Wind Ramp Events And Wind Power Forecasting Using Meteorological Measurement Field

With the large scale integration of wind power into electric power systems, the risk and effect of wind ramp events increases. Wind ramp event refers to wide fluctuation of wind power output in short-time scale. Meteorological conditions affect wind power generation. Volatility and intermittency of wind energy is the significant reason for wind ramp event. The notable fluctuation of wind power in ramp event is harmful to power balance and frequency stability, which brings hazards to the secure and stable operation of power systems. Therefore, further study of wind ramp event and high precision of wind ramp forecasting is of great importance for secure and economic operation of power systems, promoting wind power integration and using of renewable energy.This paper centers on the forecasting of wind ramp event and aims to establish model about regional wind power forecasting during ramp period and wind ramp event. Wind ramp event is related to meteorological conditions, and regional meteorological information is important for the research of ramp event. The key to the forecasting of ramp event is to set up prediction model by relating meteorological data and ramp event’s representative features.Meteorological measurement field with met towers could capture dynamic meteorological state, such as wind speed, wind direction of different location and time period, which is lacking of the other existing forecasting method. Meteorological data constitutes meteorological field, which is decomposed into various spatial modes and principal components through empirical orthogonal function (EOF) decomposition. The principal components are used in the forecasting model of ramp event.Based on the analysis of wind speed distribution and wind power characteristics, this paper analyzes the data feature of ramp event. In order to understand characteristics of ramp event, statistical analysis of wind power fluctuation in different ramp definitions is done. In addition, through correlation analysis of wind speed sequences from meteorological field, there are correlation between different wind speed sequences, which are influenced by spatial location and terrain.In order to improve the accuracy of wind power output during ramp event period, a regional wind power forecasting method during ramp periods is proposed by using meteorological measurement field data. The wind speed matrix extracted from the meteorological measurement field is decomposed into various spatial modes and principal components through empirical orthogonal function (EOF) decomposition, which encodes regional wind speed field dynamics during ramp periods. The mapping between the decomposed principal components and regional wind power is constructed by using multiple variable nonlinear regression method. To address wind speed prediction error, the interval EOF decomposition is used, adapting the proposed forecasting method to uncertain data. The forecasting results from a region with multiple wind farms demonstrate that the proposed method achieves the significant precision improvement for regional wind power forecasting during ramp periods, and has strong robustness against wind speed prediction error.The meteorological measurement field could provide meteorological data that affects wind ramp event. A wind ramp event forecasting method based on support vector machine (SVM) classification algorithm is proposed by data mining of meteorological measurement field data using SVM algorithm. Prediction model considers the meteorological factor in meteorological field, such as wind speed, wind direction and relative humidity, and uses the particle swarm optimization (PSO) algorithm to optimize the parameters of support vector machine. In order to predict ramp event, the mapping relation between regional meteorological information and ramp event’s judgment values is set up. Validating analysis was carried out on wind farm. The results show that the proposed model and method used for forecasting achieve validity and suitability.