A Modeling Method for Micro Wind Speed Prediction of Wind Turbines Based on Time Series Analysis

Abstract

Affected by meteorological conditions, terrains, locations and specific designs, wind turbines exhibit significant uncertainties and disparities in wind energy input, which makes it difficult to predict their output power. This paper aims to enhance operational control balance in wind turbines and advance more sophisticated and intelligent control at wind farms. Utilizing the autoregressive integrated moving average (ARIMA) model, a component of time series analysis, this study analyzed time series data related to the micro wind speeds of wind turbines, and examined their correlation and randomness. The study results culminated in time series modeling to represent micro wind speeds of wind turbines, which facilitated the subsequent wind speed prediction trials. Through employing the algorithm developed for micro wind speed prediction of individual wind turbines at wind farms, the proposed approach provides supporting data for the vortex-induced vibration resistance, grid connection preparation, prevention of operational risks including load impacts, and precise control, establishing a framework for performance balance across wind turbines at wind farms, refined management including service life, and efficient operation and maintenance.

Keywords

• wind turbines
• wind speed prediction
• time series analysis
• non-stationary
• autoregressive integrated moving average(arima) model