Performance Optimization of Monofacial and Bifacial Photovoltaic Systems

This thesis explores the performance optimization of monofacial and bifacial photovoltaic systems through a combination of irradiance modeling, machine learning, and geometric analysis. A data-driven approach was developed to determine the optimal tilt angles of PV panels using six years of high-resolution irradiance data for 184 land-based locations. Optimal tilt angles were computed for three adjustment strategies: yearly, seasonal, and monthly. The irradiance on tilted surfaces was estimated using the isotropic sky model, enabling efficient simulation of front- and rear-side exposure across a wide range of albedo values. The resulting tilt angles served as ground truth for training thirteen machine learning models using location coordinates and albedo as input features. Model accuracy was validated at 15 independent U.S. cities. Top-performing models for monofacial and bifacial systems, respectively, yielded angular prediction absolute errors below 1.7° and negligible irradiation discrepancies. The thesis also investigates the influence of ground albedo and tilt angle on bifacial module performance. By comparing monofacial and bifacial systems under identical conditions, the results confirm that rear-side irradiance, and hence bifacial gain, increases substantially with enhanced ground reflectivity. Moreover, a finite-element view-factor method combined with five-minute solar geometry and dynamic shadow projection was used to assess the impact of self-shading on annual GTI. The results showed that in highly reflective ground conditions, off-optimum tilt settings can lead to total GTI losses of up to 25.8% and rear-face losses exceeding 54.3%. This detailed geometric analysis highlights the importance of including shading effects when estimating bifacial energy yield and selecting tilt angles.