Innovative Short-Term Wind Generation Prediction Techniques

Abstract

This paper provides an overview of research into short-term prediction techniques to assist with the operation of windpower generators. Windpower provides a new challenge to generator operators. Unlike conventional power generation sources, windpower generators supply intermittent power, have no intrinsic ability for power storage and cannot be easily ramped up to meet requirements. However, windpower is presently the fastest growing power generation sector in the world; so these problems must be solved. To be able to operate effectively, accurate short-term forecasts are essential. Knowing the future generation output from wind turbines is useful for generators, schedulers, transmission operators, network managers and energy traders. However, the difficulties of short-term wind prediction are well documented. To solve this problem, this research introduces a novel approach - the application of an adaptive neural fuzzy inference system (ANFIS) to forecasting a wind time series. A persistence model is also created to provide a benchmark of the performance. To illustrate the techniques developed, a case study is presented based on the state of Tasmania, the major island, south of mainland Australia