Solving the PV module quality problem - BT Imaging, Thorsten Trupke

iLS-M1 advanced inspection of PV Modules Contents BT Imaging Background – EL imaging – PL imaging – Line scan imaging Introducing the LIS-M1 – Tool features – Image examples Business model Summary Founded in 2007 Spin out from University of NSW Headquarter in Sydney, Australia Product range for ➢ Bricks ➢ Wafers ➢ Cells ➢ Modules Production systems up to 7,200 wph and offline RD tools The first company to introduce luminescence imaging into the PV sector – Large Intellectual Property portfolio Module Quality is an issue M. Kntges et al., IEA PVPS Task 13 report 2014 10 cell failures Luminescence imaging on modules - EL imaging EL imaging alone not sufficient Luminescence imaging Electroluminescence EL vs Photoluminescence PL EL imaging Simple / cheap Electric Current used for excitation – Limited to cells and modules PL imaging Technically challenging Illumination used for excitation – Bricks, wafers, cells modules Photoluminescence imaging Area scanning vs Line scanning Area scanning Line scanning Sample stationary Full area illumination Stop and Go Sample in motion Partial illumination On the fly acquisition Courtesy I. Zafirovska PL imaging vs EL imaging does it matter EL Line scan PL i. Isolated area with defect ii. Broken finger iii. Two broken fingers iv. Recombination Defect I. Zafirovska et al., IEEE J. Photov. 7, 1496-1502 2017 All defects appear dark Contrast inversion for Rs features PL imaging vs EL imaging does it matter All defects appear dark in EL images – Differentiation of defects in a single EL image difficult / impossible – Automatic defect classification difficult Inverted contrast in line scan PL images – Unambiguous differentiation of defect types – Reliable automatic defect classification Some defects e.g. cell interconnection are detectable only with EL Other defects only detectable with PL Combination of EL line scanning PL is ideal Image quality matters Poor image quality in EL imaging systems that are used in production Noisy data Out of focus Image distortion Stitching issues FF calibration issues For reliable defect analysis and module classification we need High Signal to noise High resolution No image artefacts Undistorted images BT imaging’s Solution LIS-M1 Imaging system for c-Si modules – Up to 2.2m x 1.2m samples size – Mono, multi, shingle, half cell – Adaptation to other module types CdTe, CIGS feasible – 1 Mpixel cell in c-Si modules – 1-2 Minute cycle time, depending on module type – Exceptional image quality – Advanced image analysis and classification Image quality matters EL ELPL ELPL ELPL ELPL EL EL Module classification Module classification into three classes based on Module uniformity Cell Uniformity Bypass diode test Black Cells Partially isolated area Fully isolated area Cracks mono Class 1 No defects Class 2 Some defects but acceptable Class 3 Reject 0 hours Degradation studies Courtesy I. Zafirovska 1 hour Courtesy I. Zafirovska Degradation studies 2 hours Courtesy I. Zafirovska Degradation studies 4 hours Courtesy I. Zafirovska Degradation studies 8 hours Courtesy I. Zafirovska Degradation studies 16 hours Courtesy I. Zafirovska Degradation studies 33 hours Courtesy I. Zafirovska Degradation studies