The long-term analysis of wind turbine performance is a complex task, because the extracted power has a multivariate dependence on ambient conditions and working parameters, which in general changes during the machine lifetime due to aging effects and-or technology optimization. For this reason, the aim of this study is developing appropriate techniques based on SCADA data analysis. A real-world test case is discussed: a decade (2011–2020) of data has been analyzed for seven 2-MW wind turbines owned by ENGIE Italia, which underwent a control optimization in 2018. Four wind turbines in the long run display even a positive drift, which is due to the control upgrade; two have a slight performance decline with age which is compatible with literature estimates; one wind turbine has a noticeable worsening in time, which is exacerbated rather than compensated by the control optimization. It is therefore argued that the long-term performance trend depends on the history of each wind turbine and the resulting scenario is complex. The control optimization provides the expected improvement on six wind turbines out of seven, but leads to a worsening on the worst wind turbine: therefore, the health status of each machine should be monitored before altering the operation.
SCADA data analysis for long-term wind turbine performance assessment: A case study
Astolfi D.;
2022-01-01
Abstract
The long-term analysis of wind turbine performance is a complex task, because the extracted power has a multivariate dependence on ambient conditions and working parameters, which in general changes during the machine lifetime due to aging effects and-or technology optimization. For this reason, the aim of this study is developing appropriate techniques based on SCADA data analysis. A real-world test case is discussed: a decade (2011–2020) of data has been analyzed for seven 2-MW wind turbines owned by ENGIE Italia, which underwent a control optimization in 2018. Four wind turbines in the long run display even a positive drift, which is due to the control upgrade; two have a slight performance decline with age which is compatible with literature estimates; one wind turbine has a noticeable worsening in time, which is exacerbated rather than compensated by the control optimization. It is therefore argued that the long-term performance trend depends on the history of each wind turbine and the resulting scenario is complex. The control optimization provides the expected improvement on six wind turbines out of seven, but leads to a worsening on the worst wind turbine: therefore, the health status of each machine should be monitored before altering the operation.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
