Research on adaptive control technology of wind power generation based on AI data mining

Abstract

This paper introduces the current achievement and control technology of wind power generation industry. By analysing the characteristics and influencing factors of wind power generation, we proposed an ultra-short-term wind power forecasting network structure which can work well in wind power forecasting. Based on the characteristics of the datasets, a variety of mainstream neural network structures are compared, our proposed method is suitable for high volatility series is further developed. The structure integrates CNN, RNN and attention mechanism, which alleviates the problem of prediction curve lag to a certain extent and effectively avoids virtually useless prediction.