Uncertainty: P50, P90 and P99
The simulation uses a typical year, but the real year can be better or worse. The uncertainty quantifies by how much.
How it is calculated
The simulated production is taken as the P50 value (the most likely). The sources of uncertainty, expressed as a % of the annual energy, are assumed to be independent and normally distributed, and are combined in quadrature:
σ_total = √( σ_interannual² + σ_resource² + σ_model² + σ_losses² )
E_Pxx = E_P50 · (1 − z · σ_total)
| Level | z | Meaning |
|---|---|---|
| P50 | 0 | Exceeded in half of the years |
| P90 | 1,2816 | The energy exceeded with a 90 % probability |
| P99 | 2,3263 | Exceeded with a 99 % probability |
Default values
| Source | Default | What it covers |
|---|---|---|
| Year-to-year variability | 4,0 % | Variation of the irradiation from one year to the next (southern Europe) |
| Weather data | 3,0 % | Error of the satellite data or the TMY |
| Model and equipment | 3,0 % | Simulation model and equipment data |
| Losses and availability | 2,0 % | Shading, soiling, faults… |
With these values: σ_total = √(16 + 9 + 9 + 4) ≈ 6.2 %, that is, P90 ≈ 92 % of P50 and P99 ≈ 86 %.
All of them are editable in 6 · Advanced settings → Uncertainty. With data measured at the site you can reduce the uncertainty of the weather data; with a complex shading model, you should perhaps increase that of the losses.
Economics at each level
With the energy of each level the cash flows are recalculated (the savings are scaled with the energy, which is an approximation: self-consumption is not exactly linear), giving the NPV, IRR and payback of each level.
This answers «what happens if the year is bad?». If the project is still profitable at P90, it is robust. Banks and companies that finance projects usually look at the P90.
Where to see it
- In 7 · Scenarios and report → Production and profitability by probability.
- In the PDF report.
What it does not cover
The uncertainty is a simple statistical model. It does not include one-off events (a long fault, new shade from a neighbouring building) or changes in consumption.