Uncertainty analysis method and system for wind power plant prediction

[0065] figure 1 It is a flow chart of an uncertainty analysis method for wind farm prediction in the present invention, such as figure 1 shown, including:

[0066] S1. Based on the pre-acquired wind farm data processing, the turbulence intensity set is obtained;

[0067] S2. Dividing the turbulence intensity set according to a preset division threshold to obtain a turbulence intensity level;

[0068] S3. Combining the turbulence intensity levels and the conditional quantile regression model constructed based on each turbulence intensity level, to obtain uncertainty information of wind farm prediction under each turbulence intensity level.

[0069] The algorithm basis of the present invention is quantile regression estimation, that is, adopting the idea of ​​regression model between the conditional quantiles of the wind power prediction variable X and the wind power measured variable Y, adding the micro-regional turbulence intensity variable Z formed by the measured wind speed,

[0077] The input wind farm data involved in the present invention include: wind farm anemometer tower data, predicted power and measured power, wherein in this embodiment, the first sampling period is set to 5min, and the second sampling period is set to 15min, i.e. time resolution The time is 15 minutes. Among them, the wind farm anemometer tower data and the measured power are all from reliable sampling points of the wind farm. The specific technical steps used include:

[0078] S1. Based on the pre-acquired wind farm data processing, the turbulence intensity set is obtained;

[0079] (1) Basic data preparation.

[0080] According to the start and end time of the basic data set for wind power prediction uncertainty analysis, real-time meteorological observation data of wind farms are extracted, that is, information such as instantaneous wind speed, wind direction, and sampling period given by the anemometer at the hub height of the wind farm fan.

[0081] The data set is obtai