Method for judging and processing wind farm power prediction input data

Abstract

The invention discloses a method for judging and processing wind farm power prediction input data. The wind farm power prediction input data is historical data of the wind farm, and contains normal data, damaged data and distorted data, wherein the damaged data and the distorted data has harmful effect on the reliability of a sample set of training data of a wind farm power prediction mode, and further influences the precision of the wind power prediction. The invention provides a preprocessing method for wind farm historical data, in the method, the normal data and the bad data in the data are distinguished, and the damaged data or the bad data in the data is increased or changed, so that the reliability of the sample set of the training data of the prediction model is enhanced, and the accuracy of wind farm power prediction is improved.

Description

Technical field [0001] This application belongs to the technical field of power systems, and specifically relates to a method for judging and processing wind farm power prediction input data Background technique [0002] When forecasting wind power, there are two types of data that directly affect the accuracy of the forecast. One is meteorological data, and the other is historical power data. The former is directly obtained from the Meteorological Bureau, and relevant processing is not performed here. We focus on the prediction of historical power data. deal with. Due to wind turbine wind measurement equipment failure, normal or abnormal shutdown of wind turbine, communication failure, channel noise, etc., historical power data often has the following three types of problems: 1) missing data; 2) distorted data; 3) noise pollution. [0003] The historical data of a wind farm includes measured data (including wind speed, wind direction, pressure, temperature, historical power) and ...

AI Technical Summary

Problems solved by technology

Due to the failure of the fan wind measuring instrument, normal or abnormal shutdown of the fan, communication failure, channel noise, etc., the following three types of problems often occur in historical power data: 1) missing data; 2) distorted data; 3) noise pollution

Therefore, the accuracy of the historical data of wind farms in the sample set is directly related to the accuracy of wind power forecasting, and the missing data and distorted data in the historical data will have unpredictable effects on the validity of the sample set and the accuracy of prediction

At present, there is no in-depth research on the preprocessing of wind farm historical data in China, and there is no processing method for this type of data

Images