Key Technologies Research On Electroluminescence Test System Of Crystalline Solar Cells

The crystalline silicon solar cells produce various defects during the production and preparation process which greatly affect the photoelectric conversion efficiency of the solar cells.Therefore it is necessary to detect and filter the defective cells.Some defects can be identified by manual inspection and AOI systems while there are some internal defects that cannot be detected by such methods.At present,it becomes popular to detect internal defects of solar cells in electroluminescence images.However,the existing EL testers can only complete offline detection so that it is difficult to meet the production requirements.The main task of this paper is to study the crystalline silicon solar cell EL test system that can be applied to online production.According to the characteristics of the crystalline solar cells printing line,the overall design of the EL testing system is carried out.And the core hardwares of platform alignment system and image acquisition system are selected and analyzed.Based on the modular idea,the software the test system of this subject is designed.On the base of this,the visual aligning system of parallel alignment is studied,including parallel platform kinematics modeling,visual system calibration and cells positioning algorithm.After studying the error of the test system and the characteristics of the parallel platform alignment,interpolation correction algorithm based on attitude calibration data is proposed to meet the rapid requirements of production.Then,defect detection algorithm solar cells in electroluminescence images is studied which is the another key problem of this paper.After analyzing the characteristics of defects in electroluminescence images,pre-processing studies of the black spots,broken fingers and crack defects are studied.The defect signal is enhanced by combining methods in spatial and frequency domains,and the defects are finally extracted based on features such as area,gray scale and inclination of minimum circumscribed rectangle.Finally,the experimental research is carried out by using the established hardware platform and software system.The experimental results show that the correction algorithm proposed in this paper can meet the requirements of the alignment accuracy of the test system.In addition,the detection time and detection accuracy of the system can meet the actualproduction requirements.