Multi-dimensional wind power prediction based on time-series characterization analysis and VMD-EMD quadratic decomposition DOI
Zhe Zhang, Jing Gao, Yongsheng Wang

et al.

Published: July 21, 2023

Wind power data receive wind volatility and have strong non-smoothness, making it difficult to achieve high accuracy in prediction. To address this challenge, paper proposes a multi-step prediction model combining VMD (Variational Modal Decomposition), EMD (Empirical Hurst analysis temporal entropy values. Firstly, the first decomposition of historical is carried out by ; then performed on components decomposition, with low regularity are decomposed twice using decomposition; second further filtered permutation entropy, values compared The secondary formed into randomness irregular component, regular component component; for three types components, BP neural network used predict them respectively, they reorganized experiments prove that proposed has higher faster running time than current mainstream models, can more efficient

Language: Английский

A dual-scale deep learning model based on ELM-BiLSTM and improved reptile search algorithm for wind power prediction DOI

Jinlin Xiong,

Peng Tian,

Zihan Tao

et al.

Energy, Journal Year: 2022, Volume and Issue: 266, P. 126419 - 126419

Published: Dec. 13, 2022

Language: Английский

Citations

154
Enhancing short-term wind speed prediction based on an outlier-robust ensemble deep random vector functional link network with AOA-optimized VMD DOI
Chu Zhang, Zhengbo Li,

Yida Ge

et al.

Energy, Journal Year: 2024, Volume and Issue: 296, P. 131173 - 131173

Published: April 1, 2024

Language: Английский

Citations

23
Prediction of NOx emission concentration from coal-fired power plant based on joint knowledge and data driven DOI
Zheng Wu, Yue Zhang, Ze Dong

et al.

Energy, Journal Year: 2023, Volume and Issue: 271, P. 127044 - 127044

Published: March 2, 2023

Language: Английский

Citations

37
Short-term wind power prediction based on modal reconstruction and CNN-BiLSTM DOI Creative Commons
Zheng Li,

Ruosi Xu,

Xiaorui Luo

et al.

Energy Reports, Journal Year: 2023, Volume and Issue: 9, P. 6449 - 6460

Published: June 16, 2023

Accurate prediction of short-term wind power plays an important role in the safe operation and economic dispatch grid. In response to current single algorithm that cannot further improve accuracy, this study proposes a combined model based on data processing, signal decomposition, deep learning. First, outliers original can affect accuracy. This detects by Z-score method fills them with cubic spline interpolation ensure integrity data. Second, for volatility power, time series is decomposed using complete ensemble empirical modal decomposition adaptive noise (CEEMDAN). The component complexity calculated sample entropy (SE), components are reconstructed according SE size Finally, traditional convolutional neural network (CNN) structure improved bi-directional long memory (BiLSTM) used extract feature links between superimpose results each obtain final value. experimental demonstrate hybrid proposed has better performance terms performance.

Language: Английский

Citations

25
New formulation for predicting total dissolved gas supersaturation in dam reservoir: application of hybrid artificial intelligence models based on multiple signal decomposition DOI Creative Commons
Salim Heddam, Ahmed M. Al‐Areeq, Mou Leong Tan

et al.

Artificial Intelligence Review, Journal Year: 2024, Volume and Issue: 57(4)

Published: March 9, 2024

Abstract Total dissolved gas (TDG) concentration plays an important role in the control of aquatic life. Elevated TDG can cause gas-bubble trauma fish (GBT). Therefore, controlling fluctuation has become great importance for different disciplines surface water environmental engineering.. Nowadays, direct estimation is expensive and time-consuming. Hence, this work proposes a new modelling framework predicting based on integration machine learning (ML) models multiresolution signal decomposition. The proposed ML were trained validated using hourly data obtained from four stations at United States Geological Survey. dataset are composed from: ( i ) temperature T w ), ii barometric pressure BP iii discharge Q which used as input variables prediction. strategy conducted two steps. First, six singles model namely: multilayer perceptron neural network, Gaussian process regression, random forest iv vector functional link, v adaptive boosting, vi Bootstrap aggregating (Bagging), developed , their performances compared. Second, was introduced combination empirical mode decomposition (EMD), variational (VMD), wavelet transform (EWT) preprocessing algorithms with building hybrid models. signals decomposed to extract intrinsic functions (IMFs) by EMD VMD methods analysis (MRA) components EWT method. Then after, IMFs MRA selected regraded integral part thereof. single prediction compared several statistical metrics namely, root mean square error, absolute coefficient determination R 2 Nash–Sutcliffe efficiency (NSE). times high number repetitions, depending kind modeling process. results gave good agreement between predicted situ measured dataset. Overall, Bagging performed better than other five NSE values 0.906 0.902, respectively. However, extracted EMD, have contributed improvement models’ performances, significantly increased reaching 0.996 0.995. Experimental showed superiority more importantly improving predictive accuracy TDG. Graphical abstract

Language: Английский

Citations

5
1-Hour Ahead Wind Power Prediction Based on Multi-model Fusion Strategy DOI
Tengyu Zhang, Shuai Di, Xu‐Sheng Wang

et al.

