An attention mechanism based deep nonlinear ensemble paradigm of strengthened feature extraction method for wind power prediction DOI Open Access
Jujie Wang,

Yafen Liu

Journal of Renewable and Sustainable Energy, Год журнала: 2023, Номер 15(6)

Опубликована: Ноя. 1, 2023

The inherent uncertainty of wind power always hampers difficulties in the development energy and smooth operation systems. Therefore, reliable ultra-short-term prediction is crucial for energy. In this research, a two-stage nonlinear ensemble paradigm based on double-layer decomposition technology, feature reconstruction, intelligent optimization algorithm, deep learning suggested to increase accuracy power. First, using two different signal techniques processing can further filter out noise original fully capture features within it. Second, multiple components obtained through double are reconstructed sample entropy theory reassembled into several subsequences with similar complexity simplify input variables model. Finally, idea strategy, cuckoo search algorithm attention mechanism optimized long- short-term memory model applied integration, respectively, obtain final results. Two sets data from farms Liaoning Province, China used simulation experiments. empirical findings indicate that, comparison other models, has greater accuracy.

Язык: Английский

Wind Speed Prediction Using Chicken Swarm Optimization with Deep Learning Model DOI Creative Commons

R. Surendran,

Youseef Alotaibi, Ahmad F. Subahi

и другие.

Computer Systems Science and Engineering, Год журнала: 2023, Номер 46(3), С. 3371 - 3386

Опубликована: Янв. 1, 2023

High precision and reliable wind speed forecasting have become a challenge for meteorologists. Convective events, namely, strong winds, thunderstorms, tornadoes, along with large hail, are natural calamities that disturb daily life. For accurate prediction of overcoming its uncertainty change, several approaches been presented over the last few decades. As series higher volatility nonlinearity, it is urgent to present cutting-edge artificial intelligence (AI) technology. In this aspect, paper presents an intelligent using chicken swarm optimization hybrid deep learning (IWSP-CSODL) method. The IWSP-CSODL model estimates hyperparameter optimizer. model, process performed via convolutional neural network (CNN) based long short-term memory autoencoder (CBLSTMAE) model. To optimally modify hyperparameters related CBLSTMAE (CSO) algorithm utilized thereby reduces mean square error (MSE). experimental validation tested data under three distinct scenarios. comparative study pointed out better outcomes other recent models.

Язык: Английский

Процитировано

28
A novel decision support system for enhancing long-term forecast accuracy in virtual power plants using bidirectional long short-term memory networks DOI Creative Commons
Reza Nadimi, Mika Goto

Applied Energy, Год журнала: 2025, Номер 382, С. 125273 - 125273

Опубликована: Янв. 13, 2025

Язык: Английский

Процитировано

2
A parallel differential learning ensemble framework based on enhanced feature extraction and anti-information leakage mechanism for ultra-short-term wind speed forecast DOI
Jujie Wang,

Yafen Liu,

Yaning Li

и другие.

Applied Energy, Год журнала: 2024, Номер 361, С. 122909 - 122909

Опубликована: Фев. 27, 2024

Язык: Английский

Процитировано

9
Short-term wind speed forecasting based on two-stage preprocessing method, sparrow search algorithm and long short-term memory neural network DOI Creative Commons

Xue-Yi Ai,

Shijia Li, Haoxuan Xu

и другие.

Energy Reports, Год журнала: 2022, Номер 8, С. 14997 - 15010

Опубликована: Ноя. 1, 2022

Wind energy, as an environment-friendly and renewable energy source, has become one of the most effective alternatives to conventional power sources. However, intermittent nature wind speed interference noise signal bring several challenges safety reliability grid operation. To tackle this issue, a two-stage preprocessing strategy is designed, short-term prediction model based on long memory (LSTM) proposed. Firstly, singular spectrum analysis (SSA) introduced extract target data filter data. Next, denoised sequence decomposed by variational mode decomposition (VMD) into multiple intrinsic functions (IMFs), which are further aggregated sample entropy (SE). Besides, hyper-parameters LSTM neural network optimized newly sparrow search algorithm (SPSA) possessing excellent global optimization ability. Subsequently, sequences coupled with SPSA-LSTM modules synchronously. The ultimate forecasting results obtained superimposing predicted values all sequences. In order evaluate effectiveness proposed approach, two case studies conducted datasets collected from different sites 10-min 1-hour intervals comparing seven relevant models. experimental demonstrate that SSA-VMD-SE-SPSA-LSTM can adequately inherent features series, thus achieving higher accuracy.

Язык: Английский

Процитировано

25
A Study of Optimization in Deep Neural Networks for Regression DOI Open Access
Chieh-Huang Chen, Jung-Pin Lai, Yu-Ming Chang

и другие.

