0
By clicking the button, I accept the Terms of Use of the service and its Privacy Policy, as well as consent to the processing of personal data.
Don’t have an account? Signup
Powered by :
Researchers Develop Single-Image Method To Quantify PV Module Degradation Photograph: (Archive)
Researchers have developed a new luminescence-based method that can quantify photovoltaic (PV) module degradation using a single image, potentially enabling faster inspection of large solar installations and reducing the need for multiple measurements during field diagnostics.
The method, developed by researchers from Zhejiang University, Quantified Energy and the Australian Centre for Advanced Photovoltaics, combines machine-learning-assisted degradation classification with a physics-based reconstruction model. It converts a single electroluminescence (EL) or photoluminescence (PL) image into spatial maps of power loss and device parameters. 
The researchers said conventional quantitative luminescence diagnostics generally require multiple images captured under different electrical bias or illumination conditions. While a single luminescence image can reveal defects, multiple operating points have traditionally been needed to distinguish between different sources of performance loss and quantify their impact. 
The new approach addresses this by first classifying degradation into two broad mechanisms — recombination-driven and resistance-driven losses. A lightweight machine-learning classifier identifies the dominant mechanism, after which a physics-based inversion model uses the luminescence image to reconstruct local current-voltage behaviour and power-loss distributions. 
The researchers said the approach can account for degradation mechanisms including light-induced degradation (LID), light- and elevated-temperature-induced degradation (LeTID), ultraviolet-induced degradation (UVID) and potential-induced degradation (PID), along with transport-related defects such as wafer cracks and grid breaks.
The method was validated on 300 crystalline-silicon modules rated at 575 W and retrieved from field operation. The modules covered multiple manufacturing batches, operating environments and degradation modes. According to the study, the reconstructed power values closely matched conventional I-V measurements. The overall root-mean-square error (RMSE) across the 300 modules was 0.5%, while most of the modules showed errors within ±1.5%. 
At the individual-module level, the researchers compared their reconstruction with conventional I-V testing on two representative 575 W modules. For a recombination-driven degradation case, the method estimated maximum power at 532.7 W against 532.3 W from the electrical measurement. For a resistance-driven case, the corresponding values were 536.7 W and 541.3 W. The researchers said using a single fixed operating point could also reduce errors associated with differences between multiple images, including image misregistration and noise amplification. 
The technique was also tested at a utility-scale PV installation comprising 342 crystalline-silicon modules rated at 545 W each, giving the installation a capacity of about 186 kW. For the field demonstration, the researchers captured a night-time EL image using a drone while the PV strings were forward-biased at 0.22 times their short-circuit current. The drone operated approximately 5–10 metres above the array.  
The resulting image was used to map degradation at the intra-cell level and then aggregate the results to individual modules and strings. In the analysed section of the plant, the average module power loss was 7.53%, while a severely degraded cell region showed localised power loss of approximately 25%. 
The study also assessed three strings affected predominantly by different degradation mechanisms, including mild cracking, severe cracks and grid breaks, and LID. The analysis showed that the method could capture differences in power-loss distributions between strings as well as module-to-module variations. 
The researchers said the technology could eventually support routine PV operations and maintenance by moving beyond simple defect detection to quantifying the amount of power lost by individual modules. Such information could help plant operators prioritise maintenance, repair and module replacement based on the severity and location of degradation. 
Although the demonstration primarily used EL imaging, the researchers said the framework can also be applied to calibrated PL imaging. They further noted that, with suitable spectral selection and calibration, the approach could potentially be extended to other PV technologies, including perovskite and tandem modules. 
We are India’s leading B2B media house, reporting full-time on solar energy, wind, battery storage, solar inverters, and electric vehicle (EV)
Quick Links
© 2025 Saur Energy. All Rights Reserved.

source