Machine Learning–Based Performance Prediction and Health Assessment for Optoelectronic Devices Using Optical–Electrical Feature Fusion

Shaoyi Sun; Chunyu Ma

doi:10.23977/autml.2025.060209

Machine Learning–Based Performance Prediction and Health Assessment for Optoelectronic Devices Using Optical–Electrical Feature Fusion

Download as PDF

DOI: 10.23977/autml.2025.060209 | Downloads: 7 | Views: 143

Author(s)

Shaoyi Sun ¹, Chunyu Ma ¹

Affiliation(s)

¹ School of Electronic and Optical Engineering, Nanjing University of Science and Technology, Nanjing, China

Corresponding Author

Shaoyi Sun

ABSTRACT

Optoelectronic devices such as light-emitting diodes, photodetectors, and small photovoltaic modules are widely used in lighting, sensing, and energy conversion. However, their performance and health status are strongly influenced by manufacturing variability, operating conditions, and gradual degradation, which makes traditional threshold-based evaluation methods inaccurate and labor-intensive. In this work, we propose a machine learning–based framework for performance prediction and health assessment of optoelectronic devices using fused optical–electrical features. First, optical spectra, luminous flux, and chromaticity coordinates are combined with electrical characteristics such as I–V curves, input power, and operating temperature to construct a comprehensive feature set. Principal component analysis and normalization are then applied to reduce redundancy and stabilize the input space. On top of these features, we systematically compare several regression and classification models, including Random Forest, gradient boosting–based methods, and deep neural networks. A multi-task learning strategy is further introduced to jointly predict key performance indicators (e.g., efficiency, output power) and discrete health states (healthy, mildly degraded, severely degraded), using a hybrid loss that balances regression accuracy and classification robustness. Experimental results on a mixed dataset of simulated and measured devices show that the proposed framework achieves higher coefficient of determination and lower error metrics than conventional single-feature or single-model baselines, while providing interpretable feature importance that is consistent with physical intuition. The study demonstrates the feasibility of integrating optical–electrical measurements with machine learning for intelligent monitoring and remaining-life assessment of optoelectronic devices.

KEYWORDS

Optoelectronic devices; machine learning; optical–electrical feature fusion; performance prediction; health assessment; multi-task learning

CITE THIS PAPER

Shaoyi Sun, Chunyu Ma, Machine Learning–Based Performance Prediction and Health Assessment for Optoelectronic Devices Using Optical–Electrical Feature Fusion. Automation and Machine Learning (2025) Vol. 6: 63-73. DOI: http://dx.doi.org/10.23977/autml.2025.060209.

