Real versus simulated production
Once the installation is running, you can check whether the model is right: upload what it has produced and compare it month by month with the simulation.
Which file to upload
CSV or Excel from the inverter, the meter or the monitoring portal, with date-time and energy produced. kWh per interval and other equivalent units are accepted.
How it compares
- Only months with almost complete data count: a month with gaps is not compared.
- If there are several years of the same month, the real production is averaged.
- For each month it shows the real and simulated production, number of years and deviation (%).
- The summary gives the total deviation of the compared months and, if all 12 months are present, the real and simulated specific yield in kWh/kWp.
How to interpret the deviation
The simulation uses a typical weather year: a real year can come out above or below without the model being wrong. The more years you compare, the more significant the deviation.
| Deviation | Reading |
|---|---|
| Within the uncertainty (for example, ±6 %) in one year | Normal |
| Always in the same direction and of similar size, over several years | Probable bias of the model or the input data |
| Concentrated in specific months | Check shading, export limitation, faults or consumption |
| Very large in every month | Check installed power, orientation and meter data |
What it is for
- Detecting problems in the installation (soiling, failed strings, limited inverter).
- Calibrating your confidence in the model for similar clients.
- Giving the client an objective check of the promised production.
Fair comparison
Compare the real production with the simulation of the same project and with the real equipment and orientation data. Do not use real consumption or prices to assess the physics of the model: they are different things.
Is something missing or wrong? Write to us at hola@radiaris.com. Last reviewed: 26 October 2026.