The historical temperature data logged in the supervisory control and data acquisition (SCADA) system contains a wealth of information that can assist with the performance optimization of wind turbines (WTs). However, mining and using these long-term data is difficult and time-consuming due to their complexity, volume, etc. In this study, we tracked and analyzed the 5-year trends of major SCADA temperature rise variables in relation to the active power of four WTs in a real wind farm. To uncover useful information, an extended version of the bins method, which calculates the standard deviation (SD) as well as the average, is proposed and adopted. The implications of the analysis for engineering practice are discussed from multiple perspectives. The research results demonstrate a change in the patterns of the main temperature rise variables in a real wind farm, completeness of the monitoring of the WT internal temperature state, influence of wind turbine aging on temperature signals, a correlation between different measurement points, and a correlation between signals from different years. The knowledge gained from this research provides a reference for the development of more practical and comprehensive condition monitoring systems and methods, as well as better operation maintenance strategies.

记录在监控和数据采集 (SCADA) 系统中的历史温度数据包含丰富的信息，可帮助优化风力涡轮机 (WT) 的性能。然而，由于其复杂性、体积等原因，挖掘和使用这些长期数据既困难又耗时。 在本研究中，我们跟踪和分析了与有功功率相关的主要 SCADA 温升变量的 5 年趋势实际风电场中的四个 WT。为了发现有用的信息，提出并采用了计算标准偏差 (SD) 和平均值的 bins 方法的扩展版本。从多个角度讨论了分析对工程实践的影响。研究结果表明，真实风电场主要温升变量的变化规律，WT内部温度状态监测的完整性，风机老化对温度信号的影响，不同测点之间的相关性，不同年份的信号之间的相关性。从本研究中获得的知识为开发更实用、更全面的状态监测系统和方法以及更好的运行维护策略提供了参考。