Results summary
Up to SolarFarmer version 1.0.225.0, the validation was based on the local calculation only. Except for one validation site with bifacial modules, the 3D model has been used to compare SolarFarmer results with measured site data.
Now that the 3D API is available in version 1.0.229.0, all validation sites have also been simulated in the cloud, using SF-Core version 0.2.208.
The \(\mathit{rMBE}\) and \(\mathit{rMSE}\) results aggregated over the entire data period for the sites used for SolarFarmer validation are summarised in the tables below for the local and the cloud simulations.
Number of sites | 5 | ||
---|---|---|---|
Local simulations: | |||
Minimum | Maximum | Average | |
rMBE (%) | - 1.0 | + 1.2 | - 0.2 |
rRMSE (%) | 7.9 | 17.7 | 12.9 |
Cloud simulations: | |||
Minimum | Maximum | Average | |
rMBE (%) | - 0.8 | + 1.4 | 0.4 |
rRMSE (%) | 7.9 | 17.8 | 12.6 |
The differences in results when using the local and cloud simulations are mainly caused by the 3D API not yet including shading effects caused by the terrain. Additionally, the site with bifacial modules has been simulated with the 3D model in the cloud while this site is limited to the 2D model in the local simulation.
The progress of the validation by SolarFarmer version, indicating the average \(\mathit{rMBE}\) and the number of sites used, is shown in the chart below. Results for the cloud simulations are only available since version 1.0.229.0.