Analysis of multi-objective optimization models for sizing solar photovoltaic systems under a low-carbon grid scenario

The optimization of solar photovoltaic (PV) systems has been proposed to reduce greenhouse gas emissions in building energy use. However, despite the environmental concerns shown, most of the studies optimize technical and/or economic objectives. Recent studies show that optimizing costs rather than carbon can worsen life cycle emissions, driven by battery operation. Whether this conflict remains in storage-free PV systems in buildings remains unexplored. Therefore, this study aims to quantify the life cycle global warming potential (GWP) effects of selecting different optimization objectives in a PV capacity optimization model. To achieve this, three multi-objective optimization models of a PV system are compared, using the Non-dominated Sorting Genetic Algorithm-II (NSGA-II): (1) maximizing self-sufficiency and self-consumption, (2) maximizing the internal rate of return and net present value (NPV), and (3) minimizing global warming potential and maximizing NPV. The results show that the solutions provided by models with only technical or economic objectives do not always lead to a reduction in GWP. Furthermore, optimizing technical or economic objectives can increase GWP by up to 10.3% or reduce it by up to 5.3%, depending on the embodied impact of the PV system. These results are obtained under Swedish conditions, where the electricity supplied by the grid has a low carbon intensity. Nevertheless, this study shows that optimizing PV generation capacity for technical or economic objectives does not necessarily reduce life-cycle GWP.