Lecture notes in electrical engineering, Journal Year: 2025, Volume and Issue: unknown, P. 263 - 274

Published: Jan. 1, 2025

Language: Английский

Citations

0
Wind power prediction with accurate identification of ramp events based on Interval-SMOTE oversampling and ensemble learning DOI
Ying Han,

Xuemeng Wang,

Kun Li

et al.

Wind Engineering, Journal Year: 2025, Volume and Issue: unknown

Published: Jan. 30, 2025

Wind power ramp events (WPREs) are small probability with serious wind fluctuations, and it is one of the important factors leading to security accidents in grid. Firstly, given small-sample nature WPREs, this paper introduces an Interval-SMOTE oversampling method increase data points for events; generated confined within a dynamically adjusted interval that evolves each iteration, thereby ensuring maximum preservation original trends. Then, order improve detection efficiency integration Swinging Door Trending (SDT) algorithm proposed accurately identify existing non-ramp sequence. Moreover, considering different types two modeling methods Stochastic Configuration Networks (SCNs) Bidirectional Long Short-term Memory (BiLSTM) employed handle problem. Due stochastic configuration supervised mechanism key parameters, SCNs can provide significant advantages handling large samples, so applied build model as unique structures bidirectional processing information, BiLSTM has better ability mining sample events. The prediction results from models then weighted obtain final results. Experimental demonstrate sampling enhances accuracy metrics by 0.43% 3.72% farms; specifically, regarding measured RMSE, SCNs-BiLSTM outperforms comparative 3.88% 15.49% across various farms.

Language: Английский

Citations

0
A Short-Term Wind Power Forecasting Model Based on 3D Convolutional Neural Network–Gated Recurrent Unit DOI Open Access
Xiaoshuang Huang,

Yinbao Zhang,

Jianzhong Liu

et al.

Sustainability, Journal Year: 2023, Volume and Issue: 15(19), P. 14171 - 14171

Published: Sept. 25, 2023

Enhancing the accuracy of short-term wind power forecasting can be effectively achieved by considering spatial–temporal correlation among neighboring turbines. In this study, we propose a model based on 3D CNN-GRU. First, data and meteorological 24 surrounding turbines around target turbine are reconstructed into three-dimensional matrix inputted CNN GRU encoders to extract their features. Then, predictions for different horizons outputted through decoder fully connected layers. Finally, experimental results SDWPT datasets show that our proposed significantly improves prediction compared BPNN, GRU, 1D CNN-GRU models. The performs optimally. For horizon 10 min, average reductions in RMSE MAE validation set about 10% 11%, respectively, with an improvement 1% R. 120 6% 8%, 14%

Language: Английский

Citations

6
Sparrow Search Optimization with Deep Belief Network based Wind Power Prediction Model DOI

M. Aruna,

Patel Badari Narayana,

S. Narasimha Kumar

et al.

Published: Jan. 5, 2023

Wind power is a clear feature of intermittent, nonstationary, and difficult fluctuations, making it challenging for achieving consistent wind generation. Assuming the restricted nature typical energy resources developing difficulties environmental problems, several countries are starting with novel which considered renewable safety. Amongst resources, was abundant, doesn't cause pollution, has minimum cost, does not deplete. Accurate predictive enhance reliability safety grid function. Therefore, this study presents sparrow search optimization deep belief network prediction (SSODBN-WPP) technique. The SSODBN-WPP technique follows two stage process namely parameter tuning. At initial stage, employs DBN method prediction. Next, SSO algorithm used to adjust core hyperparameters algorithm. efficacy tested on comprehensive set simulations that take place dataset. A comparison reported its betterment over other approaches.

Language: Английский

Citations

3
Wind-storage combined system based on just-in-time-learning prediction model with dynamic error compensation DOI Open Access
Wei Yang, Li Jia, Yue Xu

et al.

Journal of Energy Storage, Journal Year: 2023, Volume and Issue: 68, P. 107658 - 107658

Published: May 22, 2023

Language: Английский

Citations

3