This thesis evaluates moisture transport in polymers in experiments and finite element simulations, focusing on the encapsulation of photovoltaic modules. First, we measure moisture ingress with encapsulated miniature temperature and relative humidity sensors in controlled conditions in a climatic chamber. The sensors themselves thereby impact the measurements. Comparison with an in-situ gravimetric setup enables quantification of this influence, which can be included in the simulations. They are based on the Fickian diffusion model, until now the only model applied in polymers for photovoltaic applications. While this simple model can describe some materials with reasonable accuracy, it is lacking in others. Thus, we develop a novel model based on an inhomogeneous morphology in the simulations – the dual-transport model. It is far more accurate at the cost of a higher number of required parameters and longer simulation times. We devise a system of differential equations to alleviate the latter issue, describing the same process and being applicable to a homogeneous mesh in the simulations – the analytical dual-transport model.
Moisture diffusion is thereby strongly dependent on the temperature, but independent of the moisture concentration. The possible impact of another crucial factor for the degradation of polymers, ultraviolet light, is evaluated with a custom-built light source for use within a climatic chamber, based on a light emitting diode array. The experiment shows no such impact.
Moisture sorption in the field is measured with miniature temperature and relative humidity sensors encapsulated in one-cell mini-modules. Five similar setups are installed in various locations and climates around the world. These measurements enable a quantification of the difference between the microclimatic conditions of the module and the macroclimates of the regions. Finite element simulations based on the Fickian model show a good fit with the measurements, enabling an accurate extraction of the moisture profile in the modules.
|