Ultra-short Term Prediction Model of Wind Power Generation Based on Hybrid Intelligent Method

Wei Dong,Qiang Yang
DOI: https://doi.org/10.23919/chicc.2018.8482750
2018-01-01
Abstract:Accurate and reliable prediction of power generation at multiple time-scales is of paramount importance for safe and economical operation of wind farms. However, the prediction of ultra-short time period, i.e. from 15 minutes to a couple of hours, is a non-trivial task. This paper presents an efficient algorithmic solution to address the challenge of ultra-short-term power generation of wind farms based on historical data and numerical weather prediction (NWP). The suggested approach firstly adopts a collection of statistical techniques, including feature selection, attribute reduction as well as subsets partitioning, to obtain more effective data inputs for adaptive neutron-fuzzy inference system (ANFIS). The ANFIS parameters are further trained to improve the prediction performance by the use of particle swarm optimization (PSO) algorithm. The proposed solution is evaluated through a case study of realistic wind farms and the numerical results clearly demonstrate its effectiveness and prediction accuracy for ultra-short period.
What problem does this paper attempt to address?