PV modules on the flow

Finite element analysis (FEA) as part of computer-aided engineering is used in many branches, especially within R&D departments. State-of-the-art simulation software enables the application of simplified real-life conditions to photovoltaic components, modules and even arrays of modules, even without requiring a physical prototype. This leads to faster development cycles and a more accurate estimation of lifetime and testing under (almost) real operating conditions.
In order to simulate realistic wind loads on PV modules, and to investigate the influence of air flow on a module’s mechanics, temperature computational fluid dynamics (CFD) are used for simulation analyses at the Fraunhofer Center for Silicon Photovoltaics (CSP) in Halle (Saale), Germany. Present investigations focus on module positioning (standalone or in an array) and orientation (inclination and wind direction) and, for example, how these parameters can affect the module’s mechanical probability of failure and the temperature distribution within PV modules due to forced convection.
Current testing standards for PV modules regarding their mechanical strength under wind pressure – IEC 61215 respectively IEC 61646 – describe homogeneously distributed testing loads on the module’s surface. However, experiments in wind tunnels have shown a lack of homogeneity in pressure distribution on the typically inclined geometry of PV modules.
For R&D departments in the photovoltaic industry, the costs of purchasing and maintaining such wind tunnels bear no sensible economic relation to the benefits of testing prototype modules under real wind conditions. Computer-aided engineering and simulation, more precisely computational fluid dynamics, offer an interesting alternative, as demonstrated in structural mechanics. Modern simulation software enables the modeling of complete PV modules and arrays under boundary conditions close to reality.
Furthermore, coupling with different simulation physics allows, for example, statements about mechanical stresses in the module’s components or about temperature distribution within modules influenced by irradiance, environmental temperature and ambient air flow. And in getting all of this information, it is then not necessary to even build a prototype of a PV module.

Significant lack of homogeneity

First investigations at Fraunhofer CSP using CFD showed a significant lack of homogeneity of the pressure distribution as expected from wind tunnel experiments. Aiming not only at the fluid mechanical effects but also its direct influence on the structural mechanics of the solar module, a complete module – containing glass top, encapsulant, wafer as simplified silicon cells and back sheet (or alternatively a second glass plate for double-glass modules) – was modeled as a solid domain with ambient air as the fluid domain. With the resulting static flow pattern around the module recorded as average over time, the resulting pressure distribution on the module’s front and back surface was a subsequent structural-mechanical analysis that gave information on mechanical deformation and the stresses of all the module components and the mounting structure. Depending on the type of mounting, itsinclination, and the wind direction, the resulting mechanical probability of failure could be derived from stress distribution of glass and silicon cells. These data, compared with simulations with homogeneously distributed loading according to IEC61215, respectively IEC 61646, show significant differences in results, which leads to the question of the quality of these testing standards. In addition, this supports computer-aided engineering with simulation in terms of development due to its advantage of customization and adoption of almost every boundary condition.
In Figure 1a, the distribution of 1st principal stresses of solar cells is shown resulting from the IEC testing load of homogeneously distributed 2,400 Pa surface load. Due to flow pattern, influenced by the inclination and orientation of modules in the field, air flow around the module (see Figure 1b) leads to a non homogeneous pressure distribution on the module’s surface. Subsequently, Figure 1c shows the 1st principal stresses of cells resulting from such an inhomogeneously distributed wind load at 130km/h wind speed and 30° inclination. Results show that IEC testing induces stresses in a similar range with a certain safety margin, but additional wind squall and temperatures, apart from room temperature, are not considered within this range.

Temperature distribution

In addition to structural-mechanical analysis, CFD can be coupled with thermal analysis. Due to the power loss of solar cells, they are assumed as heat sources and, with given cell efficiency and irradiance temperature, distribution in all components of a PV module can be simulated. Environment temperature, irradiance and wind speed are the main influences on module temperature. Wind speed and its effect on forced convection can be investigated with respect to module positioning. With different positioning – as either a standalone module for off-grid solutions or as modules in an array – the flow patterns differ and, thus, varying convection leads to different temperature distributions (see Figure 2).
The application of CFD simulation in photovoltaic research and development gives detailed information on real loading situation and temperature distribution, as has been exemplarily shown. This information is especially useful for more regional module and material design as well as the analysis of building-integrated PV (BIPV). Verification experiments can be planned based on simulation results, limiting the necessary effort to a minimum.