Improving Ultraviolet Fluorescence Analysis for Photovoltaic Encapsulant Assessment

Fluorescence spectroscopy could be used for lamination control of the PV module, but the mapping is time-consuming. Camera-based ultraviolet fluorescence (UVF) imaging has potential for rapid crosslinking degree mapping but is unreliable due to the lack of uncertainty control. Temporal instability and spatial non-uniformity of UV excitation, camera nonlinearity and noise, keep current UVF largely qualitative and limit reliable per-location quantification of lamination quality from fluorescence intensity. To address these issues, an improved methodology is presented: multi-frame acquisition and averaging reduce camera-noise uncertainty by approximately 10 times; an automated saturation-avoidance algorithms prevents the sensor from entering nonlinear response; 1 hour pre-operation of the UV excitation reduced a maximum of 18.4% fluorescence change during measurement. Furthermore, with the per-location linear response model and uniform-field simulation, errors driven by excitation-field non-uniformity are reduced from up to 60.72% (worst-case responsivity) to near zero. These combined improvements provide a reliable foundation for mapping local encapsulant crosslinking using camera-based UVF.