Design of Wind Turbine Dynamic Trip-Off Risk Alarming Mechanism for Large-Scale Wind Farms

The wind power in China is being developed mainly in terms of large scale, long distance, and high clustering. In this scenario, the uncertainty of wind farm operation gradually becomes a significant factor that power grid dispatcher needs to deal with. Among which, trip-off risk of wind turbine results from system disturbance, fault and fault removal of wind farm, etc. Besides, the geographical distribution of wind turbine usually also has the impact on the trip-off risk of wind farm. In view of above factors, existing security analysis methodology is difficult to satisfy the requirement of real-time alarming due to extreme analytical difficulty and huge computing complexity. Therefore, reasonable design of trip-off risk measure index of wind turbine becomes a key issue to implement a dynamic trip-off real-time alarming system. In this paper, a wind turbine trip-off decision tree system is constructed with the discrete characteristics of wind turbine distribution in wind farms taken into account. The decision trees can perform data mining using online information and make fast prediction on wind turbine voltage out-of-limit and trip-off situations under anticipated accident sets. The trip-off risk measure indexes are then output in light of the prediction results, which can provide intuitive risk level information and decision references for operators in wind farms and power systems. The results of case studies show that the proposed method can meet the requirement of prediction accuracy and resolve the contradictions between the amount of modeled nodes, the diversity of anticipated faults and the computation time when performing over-voltage early warning and control strategies.