基于SCADA数据挖掘的风电机组状态在线识别与预警

本项目面向大型风电机组安全可靠经济运行的现实需求，充分利用风电场已有的风电机组运行SCADA数据，围绕SCADA数据预处理、SCADA数据挖掘分析建模两个关键问题，开展风电机组状态在线识别与预警研究，主要研究内容是：（1）风电机组运行数据关联特性分析；（2）风电机组运行数据随机性分析；（3）基于物理关联特性的风电机组运行数据挖掘；（4）基于数据挖掘的风电机组运行状态预测与异常状态预警。项目特色是以能量流理论对SCADA数据时空特性进行预处理；以风速风向数据随机性分布特征进行SCADA数据随机性预处理，提出基于风电机组不同运行工况阶段的分段数据挖掘建模方法，建立物理模型与智能算法相结合的时序数据挖掘模型，实现风电机组状态预测、异常状态识别与预警。通过上述研究丰富机电设备运行状态识别与预警方法，减少风电机组运行维护成本并延长服役寿命，这对于提升我国风电产业核心竞争力具有重要的意义。

Aiming at the real demand of safe, reliable, economic operation for large wind turbines, the SCADA data in wind farm will be used adequately in this project to carry out the online identification and early warning research of wind turbines. Two key problems of this project are preprocess of SCADA data and SCADA data mining analysis model. The main research contents are: (1) correlation characteristic analysis of wind turbine operation data; (2) stochastic analysis of wind turbine operation data; (3) data mining based on physical correlation characteristics; (4) state prediction and abnormal state warning of wind turbines based on data mining. The project features and innovations are to preprocess the temporal and spatial characteristics of SCADA data based on energy flow theory, to preprocess the SCADA data randomness based on the random distribution characteristics of wind speed data and wind direction data, to propose the segmented data mining modeling method based on the different operating conditions of wind turbines, to establish the time series data mining model combing the physical model with the intelligent algorithm, and to realize the state prediction, abnormal state identification and early warning of wind turbines. Through the above research the operation state identification and early warning methods of electromechanical equipments will be enriched, the operation and maintenance costs of wind turbines will be reduced, and the service life of wind turbines will be prolonged. It is of great significance to enhance the core competitiveness of China's wind power industry.