EN
ENWhile simplified geometrical models are sufficient for many photovoltaic systems, they reach their limits when spatial variability and geometry-driven effects play a dominant role. Applications involving complex terrain, surrounding structures, bifacial modules, or agrivoltaic layouts require an explicit three-dimensional representation to evaluate irradiance reliably.
LuSim addresses these cases through a GPU-based 3D modelling approach designed to capture geometric interactions and spatial heterogeneity without resorting to overly simplified assumptions.
LuSim is built on an explicit three-dimensional representation of photovoltaic systems and their surroundings. Terrain, support structures, modules, and nearby objects are modelled directly in space, rather than being reduced to simplified two-dimensional assumptions.
This approach is intended for cases where geometry-driven effects and spatial variability have a direct impact on irradiance distribution and system performance. Shading, mutual interactions between surfaces, and local environmental effects are therefore treated as first-order modelling elements rather than secondary corrections.
Evaluating these effects at high spatial resolution introduces significant computational complexity. In LuSim, this complexity is addressed through a GPU-based simulation framework that allows geometric interactions to be evaluated efficiently while remaining compatible with engineering workflows.
The GPU-based approach adopted in LuSim draws on computational techniques originally developed for real-time 3D graphics and video games, where complex scenes must be rendered efficiently while preserving geometric accuracy.
In computer graphics, these techniques are designed to evaluate visibility, occlusions, and surface interactions across large and detailed three-dimensional environments. The underlying computational problem is structurally similar to irradiance modelling in complex PV systems, where the exchange of radiation depends on geometry, orientation, and mutual visibility between surfaces.
By adapting these mature and highly optimised GPU-based techniques to solar energy modelling, LuSim leverages a computational paradigm that is well suited to handling complex geometries at high spatial resolution. This makes it possible to perform detailed irradiance evaluations while keeping computation times compatible with parametric studies, batch simulations, and uncertainty analyses.
At the core of LuSim lies a view factor based formulation of irradiance exchange between surfaces. Each surface element interacts with the sky, the ground, and surrounding objects according to their relative orientation, visibility, and optical properties.
This formulation is applied consistently in three dimensions and at high spatial resolution. It is used for direct, diffuse, and reflected irradiance components, allowing irradiance to be evaluated locally on photovoltaic modules, ground surfaces, crops, or other targets, rather than being limited to aggregated or lumped values.
By working with spatially resolved irradiance quantities, LuSim makes it possible to analyse heterogeneity effects that are often averaged out in simplified models, such as partial shading, non-uniform albedo, localised bifacial gains, or complex interactions between structures and terrain.
LuSim evaluates irradiance with explicit spatial and temporal resolution. Targets can be defined at different scales, from individual module regions to aggregated zones representing larger system components.
This enables the analysis of time-dependent effects such as shading evolution, tracking behaviour, and seasonal variability, as well as spatial patterns that influence system performance, mismatch losses, degradation mechanisms, and uncertainty.
Resolution choices are made to balance physical relevance and computational efficiency, depending on the purpose of the analysis.
LuSim supports complete photovoltaic energy yield assessments within a single modelling framework, and is particularly well suited to systems where geometry-driven effects and spatial variability play a dominant role.
Compared to ray tracing approaches, the GPU-based view factor methodology adopted in LuSim offers a different balance between physical detail and computational cost. This makes it particularly well suited for systematic design exploration, multiparameter analyses, and uncertainty studies in complex environments.
The choice of modelling approach should always be guided by the nature of the system being modelled and the questions being addressed.
English English