AI-driven state of power prediction in battery systems: A PSO-optimized deep learning approach with XAI DOI
Sadiqa Jafari,

Yung-Cheol Byun

Energy, Journal Year: 2025, Volume and Issue: unknown, P. 136764 - 136764

Published: June 1, 2025

Language: Английский

Parrot optimizer with multiple search strategies for parameters estimation of proton exchange membrane fuel cells model DOI Creative Commons

Lakhdar Chaib,

Fatima Zahra Khemili,

Mohammed Tadj

et al.

Scientific Reports, Journal Year: 2025, Volume and Issue: 15(1)

Published: March 14, 2025

Proton exchange membrane fuel cells (PEMFCs) have emerged as a promising renewable energy source, generating significant interest in recent years due to their high efficiency, low operating temperature, and durability. Accurately estimating seven unknown parameters the PEMFC electrochemical model is crucial for developing more precise model, thereby improving efficiency performance of systems. For this reason, new optimization method inspired by parrots' (pyrrhura molinaes') behavior, named Parrot Optimizer (PO), introduced here address problem optimal parameter identification ( $$\:{\zeta\:}_{1},\:\:{\zeta\:}_{2},\:{\zeta\:}_{3},\:{\zeta\:}_{4},\:{\lambda\:}_{m},\:{R}_{C},\:\text{a}\text{n}\text{d}\:\beta\:$$ ) models. The estimate these characteristics treated challenging, nonlinear issue that has be addressed with strong technique. paper outlines two improvements basic PO algorithm: first involves employing Opposition-based Learning boost search refine candidate solution generation. second integrates Local Escaping Operator exploration capabilities mitigate risk getting trapped local optima, enhance overall convergence behavior. IPO was rigorously validated through application benchmark functions assess its performance. Three distinct stacks, NedStackPS6, BCS Stack, Ballard Mark V, been used empirically demonstrate efficacy improved optimizing model. Several recognized modeling approaches from literature are comprehensive examination show method's dependability. V units, corresponding SQE values 2.065816 0.012457 0.814325 V. demonstrates 12.87% improvement best measure an 88.37% reduction standard deviation compared PO. results designed approach, including sensitivity analysis, correctly characterizes effectively achieves lowest consistent trajectories.

Language: Английский

Citations

0
Efficient identification of photovoltaic cell parameters via Bayesian Neural Network-artificial ecosystem optimization algorithm DOI Creative Commons
Bo Yang,

Ruyi Zheng,

Yucun Qian

et al.

Global Energy Interconnection, Journal Year: 2025, Volume and Issue: unknown

Published: March 1, 2025

Language: Английский

Citations

0
An evolutionary multitasking algorithm for multi-objective feature selection using dual-perspective reduction DOI

Mengyue Wang,

Hongwei Ge, Xia Wang

et al.

Engineering Applications of Artificial Intelligence, Journal Year: 2025, Volume and Issue: 152, P. 110764 - 110764

Published: April 15, 2025

Language: Английский

Citations

0
Identifying the unknown parameters of PEM fuel cells based on a human-inspired optimization algorithm DOI
Badis Lekouaghet, Mohammed Haddad, M. Benghanem

et al.

International Journal of Hydrogen Energy, Journal Year: 2025, Volume and Issue: 129, P. 222 - 235

Published: April 25, 2025

Language: Английский

Citations

0
A Sequoia-Ecology-Based Metaheuristic Optimisation Algorithm for Multi-Constraint Engineering Design and UAV Path Planning DOI Creative Commons

Shijie Fan,

Ruichen Wang, Kang Su

et al.

Results in Engineering, Journal Year: 2025, Volume and Issue: unknown, P. 105130 - 105130

Published: April 1, 2025

Language: Английский

Citations

0
Exploring the role of negative emission technologies in regional power system planning toward carbon net zero -- A Case Study for the Province of Saskatchewan, Canada DOI
Xu Yang, Guohe Huang, Yanyan Liu

et al.

Energy, Journal Year: 2025, Volume and Issue: unknown, P. 136351 - 136351

Published: April 1, 2025

Language: Английский

Citations

0
Random Walk‐Based GOOSE Algorithm for Solving Engineering Structural Design Problems DOI Creative Commons

S. Mounika,

Himanshu Sharma, A. Krishna

et al.

Engineering Reports, Journal Year: 2025, Volume and Issue: 7(5)

Published: April 30, 2025

ABSTRACT The proposed Random Walk‐based Improved GOOSE (IGOOSE) search algorithm is a novel population‐based meta‐heuristic inspired by the collective movement patterns of geese and stochastic nature random walks. This includes inherent balance between exploration exploitation integrating walk behavior with local strategies. In this paper, IGOOSE has been rigorously tested across 23 benchmark functions where 13 benchmarks are varying dimensions (10, 30, 50, 100 dimensions). These provide diverse range optimization landscapes, enabling comprehensive evaluation performance under different problem complexities. various parameters such as convergence speed, magnitude solution, robustness for dimensions. Further, applied to optimize eight distinct engineering problems, showcasing its versatility effectiveness in real‐world scenarios. results these evaluations highlight competitive tool, offering promising both standard complex structural problems. Its ability effectively, combined deal positions valuable tool.

Language: Английский

Citations

0
Investigation of the critical buckling load of a column with linearly varying moment of inertia using analytical, numerical, and hybrid machine learning approaches DOI
Ayşe Polat, Aiman Tariq, Fuad Okay

et al.

The Journal of Strain Analysis for Engineering Design, Journal Year: 2025, Volume and Issue: unknown

Published: May 10, 2025

This study investigates the buckling behavior of columns with variable cross-sections using analytical, numerical, and hybrid machine learning (ML) approaches. Initially, power series method is employed to calculate loads both constant varying under diverse boundary conditions. Then a finite element (FE) analyses are performed obtain results validate by comparing them from method. Once validated, FE model used generate large dataset encompassing wide range cross-sections, lengths, material properties, as per samples obtained through Sobol sampling A ML then developed integrating XGBoost algorithm particle swarm optimization (PSO) technique for hyperparameter tuning. PSO-XGBoost trained predict cross-sections. Its performance input parameters outside training evaluated statistical metrics scatter plots. The demonstrate excellent agreement between analysis method, confirming reliability achieved remarkable predictive accuracy, R 2 values 0.999 0.996 testing datasets, respectively. Furthermore, SHapley Additive exPlanations (SHAP) conducted explore influence interactions on loads, providing valuable insights into model’s interpretability underlying mechanics column buckling.

Language: Английский

Citations

0
Multi-objective Optimization of Industrial Batch Balhimycin and Fed-Batch Lysine Biochemical Processes DOI
Swaprabha P. Patel, Ashish M. Gujarathi

Process Biochemistry, Journal Year: 2025, Volume and Issue: unknown

Published: May 1, 2025

Language: Английский

Citations

0
An Independent Discriminant Network Towards the Identification of Counterfeit Images and Videos DOI

Shayantani Kar,

B. Shresth Bhimrajka,

Aditya Kumar

et al.

IETE Journal of Research, Journal Year: 2025, Volume and Issue: unknown, P. 1 - 13

Published: May 28, 2025

Language: Английский

Citations

0