Hybrid firefly algorithm–neural network for battery remaining useful life estimation DOI Creative Commons
Zuriani Mustaffa, Mohd Herwan Sulaiman

Clean Energy, Journal Year: 2024, Volume and Issue: 8(5), P. 157 - 166

Published: Aug. 27, 2024

Abstract Accurately estimating the remaining useful life (RUL) of batteries is crucial for optimizing maintenance, preventing failures, and enhancing reliability, thereby saving costs resources. This study introduces a hybrid approach RUL battery based on firefly algorithm–neural network (FA–NN) model, in which FA employed as an optimizer to fine-tune weights hidden layer biases NN. The performance FA–NN comprehensively compared against two models, namely harmony search algorithm (HSA)–NN cultural (CA)–NN, well single autoregressive integrated moving average (ARIMA). comparative analysis mean absolute error (MAE) root squared (RMSE). Findings reveal that outperforms HSA–NN, CA–NN, ARIMA both metrics, demonstrating superior predictive capabilities battery. Specifically, achieved MAE 2.5371 RMSE 2.9488 with HSA–NN 22.0583 34.5154, CA–NN 9.1189 22.4646, 494.6275 584.3098. Additionally, exhibits significantly smaller maximum errors at 34.3737 490.3125, 827.0163, 1.16e + 03, further emphasizing its robust minimizing prediction inaccuracies. offers important insights into health management, showing proposed method promising solution precise predictions.

Language: Английский

Processor in the loop implementation of artificial neural network controller for BLDC motor speed control DOI Creative Commons
Meriem Megrini, Ahmed Gaga,

Youness Mehdaoui

et al.

Results in Engineering, Journal Year: 2024, Volume and Issue: 23, P. 102422 - 102422

Published: June 15, 2024

This paper's goal is to use a proportional integral derivative controller and an artificial neural network (ANNC) control the speed of BLDC motor. The outcomes MATLAB/SIMULINK simulation are used as basis for comparing two controllers. On other hand, ANNC makes motor's reaction more consistent dependable. In words, it reduces peak time, overshoot, settle time while speeding up system responses. implemented in Processor Loop (PIL) experimenting with Arduino Mega. It based on generated C-code executed embedded card. experiment's match those simulation.

Language: Английский

Citations

8
Short-Term Load Forecasting for Residential Buildings Based on Multivariate Variational Mode Decomposition and Temporal Fusion Transformer DOI Creative Commons

H. Ye,

Qiuyu Zhu, Xuefan Zhang

et al.

Energies, Journal Year: 2024, Volume and Issue: 17(13), P. 3061 - 3061

Published: June 21, 2024

Short-term load forecasting plays a crucial role in managing the energy consumption of buildings cities. Accurate enables residents to reduce waste and facilitates timely decision-making for power companies’ management. In this paper, we propose novel hybrid model designed predict series multiple households. Our proposed method integrates multivariate variational mode decomposition (MVMD), whale optimization algorithm (WOA), temporal fusion transformer (TFT) perform one-step forecasts. MVMD is utilized decompose into intrinsic functions (IMFs), extracting characteristics at distinct scales. We use sample entropy determine appropriate number levels penalty factor MVMD. The WOA optimize hyperparameters MVMD-TFT enhance its overall performance. generate two cases originating from BCHydro. Experimental results show that our has achieved excellent performance both cases.

Language: Английский

Citations

7
Inventory optimization model using Artificial Neural Network method and Continuous Review (s,Q) DOI Creative Commons

Hanny Setyaningrum,

Iphov Kumala Sriwana,

Ilma Mufidah

et al.

SINERGI, Journal Year: 2025, Volume and Issue: 29(1), P. 143 - 143

Published: Jan. 3, 2025

The medical device industry company experienced the problem of prolonged accumulation finished goods in warehouse, causing one safety box items to be defective and damaged. Therefore, this study aims plan demand forecasting design inventory policies that consider repair caused during buildup warehouse minimize total costs using ANN Continuous Review (s,Q) methods. Demand is carried out for next 20 months, from May 2023 December 2024, model with a 17936 units inner 3370 outer items. After that, policy calculation uses continuous review method. results show decrease cost on by 83% 79%. forecasting, there was also initial 81% 80%. This research develops an optimization considers due integrating holding cost, ordering variables develop more effective efficient utilize damaged products resale. limitation it only gets months because started operating September 2021 limited data access. It hoped future researchers can strategy 10 years, focusing warehouse.

