.1042 m/s, 4.63% and 0.1309 m/s after correction, decreased by 94.13%, 91.75% and 93.93%, respectively, compared to those without correction.

Short-term wind speed prediction: Hybrid of ensemble empirical mode decomposition, feature selection and error correction

Short-Term Wind Speed Forecasting Using Empirical Mode Decomposition and Feature Selection

Wind Speed Multi-Step Forecasting Using High-Resolution and Multi-Objective Ensemble Model with Error Correction

Wind Speed Short-Term Prediction Based on Empirical Wavelet Transform, Recurrent Neural Network and Error Correction

A New Hybrid Framework for Probabilistic Wind Speed Prediction Using Deep Feature Selection and Multi-Error Modification

Decomposition-Selection-Ensemble Prediction System for Short-Term Wind Speed Forecasting

A new two-stage decomposition and integrated hybrid model for short-term wind speed prediction

A Short-Term Wind Speed Prediction Method Utilizing Novel Hybrid Deep Learning Algorithms to Correct Numerical Weather Forecasting

Short-Term Wind Speed Prediction Using EEMD-LSSVM Model

Short-Term Wind Speed Forecasting Based on Signal Decomposing Algorithm and Hybrid Linear/Nonlinear Models

Short-term Wind Speed Combined Prediction for Wind Farms

A Hybrid System for Short-Term Wind Speed Forecasting

A new decomposition-ensemble strategy fusion with correntropy optimization learning algorithms for short-term wind speed prediction

A Novel Hybrid Forecasting Model with Feature Selection and Deep Learning for Wind Speed Research

A Novel Wind Speed Prediction Method: Hybrid of Correlation-Aided DWT, LSSVM and GARCH

An Ensemble Model of Wind Speed Forecasting Based on Variational Mode Decomposition and Bare-Bones Fireworks Algorithm

An adaptive hybrid model for short term wind speed forecasting

A Novel Hybrid Model for Short-Term Wind Power Forecasting

Ensemble optimization approach based on hybrid mode decomposition and intelligent technology for wind power prediction system

Hybrid wind speed forecasting model based on multivariate data secondary decomposition approach and deep learning algorithm with attention mechanism

A novel ensemble system for short-term wind speed forecasting based on hybrid decomposition approach and artificial intelligence models optimized by self-attention mechanism