As one of the most sustainable alternatives regarding environmental impact, cost-effectiveness, and social integration, solar energy is expected to become an ever more ubiquitous part of our intricate human world. Dropping prices in photovoltaics that can harvest clean energy in a decentralized, safe, and modular manner are making it more viable for solar devices to be implemented in complex environments, such as urban settings. These scenes involve more constrained and dynamic conditions, encouraging the use of solar devices that can adopt arbitrary positions and personalized tracking behaviors to make the most of available resources. In modeling the incoming solar radiation for such conditions, some common simplifying assumptions may be too limiting, in particular, not considering the anisotropic nature of diffuse shadows. We develop a variety of shadow modeling approaches for all anisotropic components of the radiation; four approaches for Beam radiation and three for Diffuse components. Through thousands of simulations in urban scenes of varying complexity, these approaches are tested, characterized, and compared in terms of accuracy, precision, run-time efficiency, and practicality. Critical trade-offs are revealed between accuracy and run-time as a function of the type of approach and resolution. Our characterizations support the development and selection of modeling frameworks that are better suited to the application. This may be useful in the design, optimization, control, and forecasting of more widely adopted solar harvesting in the challenging human environment.
|