← 返回
大型语言模型用于太阳能电池缺陷检测:零样本分类方法
Flat-Panel-Rectenna With Broad RF Energy Harvesting Coverage for Wireless-Powered Sensor Applications
| 作者 | Yan Han · Eunji Kim · Han Lim Lee |
| 期刊 | IEEE Access |
| 出版日期 | 2025年1月 |
| 技术分类 | 储能系统技术 |
| 技术标签 | 储能系统 SiC器件 |
| 相关度评分 | ★★★★★ 5.0 / 5.0 |
| 关键词 | 平板整流天线 无线电力充电 传感器应用 能量捕获 RF - DC转换效率 |
语言:
中文摘要
太阳能电池质量检测对光伏组件性能至关重要,传统缺陷检测依赖大量标注数据。本文提出基于大语言模型的零样本缺陷分类方法,无需训练数据即可识别电池片裂纹、热斑、栅线断裂等缺陷类型,通过视觉-语言模型实现智能化质检。
English Abstract
To circumvent the issue of significant space occupation associated with omnidirectional antenna-based rectennas and to seamlessly integrate into self-powered applications with minimal uncovered zones, this paper presents a new flat-panel rectenna (FPR) designed specifically for microwave-based wireless power charged sensor applications. This innovation aims to optimize space efficiency and enhance energy capture capabilities, providing a more effective solution for the continuous and reliable operation of wireless powered devices in various sectors. The proposed compact rectenna is designed with a planar antenna featuring wide beamwidth and an RF-to-DC conversion network that includes a dual-voltage rectifier architecture and a customized matching network to optimize RF-to-DC conversion efficiency. The FPR is then integrated with a batteryless heart rate sensor, enabling the sensor to operate and expand its physical orientation while maintaining accurate monitoring. The performance of the proposed RF energy harvester was evaluated at 5.8 GHz, showing a half-power coverage of 144.6° with respect to a peak antenna gain of 3.34 dBi, and a peak RF-to-DC conversion efficiency of 51.8%. Additionally, the batteryless sensor was shown to function effectively with a maximum RF energy incident angle of 140°, despite the flat-panel architecture.
S
SunView 深度解读
该零样本缺陷检测技术可直接应用于阳光电源光伏组件的生产线质检。通过大模型实现免训练的缺陷识别,降低质检系统部署成本,提升SG系列光伏逆变器配套组件的质量控制水平,减少电池片缺陷导致的系统性能损失。