Artificial Intelligence-Based Quality Risk Prediction Framework for Energy Storage Manufacturing using ISFLA-Optimized LightGBM

Sam Yaw Wing

GBP Proceedings Series, 2026, vol. 30, 63-75

Abstract: This paper addresses the critical quality control challenges currently faced in energy storage cell manufacturing, which primarily include complex process coupling, significant detection lag, and limited capabilities for early risk identification. To overcome these substantial industrial hurdles, we propose a comprehensive, artificial intelligence-driven quality risk prediction system that seamlessly integrates the industrial Internet of Things (IoT), high-speed 5G communication, advanced edge computing, and a highly optimized Light Gradient Boosting Machine (LightGBM) model. Furthermore, a robust three-layer, two-network industrial architecture is designed to enable continuous, real-time data collection and rigorous analysis across nine core manufacturing processes. To maximize predictive performance, we introduce an Improved Shuffled Frog Leaping Algorithm (ISFLA) that significantly enhances LightGBM hyperparameter optimization. This is achieved through the implementation of adaptive weights and cosine perturbation mechanisms, effectively preventing premature convergence during model training. Utilizing a massive dataset of 1.2 million real production samples derived from 42 key process parameters, our proposed ISFLA-LightGBM model demonstrates exceptional performance. It achieves a remarkable 95.72% classification accuracy for risk levels, alongside a Mean Absolute Error (MAE) of 28.51 mAh and a Root Mean Square Error (RMSE) of 36.29 mAh for capacity prediction. These results consistently outperform traditional Support Vector Machines (SVM), Back Propagation Neural Networks (BPNN), and baseline LightGBM models by margins of 4% to 13%. Subsequent industrial deployment of this system successfully increased overall production yield by 3.2% and provided crucial 4.5-hour early warnings, thereby significantly advancing intelligent manufacturing paradigms for modern energy storage and smart factories.

Keywords: risk prediction; energy storage; optimization algorithms; artificial intelligence; smart manufacturing (search for similar items in EconPapers)
Date: 2026
References: Add references at CitEc
Citations:

Downloads: (external link)
https://soapubs.com/index.php/GBPPS/article/view/2357/2161 (application/pdf)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:axf:gbppsa:v:30:y:2026:i::p:63-75

Access Statistics for this article

More articles in GBP Proceedings Series from Scientific Open Access Publishing
Bibliographic data for series maintained by Yuchi Liu ().