Research On Short Term Wind Power Prediction And Uncertainty Analysis Based On Cloud Theory

Wind power forecasting is an important foundation and necessary measurement to solve the impact of wind power uncertainty.The conditions of high proportion wind power integration require more strict prediction accuracy at each moment.It has great significance to improve the accuracy of wind power forecasting results and the reliability of its uncertainty analysis.The training samples are one of the key factors that affect the prediction accuracy.The outputs of wind turbines have different characteristics under different wind scenarios which include stochastic fluctuations.Due to the diversity and vagueness of wind conditions,it is crucial to select the training samples with similarity in characteristics of fluctuation of wind conditions during the dispatch period to prediction accuracy.Therefore,based on the cloud theory,this paper studies the short-term wind power forecasting and its uncertainty analysis method for directional selection of training samples,which can learn and model the wind-induced stochastic volatility and generic fuzziness characteristics in the“specified period”.The accuracy of the wind power short-term forecasting will be improved ultimately.The specific research contents are as follows:(1)The influence of wind speed fluctuation on the output power of wind turbines is studied.Three quantitative indicators of wind speed and power fluctuation characteristics are defined respectively.A wind conditions classification method based on k-means clustering is proposed.On this basis,the dynamic time warping algorithm is used to mine the most similar wind speed fluctuation sequence in each wind condition.Based on actual operation data of wind farm with minute resolution(including wind speed and output power of wind turbines),the power generation characteristics of wind turbine and the fluctuation law of wind turbine output power under the similar wind speed fluctuation sequence are studied.(2)A short-term wind power prediction method based on wind speed cloud model directionally selecting similar training samples is proposed.Historical wind speed cloud models considering wind speed expectation,fluctuating entropy and hyper-entropy are established.The cloud model similarity measure index is established to determine the historical wind speed cloud model that is most similar to the wind speed cloud model to be predicted.As a training sample,a short-term wind power prediction model is established which can improve the prediction accuracy through targeted filtering and refinement of historical data.The results show that the proposed method can improve the accuracy of short-term wind power prediction and with practical value.(3)A short-term wind power interval prediction model based on cloud reasoning is established.Through the discretization of wind speed data of numerical weather prediction(NWP),the cloud model of wind speed and its corresponding output power are respectively established.Based on the single-rule single-condition cloud reasoning theory,a short-term wind power interval prediction model is established.The effectiveness of the proposed method is proved through the verification and analysis of the actual operation data of a wind farm in the north of China.