Language: Английский

Citations

0
Performance Assessment of Machine Learning Techniques in Electronic Nose Systems for Power Transformer Fault Detection DOI Creative Commons
Selene Araya, Jorge Alfredo Ardila‐Rey, Miguel A. Cabra de Luna

et al.

Energy and AI, Journal Year: 2025, Volume and Issue: unknown, P. 100497 - 100497

Published: March 1, 2025

Language: Английский

Citations

0
Forecasting of natural gas based on a novel discrete grey seasonal prediction model with a time power term DOI Creative Commons
Jun Zhang, Chaofeng Shen,

Yanping Qin

et al.

Energy Strategy Reviews, Journal Year: 2025, Volume and Issue: 58, P. 101677 - 101677

Published: March 1, 2025

Language: Английский

Citations

0
A Quest for Context-Specific Stock Price Prediction: A Comparison Between Time Series, Machine Learning and Deep Learning Models DOI Creative Commons
Mugdha Shailendra Kulkarni, S. Vijayakumar Bharathi, Arif Perdana

et al.

SN Computer Science, Journal Year: 2025, Volume and Issue: 6(4)

Published: April 2, 2025

Language: Английский

Citations

0
A novel generalized nonlinear fractional grey Bernoulli model and its application DOI Creative Commons
Jun Zhang, Chaofeng Shen,

Yanping Qin

et al.

Alexandria Engineering Journal, Journal Year: 2024, Volume and Issue: 109, P. 239 - 249

Published: Sept. 6, 2024

Language: Английский

Citations

1
Türkiye'de Cinsiyete göre Obezite Öncesi Yüzdelik Dağılımının Yapay Sinir Ağı ve Zaman Serileri ile Tahmini DOI Open Access
Halil Çolak, Emre Çolak

Karadeniz Fen Bilimleri Dergisi, Journal Year: 2024, Volume and Issue: 14(3), P. 1340 - 1359

Published: Sept. 15, 2024

Obezite, artan aşırı kilolu birey oranları nedeniyle Türkiye'de önemli bir halk sağlığı sorunu teşkil etmektedir. Ancak bu sorun, sağlıklı beslenme alışkanlıklarının teşvik edilmesi, düzenli fiziksel aktivitenin desteklenmesi ve toplumsal farkındalığın artırılması gibi önlemlerle etkili şekilde ele alınabilir. Bu hedefe ulaşmak kolektif çaba ortak vizyon gerektirecektir. Obezite için alınacak tedbirlerin etkin olabilmesi açısından, obezite öncesi dönemin bilinmesi büyük önem taşımaktadır. Makine öğrenmesinin avantajlarından tanesi de geleceği tahmin etmesidir. Yapılan çalışmada Türkiye’de cinsiyete göre yüzdelik dağılım tahminleri yapılmış 2023 ile 2030 yılları arasındaki veriler edilmiştir. Bunun Levenberg-Marquardt (LM) algoritması, Bayesian Regularization (BR) ARIMA model Holt-Winters (HW) yöntemi kullanılmıştır. Çıkan sonuçlara cinsiye dağılımın yılında kadınlarda LM’e %32,79 değerinde erkeklerde ise modelin %42,73 olacağı tahminlendi.

Citations

0
Hybrid firefly algorithm–neural network for battery remaining useful life estimation DOI Creative Commons
Zuriani Mustaffa, Mohd Herwan Sulaiman

Clean Energy, Journal Year: 2024, Volume and Issue: 8(5), P. 157 - 166

Published: Aug. 27, 2024

Abstract Accurately estimating the remaining useful life (RUL) of batteries is crucial for optimizing maintenance, preventing failures, and enhancing reliability, thereby saving costs resources. This study introduces a hybrid approach RUL battery based on firefly algorithm–neural network (FA–NN) model, in which FA employed as an optimizer to fine-tune weights hidden layer biases NN. The performance FA–NN comprehensively compared against two models, namely harmony search algorithm (HSA)–NN cultural (CA)–NN, well single autoregressive integrated moving average (ARIMA). comparative analysis mean absolute error (MAE) root squared (RMSE). Findings reveal that outperforms HSA–NN, CA–NN, ARIMA both metrics, demonstrating superior predictive capabilities battery. Specifically, achieved MAE 2.5371 RMSE 2.9488 with HSA–NN 22.0583 34.5154, CA–NN 9.1189 22.4646, 494.6275 584.3098. Additionally, exhibits significantly smaller maximum errors at 34.3737 490.3125, 827.0163, 1.16e + 03, further emphasizing its robust minimizing prediction inaccuracies. offers important insights into health management, showing proposed method promising solution precise predictions.

Language: Английский

Citations

0