REFERENCES

[1] Onah E H, Lethole N L, Mukumba P. Optoelectronic devices analytics: machine-learning-driven models for predicting the performance of a dye-sensitized solar cell[J]. Electronics, 2025, 14(10): 1948. doi:10.3390/electronics14101948.
[2] Rouet-Leduc B, Barros K, Lookman T, et al. Optimisation of GaN LEDs and the reduction of efficiency droop using active machine learning[J]. Scientific Reports, 2016, 6: 24862. doi:10.1038/srep24862.
[3] Ibrahim M S, Fan J, Yung W K C, et al. Machine learning and digital twin driven diagnostics and prognostics of light-emitting diodes[J]. Laser & Photonics Reviews, 2020, 14(12): 2000254. doi:10.1002/lpor.202000254.
[4] Park I H, Lee S E, Kim Y, et al. Lifetime assessment of organic light emitting diodes by compact model incorporated with deep learning technique[J]. Organic Electronics, 2022, 101: 106404. doi:10.1016/j.orgel.2021.106404.
[5] Berghout T, Benbouzid M, Bentrcia T, et al. Machine learning-based condition monitoring for PV systems: state of the art and future prospects[J]. Energies, 2021, 14(19): 6316. doi:10.3390/en14196316.
[6] Jobayer M, Shaikat M A H, Rashid M N, et al. A systematic review on predicting PV system parameters using machine learning[J]. Heliyon, 2023, 9(6): e16815. doi:10.1016/j.heliyon.2023.e16815.
[7] Mahmood A, Wang J L. Machine learning for high performance organic solar cells: current scenario and future prospects[J]. Energy & Environmental Science, 2021, 14(1): 90–105. doi:10.1039/D0EE02838J.
[8] Meftahi N, Kachalova T, Scharber M C, et al. Machine-learning property prediction for organic photovoltaic devices[J]. npj Computational Materials, 2020, 6: 166. doi:10.1038/s41524-020-00429-w.
[9] Ahmed D R, Muhammadsharif F F. A review of machine learning in organic solar cells[J]. Processes, 2025, 13(2): 393. doi:10.3390/pr13020393.
[10] Siddiqui H, Usmani T. Interpretable AI and machine learning classification for identifying high-efficiency donor–acceptor pairs in organic solar cells[J]. ACS Omega, 2024, 9: 34445–34455. doi:10.1021/acsomega.4c02157.
[11] Lee M H, Jang J, Kwon S K, et al. Interpretable machine learning model for the highly efficient ternary organic solar cells[J]. Solar RRL, 2023, 7: 2300307. doi:10.1002/solr.202300307.
[12] dos Reis Rodrigues V, Faria F, Morgado-Dias F. Machine learning-driven prediction of organic solar cell performance: a data-centric approach to molecular design[J]. Journal of Molecular Modeling, 2025, 31: 298. doi:10.1007/s00894-025-06514-5.
[13] Yuan C C A, Fan J, Fan X. Deep machine learning of the spectral power distribution of the LED system with multiple degradation mechanisms[J]. Journal of Mechanics, 2021, 37(2): 172–183. doi:10.1093/jom/ufaa025.
[14] Blaszczak U J, Gryko L. High-quality multi-emitter LED-based retrofits for incandescent photometric A illuminant: reliability of R² evaluation[J]. Applied Sciences, 2024, 14(13): 5717. doi:10.3390/app14135717.
[15] Masita K, et al. Deep learning in defects detection of PV modules: a review[J]. Solar Energy Advances, 2025, 5: 100090. doi:10.1016/j.seja.2025.100090.
[16] Et-taleby A, Chaibi Y, Benslimane M, et al. Applications of machine learning algorithms for photovoltaic fault detection: a review[J]. Statistics, Optimization & Information Computing, 2023, 11(1): 168–177. doi:10.19139/soic-2310-5070-1537.
[17] Islam M, Mahjabeen F. Artificial intelligence in photovoltaic fault identification and diagnosis: a state-of-the-art review[J]. Energies, 2023, 16(21): 7417. doi:10.3390/en16217417.
[18] Sabati A, Bayindir R, Rahebi J. Photovoltaic panels fault detection with convolutional neural network and bitterling fish optimization (BFO) algorithm[J]. International Journal of Computational Intelligence Systems, 2025, 18(1): 239. doi:10.1007/s44196-025-00984-4.
[19] Nassreddine G, El Arid A, Nassereddine M, et al. Fault detection and classification for photovoltaic panel system using machine learning techniques[J]. Applied AI Letters, 2025, 6(2): e0115. doi:10.1002/ail2.115.
[20] Bougoffa M, Benmoussa S, Djeziri M, et al. Hybrid deep learning for fault diagnosis in photovoltaic systems[J]. Machines, 2025, 13(5): 378. doi:10.3390/machines13050378.
[21] Khandeparkar V, Shreshtha, Ramu S K. Effectiveness of supervised machine learning models for electrical fault detection in solar PV systems[J]. Scientific Reports, 2025, 15: 34919. doi:10.1038/s41598-025-18802-4.
[22] Teta A, et al. Early fault detection and diagnosis of grid-connected photovoltaic systems using a lightweight CNN optimized by energy valley optimizer[J]. Scientific Reports, 2024, 14: 69890. doi:10.1038/s41598-024-69890-7.
[23] Pedregosa F, Varoquaux G, Gramfort A, et al. Scikit-learn: machine learning in Python[J]. Journal of Machine Learning Research, 2011, 12: 2825–2830.
[24] Zhang Y, Yang Q. An overview of multi-task learning[J]. National Science Review, 2018, 5(1): 30–43. doi:10.1093/nsr/nwx105.
[25] Ji K, Lin W, Sun Y, et al. Self-supervised deep learning for tracking degradation of perovskite light-emitting diodes with multispectral imaging[J]. Nature Machine Intelligence, 2023, 5(11): 1225–1235. doi:10.1038/s42256-023-00736-z.
[26] Porowski R, Kowalik R, Szelag B, et al. Prediction of photovoltaic module characteristics by machine learning for renewable energy applications[J]. Applied Sciences, 2025, 15(16): 8868. doi:10.3390/app15168868.

Subscription

E-Mail Alert

Downloads:	4393
Visits:	208038

Machine Learning–Based Performance Prediction and Health Assessment for Optoelectronic Devices Using Optical–Electrical Feature Fusion

Author(s)

Affiliation(s)

Corresponding Author

ABSTRACT

KEYWORDS

CITE THIS PAPER

REFERENCES

RESOURCES

JOIN US

PUBLICATION SERVICES

CONTACT US