Research And Application On Key Technologies Of Wind Turbine Control System

As a non-polluting clean energy,wind energy has received a lot of attention.In order to make full use of wind energy,the research and application on key technologies of wind turbine control system is one of the hot issues.In this paper,wind speed prediction and maximum power point tracking is dis cussed.A wind speed prediction model based on deep learning is proposed.Based on the wind speed,the maximum power tracking strategy of the hill-climbing method is improved and finally realize the full development and utilization of wind energy.Firstly,This paper proposes a wind speed prediction model(FRS-CLSTM)based on deep learning.The wind speed prediction model consists of two parts: fuzzy rough set(FRS)and CLSTM neural network model.The fuzzy rough set(FRS)reduces the dimension of huge data set.The CLSTM model takes into account the excellent characteristics of CNN and RNN,Using the CNN to extract the short-term dependencies of multi-dimensional time the and the RNN to capture the time span of input time series.The application of the model shows its effectiveness on two wind farms in Chifeng and Changqing.Secondly,the key technologies of maximum power tracking are optimized.Combining with the wind speed prediction model of FRS-CLSTM,the traditional hill-climbing search method is improved from three aspects: determining the search direction,narrowing the tracking range and avoiding the frequent fluctuation of the MPP point,which makes up the deficiencies of the traditional hill-climbing method and achieves the maximum power tracking below the rated wind speed.Finally,the simulation results show that the improved strategy can effectively avoid the wrong search direction under the condition of wind speed change,significantly reduce the wind turbine oscillation at the MPP point and gre atly improve the wind energy capture efficiency.