Small wind / Buyer questions
Energy and economics
Can irregularly spaced wind records be averaged as if every row represented the same duration?
The short answer
Irregularly spaced records should not automatically receive equal weight. A row may represent a different duration from the next row, or it may be only a spot observation. The analyst must determine what each value represents before calculating an average or converting it into annual energy.
Establish the meaning of each row
Check whether timestamps identify interval means, instantaneous samples or event-driven records. Determine the valid start and end of each represented interval. The time until the next row is not necessarily a measured duration; it may include a communications gap.
Keep unknown periods separate instead of assuming the last recorded speed continued indefinitely. That assumption can be tested as a scenario, but it should not silently become measured evidence.
Apply appropriate duration weights
For an illustrative example, a mean of 3 m/s representing one hour and a mean of 9 m/s representing three hours have a duration-weighted mean of 7.5 m/s. Averaging the two row values gives 6 m/s, which assigns the periods the wrong relative importance.
Even the correct weighted mean does not replace an energy calculation using the relevant time series or distribution and turbine curve. Duration weighting fixes one data issue; it does not remove the nonlinear relationship between speed and generation.
Report coverage and calculation rules
Show the total represented duration, unrepresented gaps and any substitutions. When creating uniform intervals, document the resampling method and avoid duplicating observations or inventing detail that the original record cannot support.
Use a few manually checked intervals to verify the calculation before applying it to the full campaign. The final assessment should make clear whether its annualization uses valid measured time, estimated time or an explicitly modelled year. This allows the buyer to judge the result without mistaking a large row count for a long and complete measurement record.
Sources
- System Advisor Model documentation: wind-resource inputs and time-series limitations
- NREL wind-resource assessment report: measurement analysis and long-term correlation
Examples are illustrative. Confirm product-specific performance, compatibility and supply scope in the applicable documents.