Energy Yield and Electricity Management of Thin-Film and Crystalline Silicon Solar Cells

Abstract

In the case of high photovoltaic (PV) penetration into the electricity grid, the energy produced by a PV system that is effectively used (useful energy) depends on the energy yield and on how this energy is managed to avoid detrimental effects occurring at high PV injection, e.g. during the midday peak. The overall goal of this thesis is to provide guidelines for maximizing the useful energy of a PV system by quantifying losses incurred during operation at both the solar cell device and the system levels. Solar cells are usually optimized for the standard test conditions (STC). However, the conditions are generally different during operation. This work assesses how solar cell materials and designs can be optimized to maximize the energy yield for specific operating condition. We mainly focus on thin-film silicon solar cells because of their challenging metastable behavior. The temperature dependence of the performance of thin-film amorphous silicon (a-Si:H) and microcrystalline silicon solar cells is thus measured for different deposition parameters and cell designs. We observe that, by tuning the intrinsic layer thickness of a-Si:H cells, the cells with the best (STC) efficiency do not necessarily provide the highest energy output. We also explain the presence of a maximum in the value of the fill factor as a function of temperature. The temperature dependence study is then extended to thin-film silicon multi-junction, crystalline silicon heterojunction (SHJ) and other crystalline silicon solar cells. For thin-film silicon solar cells, spectral effects and degradation or recovery effects due to the metastable character of a-Si:H (due to the Staebler-Wronski effect) significantly impact the energy yield. Based on indoor and outdoor degradation/recovery experiments, we show that it is challenging to describe this metastability with a diode model. However, such a model with a current loss term and an additional temperature dependence for the saturation current and ideality factors accurately reproduces the current-voltage characteristics of a-Si:H solar cells over a wide range of irradiance levels and operating temperatures. On the system level, we model a PV system with local storage to evaluate several strategies to reduce the detrimental midday injection peaks. The impact of such measures on the useful energy is also investigated. We develop a simple control algorithm that minimizes the losses due to a feed-in limit and maximizes self-consumption without the need of a production forecast. We show that heat storage using a boiler or a heat pump performs as well as battery storage. In general, a feed-in limit reduces significantly peak injection but only a relatively small storage capacity is needed to reduce losses (due to this limit). Changes in tilt and orientation of modules also reduce losses resulting from feed-in limits and shrink the winter/summer production ratio by more than a factor of two. We also develop a statistical method that estimates - from loads measured every 15 min - when different electrical appliances in a household are commonly used. This model indicates that about 8% of the total load could be shifted easily to the midday period, thereby reducing the midday injection peak. Finally, we combine device and system aspects to show that varying cell technology (e.g. with different temperature response) has a limited but not negligible impact on system output.