Develop physics-based models for predicting the degradation of solar modules to support imec’s module design and yield prediction activities
Solar modules are operated in a wide variety of applications ranging from the gigawatt-scale photovoltaic (PV) power plants under hot desert climates to offshore systems floating at the sea. Operational solar modules can, therefore, be exposed to various environmental stress factors. With time, this exposure tends to trigger multiple degradation processes that evolve in parallel but can also influence each other. There are dozens of models for predicting the ensuing deterioration in the performance of solar modules. These models address either the degradation caused by a single process or the overall effect of multiple processes. The models that quantify the overall effect are typically empirical relations that link the performance degradation rate of a solar module directly to time or to easily measurable stress factors such as ambient air temperature or relative humidity. These black-box models are particularly useful for forecasting the degradation of a specific system whose performance and ambient parameters have already been measured for a few years. The models for simulating single degradation processes, in turn, often involve more physics. They aim to capture the dynamics of actual degradation mechanisms, which they relate to ambient factors by measurable parameters with a physical meaning. Though, even most of these physics-based models involve some parameters that are fitting coefficients rather than physical material properties. The incorporation of physics in the model reduces its dependence on training data and makes the produced estimates more general. This is particularly useful when optimizing module design or when assessing the degradation in a novel kind of an application.
Physical soundness has been one of the key design objectives of imec’s PV system performance modelling framework. Our goal is, therefore, to also simulate module degradation with physics-based models. The objective of this Ph.D. thesis project is to develop the required models for the most important degradation processes and implement the models in the existing framework. Each degradation model should apply the effects of stress factors only to the module parameters that the factors directly affect – not to the ultimate indicator of electrical performance such as short-circuit current or cell efficiency. Moreover, the models’ parameters should, whenever possible, consist of measurable material properties or controllable design parameters to allow the sensitivity analysis and the optimization of design factors.
During this Ph.D. project, the candidate will need to form a deep understanding of the occurrence and impact of the various degradation processes, the prior art of modelling these processes, and the underlying physics. Based on this understanding, the candidate will form the initial hypotheses, which she/he will test by laboratory and field experiments. The candidate will use the findings to develop the final models, which will be validated based on long historical datasets collected at imec or on our partners’ premises. At the end, the candidate will integrate the validated models in imec’s PV system performance modelling framework.
Required background: Physicist, Material Scientist, Electrical Engineer, or Mechanical Engineer with a background in modelling PV systems in Python.
Type of work: 20% literature review, 20% model development, 35% experimental work, 15% dissemination of results, 10% software implementation
Supervisor: Michael Daenen
Co-supervisor: Jef Poortmans
Daily advisor: Arttu Tuomiranta, Joris Lemmens