Electronics, Год журнала: 2023, Номер 12(14), С. 3071 - 3071

Опубликована: Июль 14, 2023

Due to rapid development in information technology both hardware and software, deep neural networks for regression have become widely used many fields. The optimization of (DNNR), including selections data preprocessing, network architectures, optimizers, hyperparameters, greatly influence the performance tasks. Thus, this study aimed collect analyze recent literature surrounding DNNR from aspect optimization. In addition, various platforms conducting models were investigated. This has a number contributions. First, it provides sections models. Then, elements each section are listed analyzed. Furthermore, delivers insights critical issues related Optimizing simultaneously instead individually or sequentially could improve Finally, possible potential directions future provided.

Язык: Английский

Процитировано

17
Multi-branch wind power prediction based on optimized variational mode decomposition DOI Creative Commons

Bangru Xiong,

Xinyu Meng,

Gang Xiong

и другие.

Energy Reports, Год журнала: 2022, Номер 8, С. 11181 - 11191

Опубликована: Сен. 7, 2022

Wind power has obvious characteristics of non-stationary, intermittent, and complex fluctuations, making it difficult to achieve reliable wind generation. This brings great challenges the safe stable operation grid regulation, so accurate prediction is very important. In this paper, we proposed a method for based on optimized variational mode decomposition (VMD) deep learning algorithm nonlinear weighted combination. Due low adaptability VMD, paper adopted whale optimization (WOA) automatically optimize core parameters VMD. The decomposed components historical were spliced form composite vector,and Convolutional Neural Network (CNN) Gated Recurrent Unit (GRU) used extract local feature global trend respectively. Finally, obtained features fused predict future power. experimental results showed that accuracy been greatly improved, compared with existing single combined forecasting methods, error within an acceptable range.

Язык: Английский

Процитировано

15
Forecasting and Multilevel Early Warning of Wind Speed Using an Adaptive Kernel Estimator and Optimized Gated Recurrent Units DOI Creative Commons
Pengjiao Wang,

Qiuliang Long,

Hu Zhang

и другие.

Mathematics, Год журнала: 2024, Номер 12(16), С. 2581 - 2581

Опубликована: Авг. 21, 2024

Accurately predicting wind speeds is of great significance in various engineering applications, such as the operation high-speed trains. Machine learning models are effective this field. However, existing studies generally provide deterministic predictions and utilize decomposition techniques advance to enhance predictive performance, which may encounter data leakage fail capture stochastic nature data. This work proposes an advanced framework for prediction early warning by combining optimized gated recurrent unit (GRU) adaptive kernel density estimator (AKDE). Firstly, 12 samples (26,280 points each) were collected from extensive open database. Three representative metaheuristic algorithms then employed optimize parameters diverse models, including extreme machines, a transformer model, networks. The results yielded optimal selection using GRU crested porcupine optimizer. Afterwards, AKDE, joint probability cumulative distribution function related errors could be obtained. It was applicable calculate conditional that actual speed exceeds critical value, thereby providing probabilistic-based multilevel manner. A comparison performance methods accuracy subsequent decisions validated proposed framework.

Язык: Английский

Процитировано

3
Uncertainty reduction in power forecasting of virtual power plant: from day-ahead to balancing markets DOI Creative Commons
Reza Nadimi, Mika Goto

Renewable Energy, Год журнала: 2024, Номер unknown, С. 121875 - 121875

Опубликована: Ноя. 1, 2024

Язык: Английский

Процитировано

3
Carbon price prediction based on multidimensional association rules and optimized multi-factor LSTM model DOI
Xinqi Tu, Lianlian Fu, Lingling Wang

и другие.

Energy, Год журнала: 2025, Номер unknown, С. 136768 - 136768

Опубликована: Май 1, 2025

Язык: Английский

Процитировано

0
Enhanced machine learning model via twin support vector regression for streamflow time series forecasting of hydropower reservoir DOI Creative Commons
Xinyue Fu, Zhong-kai Feng, Hui Cao

и другие.

Energy Reports, Год журнала: 2023, Номер 10, С. 2623 - 2639

Опубликована: Сен. 19, 2023

The non-stationary, complex, and non-linear characteristics of streamflow time series have a significant impact on the simulation results conventional hydrological forecasting models. To improve performances, this paper develops an enhanced machine learning model for forecasting, where twin support vector regression (TSVR) is combined with singular spectrum analysis (SSA) grey wolf optimizer (GWO). Specially, SSA method set as data preprocessing tool pattern identification; TSVR basic module each GWO used to select feasible parameter combination. Multi-step-ahead tasks are examine feasibility predictability proposed model. indicate that can yield superior compared traditional Thus, robust reliable provided under uncertainty.

Язык: Английский

Процитировано

7