Domain-Adaptive Multimodal Large Language Models for Photovoltaic Fault Diagnosis via Dynamic LoRA Routing
Processes · 13.02.2026 · 14 (4) · 653
DOI: 10.3390/pr14040653 ↗Kurz gefasst
Fachartikel in der Zeitschrift Processes, online erschienen am 13.02.2026. Verfasst von Junjian Wu und 5 weiteren Autor:innen. Die Arbeit wurde bisher 3-mal zitiert (Crossref) und nennt 76 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Science and Technology Project of State Grid Corporation of China.
Abstract (Originalsprache)
The reliability of photovoltaic (PV) equipment is vital for ensuring the safe and stable operation of power systems. While multimodal large language models (MLLMs) open up promising avenues for intelligent fault diagnosis, they often falter when confronted with the heterogeneity of PV data—where visual observations come from different sensor modalities (e.g., visible, infrared, and thermal) and display strong domain-dependent variations. Conventional Low-Rank Adaptation (LoRA) is not expressive enough to model such modality-aware differences, which can result in insufficient exploitation of informative patterns. To overcome this limitation, we propose PV-FaultExpert, a domain-adaptive MLLM designed specifically for PV equipment fault analysis. PV-FaultExpert is built upon DyLoRA (Dynamic Expert Routing with LoRA), a dynamic routing strategy that reformulates standard LoRA into a shared low-rank component coupled with multiple expert-specific adapters. A routing module then selects expert paths according to input characteristics, allowing the model to adapt to diverse modalities while maintaining parameter efficiency. Moreover, we construct a PVfault diagnosis dataset via ChatGPT-4o-assisted chain-of-thought reasoning and subsequent expert verification, which both supports model training and enables rigorous evaluation of our method. Extensive experiments demonstrate that PV-FaultExpert consistently surpasses strong baselines, including GPT-4 and Claude-3, across multiple evaluation criteria, producing fault analysis reports that are accurate, interpretable, and aligned with safety-critical requirements.
Förderung
- Science and Technology Project of State Grid Corporation of China
Zitieren
APA
Wu, J., Chen, Y., Min, Q., Chen, M., Zhao, J., & Ye, M. (2026). Domain-Adaptive Multimodal Large Language Models for Photovoltaic Fault Diagnosis via Dynamic LoRA Routing. Processes, 14(4), 653. https://doi.org/10.3390/pr14040653
Vancouver
Wu J, Chen Y, Min Q, Chen M, Zhao J, Ye M. Domain-Adaptive Multimodal Large Language Models for Photovoltaic Fault Diagnosis via Dynamic LoRA Routing. Processes. 2026;14(4):653. doi:10.3390/pr14040653
BibTeX
@article{Wu2026,
author = {Wu, Junjian and Chen, Yiwei and Min, Qihao and Chen, Ming and Zhao, Jie and Ye, Mang},
title = {Domain-Adaptive Multimodal Large Language Models for Photovoltaic Fault Diagnosis via Dynamic LoRA Routing},
journal = {Processes},
year = {2026},
volume = {14},
number = {4},
pages = {653},
doi = {10.3390/pr14040653}
}
EUVIMED