Comparison of Inline Crack Detection Systems for Multicrystalline Silicon Solar Cells

Greulich, Johannes; Demant, Matthias; Kunze, Philipp; Dost, Georg; Ramspeck, K.; Vetter, Andreas; Probst, Christian M.

doi:10.1109/jphotov.2020.2996750

Cited by 15 publications

(11 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This indicates that the recall rate in our rating was 64% [= 16 cells (consisting of nine HC cells and seven MC cells) / 25 cells]. Because this rate is comparable to those (67 ± 15%) reported in a previous article [ 60 ], there is no contradiction between the cracked cell sizes estimated from the bimodal distribution of the time constants and those presumed by the actual inspection. Consequently, we deduce that the remarkable decrease in the time constant of the carrier recombination should be recognized as a specific electrical signature for all cracked PV cells (including MC cells).…”

Section: Resultssupporting

confidence: 81%

Solar cell cracks within a photovoltaic module: Characterization by AC impedance spectroscopy

Tanahashi

Hsu²

2022

PLoS ONE

View full text Add to dashboard Cite

Various cell crack modes (with or without electrically inactive cell areas) can be induced in crystalline silicon photovoltaic (PV) cells within a PV module through natural thermomechanical stressors such as strong winds, heavy snow, and large hailstones. Although degradation in the performance of PV modules by cell cracks has been reported occasionally, the mode-dependent evolutions in the electrical signatures of cracks have not yet been elucidated. In this study, we propose that the reduction of the time constant in the AC impedance spectra, which is caused by the elevation of minority-carrier recombination in the p–n junction of a PV cell, is a ubiquitous signature of cracked PV cells encapsulated in a commercially available PV module. Several other characteristics derived from the illuminated current-voltage (I–V) and dark I–V data significantly evolved only in PV cells with inactive cell areas. We also propose that the evaluation by carrier recombination is a crucial diagnostic technique for detecting all crack modes, including microcracks, in wafer-based PV modules.

show abstract

Section: Resultssupporting

confidence: 81%

Solar cell cracks within a photovoltaic module: Characterization by AC impedance spectroscopy

Tanahashi

Hsu²

2022

PLoS ONE

View full text Add to dashboard Cite

show abstract

“…Possible solutions include using better augmentation strategy, adopting more complicated deep learning structures, gathering and labeling more EL images of defective modules with minor defects. Second, just as Greulich et al, 29 recently studied, the reproducibility problem of human labeling does exist that our four well trained people surly have divergence of some ambiguous defects and some of them were wrongly labeled or forgot to label. Although the four people would discuss about the ambiguities and an appointed experienced leader would make the decision, mistakes were not evitable.…”

Section: Application In Production Linementioning

confidence: 84%

Deep learning‐based automatic detection of multitype defects in photovoltaic modules and application in real production line

Yang

Zhan²,

Wang³

et al. 2021

Progress in Photovoltaics

View full text Add to dashboard Cite

Automatic defect detection in electroluminescence (EL) images of photovoltaic (PV) modules in production line remains as a challenge to replace time‐consuming and expensive human inspection and improve capacity. This paper presents a deep learning‐based automatic detection of multitype defects to fulfill inspection requirements of production line. At first, a database composed of 5983 labeled EL images of defective PV modules is built, and 19 types of identified defects are introduced. Next, a convolutional neural network is trained on top‐14 defects, and the best model is selected and tested, achieving 70.2% mAP50 (mean average precision with at least 50% localization accuracy). Then, through analyzing an object detection‐based confusion matrix, recognition bias and detection compensation in missed defects that restrain the best model's mAP50 are discovered to be harmless to normal/defective module classification in real production line. Finally, after setting specific screen criteria for different types of defects, normal/defective module classification is conducted on additionally collected 4791 EL images of PV modules on 3 days, and the best model achieves balanced scores of 95.1%, 96.0%, and 97.3%, respectively. As a result, this method surely has a highly promising potential to be adopted in real production line.

show abstract

“…Furthermore, it has been shown that even experts have difficulties finding all defects and the human defect detection rate can vary strongly. [ 10,32 ] This causes the labeling process to be error‐prone, and thus leads CNNs to partially identify false structures as defects. The disadvantage of the time‐consuming labeling process becomes even more important when the algorithms have to be adapted to other cell lines.…”

Section: Related Workmentioning

confidence: 99%

“…One drawback relates to the label process and to the labels themselves. Labels as well as heuristic filters are error‐prone, [ 10 ] time‐consuming, costly, and use only a part of the information contained in the measurement images because they focus, e.g., only on the detection of one defect. Another disadvantage pertains to the usability of the results.…”

Section: Introductionmentioning

confidence: 99%

Learning an Empirical Digital Twin from Measurement Images for a Comprehensive Quality Inspection of Solar Cells

Kunze

Rein

Hemsendorf³

et al. 2021

Solar RRL

Self Cite

View full text Add to dashboard Cite

Measurement images of solar cells contain information about their material‐ and process‐related quality beyond current–voltage characteristics. This information is currently only partially used because most algorithms look for human‐defined image features or defects. Herein, a purely data‐driven method is proposed to derive the essential image information in terms of the electrical quality within a comprehensive and meaningful representation. This representation is denoted as the empirical digital twin of the cell. Using it, solar cells can be classified according to their defects visible in the measurement images. For this purpose, a human‐in‐the‐loop approach to efficiently develop a classification scheme is presented. Therefore, a convolutional neural network combining various measurement data of a sample by correlating them with quality parameters is designed. The digital twin is an intermediate representation of the network capturing the quality‐relevant defect signatures from the images. Human experts can analyze this representation space to identify defect clusters that relate to different process errors, such as finger interruptions and shunts. How the representations are usable to derive sorting criteria for quality inspection is shown. Finally, how the empirical digital twin and the sorting scheme can be used for segmenting the defects without additional label effort is demonstrated.

show abstract

Comparison of Inline Crack Detection Systems for Multicrystalline Silicon Solar Cells

Cited by 15 publications

References 16 publications

Solar cell cracks within a photovoltaic module: Characterization by AC impedance spectroscopy

Solar cell cracks within a photovoltaic module: Characterization by AC impedance spectroscopy

Deep learning‐based automatic detection of multitype defects in photovoltaic modules and application in real production line

Learning an Empirical Digital Twin from Measurement Images for a Comprehensive Quality Inspection of Solar Cells

Contact Info

Product

Resources

About