Show summary
Agricultural activities occupy nearly half of the world’s habitable land, an area larger than the Americas and the European Union combined. This has largely come at the expense of natural habitats, and today, new land uses such as utility-scale photovoltaic (PV) electricity production add further pressure.
Agrivoltaic systems, in which PV modules and agricultural crops coexist on the same land, offer a pathway to greater land-use efficiency without claiming additional territory from nature. Although both PV modules and food crops rely on sunlight as their primary energy source, synergistic effects between the two have been demonstrated worldwide. PV modules can, for instance, harvest irradiance that exceeds a crop’s biological saturation point, or provide shade and shelter under harsh environmental conditions. The efficacy of these synergies, however, depends strongly on the specific crops and climates involved. In Northern Europe, short growing seasons and, in some places, strong regulatory protection of scarce agricultural land may require that crop yields are maintained or improved. In this context, vertical bifacial agrivoltaic systems have attracted particular interest due to their low physical footprint and limited crop shading. In these systems, vertically mounted PV modules arranged in rows leave the inter-row space accessible for agricultural machinery while simultaneously functioning as windbreaks that improve the local microclimate in ways that most other agrivoltaic designs do not.
This thesis develops computational frameworks for the design and evaluation of agrivoltaic systems, with particular emphasis on vertical bifacial systems in northern climates. The research is structured around methodological gaps identified in the literature. Paper I reviews the state of the art of agrivoltaic modeling, identifying weaknesses that several subsequent papers help address. Among these is the need for accurate optical modeling without the high computational cost of ray tracing tools commonly used for PV production modeling. Paper II addresses this bottleneck by identifying and benchmarking efficient matrix-based ray-tracing tools against industry-standard approaches, demonstrating that they deliver both the accuracy and speed required for autonomous optimization of agrivoltaic system design.
Paper III then utilizes one of the identified ray-tracing tools within an autonomous optimization framework. The framework identifies system designs that maximize the Land Equivalent Ratio through optimal light sharing between food crops and PV modules. Optical modeling alone, however, neglects thermodynamic and aerodynamic changes that are especially relevant where vertical systems act as windbreaks.
Paper IV therefore integrates computational fluid dynamics to analyze this windbreak effect. The results demonstrate how vertical photovoltaic rows create sheltered zones with significantly reduced wind speeds, which leads to reduced evapotranspiration and increased temperatures. The subsequent crop yield analysis shows that these beneficial mechanisms can compensate for shading, resulting in net-positive crop yields in high-latitude climates, particularly during years of drought. Finally, Paper V uses satellite remote sensing to monitor operating vertical systems across the world, offering a scalable way to independently verify agricultural productivity in commercial systems. The analysis finds no reduction in vegetation indices after installation, indicating that commercial systems maintain agricultural performance.