Techniques, regression and applications of glass strength measurements for concentrator photovoltaic (CPV) mirrors

A primary failure mode for glass failure in reflective CPV systems is the mechanical stress caused by a thermal gradient. To establish the necessary reflector specifications, it is essential to have both economic techniques to measure glass strength and an insight into the failure mechanisms. Due to the highly stochastic nature of glass fracture, large data sets are necessary for statistical validity and to provide meaningful estimates of field failure rates. This paper discusses experimental measurement techniques for both value-added reflectors and for non-value surrogate substrates which are generated as waste during the manufacturing process. Specialized tooling enables measurement by commercially available stress-strain equipment (e. g., "Instron" testers). The glass strength is calculated from the force-to-break data, sample thickness and a substrate shape dependence. These strength data are regressed using a two-parameter Weibull model, enabling calculation of the Weibull modulus, which is a measure of the distribution of flaws of a brittle material. Using a finite element analysis model of the thermal-mechanical stress to determine the critical stress, the Weibull analysis enables extrapolation of the data to predict field failure rates. The test and regression now comprise an On-going Reliability Test (ORT) that is inherently low-cost and appropriate for high-volume manufacture. For fracture, the initiating flaws are the result of glass cutting and trimming operations. There can be low-strength outliers which result from bulk glass defects, though such flawed product should be culled during the manufacturing and inspection processes. As expected and commonly known, the glass strength is very sensitive to the cutting method and resulting quality.