Small wind / Buyer questions
Energy and economics
Why must a P90 wind-energy estimate specify its time horizon?
The short answer
A P90 estimate needs a time horizon because uncertainty in one year's output is different from uncertainty in a long-term average. The label alone does not show which quantity is expected to be met or exceeded with the stated probability. Comparing values without that definition can give a misleading impression of project risk.
Identify the assessed quantity
Ask whether P90 refers to annual net energy in a particular year, an average across several years or another defined period. Record the energy boundary and all included uncertainty components. A weather-only percentile and a full project assessment should remain clearly distinguished.
The same project can legitimately have different reported P90 values for different questions. The difference does not necessarily indicate a calculation error; it can reflect different time-dependent sources of variability.
Do not remove uncertainty by simple division
A longer averaging period can reduce some variability, but that does not mean every uncertainty component shrinks with the same mathematical rule. Persistent biases, shared assumptions and correlations need their own treatment.
Do not divide a single-year uncertainty by a chosen factor and call the result a verified long-term P90 without a supported method. Ask which components change with the horizon and which remain. The assumptions are especially important when the available reference record is short or changes over time.
Compare proposals on a common basis
Use the same metric, horizon and probability convention when evaluating alternatives. Preserve the central estimate, uncertainty treatment and resulting percentile together so the lower energy figure can be understood.
For a small project, a simpler scenario analysis may be more appropriate than a poorly supported probability model. Either approach should expose the decision's sensitivity. The handover must make clear what period the risk statement covers and should not imply a future production guarantee or a financing approval that has not been established separately.
Sources
- System Advisor Model documentation: wind-energy uncertainty inputs and probability results
- Bodini and Optis: correlations between operational annual-energy uncertainty components
Examples are illustrative. Confirm product-specific performance, compatibility and supply scope in the applicable documents.