Multi-objective optimisation of agrivoltaic systems integrating photovoltaic energy, crop yield, and water use efficiency

This thesis develops a comprehensive and physics-based modeling and optimization framework for agrivoltaic (AV) systems, aimed at identifying optimal configurations that balance photovoltaic energy production, crop productivity, and irrigation water savings. Agrivoltaics represents a promising solution to the energy–food–water nexus by enabling simultaneous land use for agriculture and renewable energy generation. An integrated numerical framework is implemented in Python, coupling solar radiation modeling, photovoltaic performance, crop growth, and evapotranspiration processes within a unified simulation environment. Solar irradiance and shading dynamics are modeled using PVlib, while crop productivity is estimated through the Monteith Radiation Use Efficiency (RUE) approach based on absorbed photosynthetically active radiation (APAR). Irrigation requirements are quantified using the FAO-56 Penman–Monteith method, incorporating microclimatic modifications induced by photovoltaic structures. A discrete design space was evaluated by varying key system parameters, including module tilt angle, row spacing (pitch), and mounting height. System performance was assessed using multiple indicators, including energy yield, relative crop yield, water saving, and land equivalent ratio (LER). The results demonstrate that agrivoltaic performance is governed by structured trade-offs rather than a single optimal configuration. Row spacing (pitch) was identified as the dominant design parameter, controlling the primary balance between photovoltaic density and crop shading. Tilt angle acts as a secondary but important parameter, with values slightly below the photovoltaic optimum providing an effective compromise between energy production and crop productivity. Pareto analysis revealed that optimal solutions are concentrated within intermediate design ranges, where both energy and agricultural outputs are balanced. Land Equivalent Ratio (LER) values consistently exceed unity, confirming the advantage of agrivoltaic systems over separate land-use strategies. Furthermore, system typology plays a context-dependent role, with vertical bifacial systems showing advantages in environments with higher diffuse radiation. Their vertical geometry enables more uniform light distribution across the crop canopy, reducing shading intensity and supporting higher crop productivity, while bifacial irradiation capture allows for meaningful energy generation despite lower direct irradiance. Overall, this study demonstrates that agrivoltaic systems require a multi-objective, context-specific design approach. The proposed framework provides a robust basis for identifying optimal configurations and supports the development of efficient, resilient, and site-adapted agrivoltaic solutions.