Application of cluster analysis in short‐term wind power forecasting model
In one sentence. Clustering NWP-similar historical samples before a neural network improves short-term wind power forecasts at a northeast China farm by about 12%.
Abstract
At present, the method of predicting wind power generation is mainly based on data integration calculation. Although this method is simple, it has shortcomings in short‐term and ultra‐short‐term predictions owing to low accuracy. In this study, the clustering analysis data processing method is used to pre‐process the meteorological wind power generation data, thus improving the data quality. This method builds model samples based on historical data with similar numerical weather prediction (NWP) characteristic parameters of the original sample data and forecast date, takes the NWP information of the forecast date as the basis of similarity measurement, and extracts effective data for the neural network prediction model after the improved clustering processing. Therefore, short‐term wind power prediction analysis can be performed. Herein, the proposed data processing method is combined with the neural network model to create a software product that is applied to a wind farm in northeast China. The combined clustering data processing method of the wind power prediction model improved power prediction by ∼12% compared with that of the traditional continuous model. This demonstrates an obvious improvement in the prediction accuracy, thereby further proving the validity of the proposed method.
Labels
- Topic
- Applications & impacts
- Type
- New model or method
- Method
- Classical ML (trees, GP, linear, kernels)
- Time scale
- 6 h–3 days
- Domain
- Local / station
- Data
- NWP analysis or forecasts (IFS, GFS, ...), Station / in-situ
- Status
- Research
- Identifiers
- DOI:10.1049/joe.2018.5488 OpenAlex:W2943440041
- Added
- 2026-10-22
Labels and summary are generated by Claude and reviewed weekly. Wrong? Edit this entry.