Abstract:Due to the complexity of wind power, traditional prediction models are incapable of fully extracting the hidden features of multidimensional strong fluctuation data, which results in poor multi-step prediction performance. To predict continuous power effectively in the future, an improved wind power multi-step prediction model combining variational mode decomposition (VMD) with sequence-to-sequence (Seq2Seq) is proposed. Firstly, the wind power sequence is smoothed using VMD and the decomposition parameters of VMD are optimized by using the squirrel search algorithm (SSA) to effectively optimize the decomposition effect. Then, the subsequence obtained from decomposition, together with the original wind power data, is reconstructed into multivariate time series features. Finally, a Seq2Seq model is constructed, and convolutional neural networks (CNNs) with bidirectional gate recurrent units (BiGRUs) are used to learn the coupling and timing relationships of the input data and encode them. The gate recurrent unit (GRU) is decoded to achieve continuous power prediction. Based on the actual operating data of a wind farm, a case analysis is conducted. Experimental results show that SSA-VMD can effectively optimize the decomposition effect, and the subsequences obtained with its decomposition are highly accurate when applied to predictions. The Seq2Seq model has better multi-step prediction results than traditional prediction methods, and as the prediction step size increases, the advantages are more obvious.

What problem does this paper attempt to address?

The main problem this paper attempts to address is the poor performance of multi-step forecasting in wind power prediction. Specifically, due to the complexity of wind power generation, traditional forecasting models cannot fully extract the implicit features of multi-dimensional strong fluctuation data, leading to poor multi-step forecasting performance. To effectively predict future continuous power, this paper proposes a hybrid model combining improved Variational Mode Decomposition (VMD) and Sequence-to-Sequence (Seq2Seq) networks. ### Main Problems: 1. **Poor Multi-step Forecasting Performance**: Traditional forecasting models perform poorly in multi-step forecasting because they cannot effectively extract the implicit features of multi-dimensional strong fluctuation data. 2. **Non-stationarity and Volatility**: Wind power generation is influenced by meteorological factors, which makes it non-stationary and volatile, making accurate prediction difficult. 3. **Parameter Setting Issues**: The parameter setting of Variational Mode Decomposition (VMD) significantly affects the decomposition effect. Traditional empirical setting methods are subjective and random, affecting the prediction results. ### Solutions: 1. **Improved Variational Mode Decomposition (VMD)**: By introducing average envelope entropy as an evaluation metric and using the Squirrel Search Algorithm (SSA) to automatically find the optimal VMD decomposition parameters, the decomposition effect is improved. 2. **Sequence-to-Sequence (Seq2Seq) Model**: A new Seq2Seq model is constructed, where the encoder uses Convolutional Neural Networks (CNN) and Bidirectional Gated Recurrent Units (BiGRU) to extract coupling information and temporal information from the input data, and the decoder uses Gated Recurrent Units (GRU) for predictive output. 3. **Multi-step Forecasting**: The model can capture the dynamic changes in future wind power generation, thereby timely adjusting the dispatch plan and promoting the efficient and stable operation of the power system. ### Innovations: 1. **Optimized VMD Parameters**: Introducing average envelope entropy as an evaluation metric and using the Squirrel Search Algorithm to automatically find the optimal VMD decomposition parameters, improving the decomposition effect. 2. **Novel Seq2Seq Model**: Proposing a novel Seq2Seq model that better learns the implicit correlation features of multi-dimensional time series data through an end-to-end process, achieving effective prediction of wind power generation fluctuation trends over future time periods. 3. **Improved Encoder and Decoder Structure**: Unlike traditional models that use Recurrent Neural Networks as encoders and decoders, this paper uses CNN-BiGRU as the encoder and GRU as the decoder, fully exploring the deep correlation information between different features and the dependencies between time series data. 4. **Experimental Validation**: Testing with actual datasets demonstrates the effectiveness and robustness of the proposed model. Experimental results show that the model outperforms baseline methods in multi-step forecasting performance. In summary, this paper aims to improve the multi-step forecasting performance of wind power generation by combining improved VMD and Seq2Seq models, thereby providing more accurate forecasting information for grid dispatch and promoting large-scale grid integration of wind power.

Multi-Step Prediction of Wind Power Based on Hybrid Model with Improved Variational Mode Decomposition and Sequence-to-Sequence Network

Multi-branch wind power prediction based on optimized variational mode decomposition

An Ultra‐Short‐Term Multi‐Step Prediction Model for Wind Power Based on Sparrow Search Algorithm, Variational Mode Decomposition, Gated Recurrent Unit, and Support Vector Regression

Multi-step Wind Speed Prediction Based on an Improved Multi-Objective Seagull Optimization Algorithm and a Multi-Kernel Extreme Learning Machine.

Ultra-short term wind power prediction based on quadratic variational mode decomposition and multi-model fusion of deep learning

Short-term wind power prediction based on modal reconstruction and CNN-BiLSTM

A hybrid deep learning architecture for wind power prediction based on bi-attention mechanism and crisscross optimization

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

A Multiscale Hybrid Wind Power Prediction Model Based on Least Squares Support Vector Regression–Regularized Extreme Learning Machine–Multi-Head Attention–Bidirectional Gated Recurrent Unit and Data Decomposition

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

Ultra-short-term wind power prediction model based on fixed scale dual mode decomposition and deep learning networks

Achieving wind power and photovoltaic power prediction: An intelligent prediction system based on a deep learning approach

Short-Term Wind Power Prediction Based on CEEMDAN-SE and Bidirectional LSTM Neural Network with Markov Chain

Multistep Wind Power Prediction Using Time-Varying Filtered Empirical Modal Decomposition and Improved Adaptive Sparrow Search Algorithm-Optimized Phase Space Reconstruction–Echo State Network

Short-term wind power prediction based on ICEEMDAN decomposition and BiTCN–BiGRU-multi-head self-attention model

Deep Learning-Based Prediction of Wind Power for Multi-turbines in a Wind Farm

Multi-step wind speed forecasting based on a hybrid decomposition technique and an improved back-propagation neural network

Short-Term Multi-Step Wind Direction Prediction Based on OVMD Quadratic Decomposition and LSTM

Multiscale ultra‐short‐term wind power prediction model based on GD‐IFEM‐PSO and VMD‐BP

A multivariable hybrid prediction model of offshore wind power based on multi-stage optimization and reconstruction prediction

Multi-step short-term wind power prediction based on spatio-temporal graph convolutional networks