Combining model-based monitoring and a physics of failure approach for wind turbine failure detection
conference contributionposted on 19.05.2017 by Jannis Weinert, Simon Watson
Any type of content contributed to an academic conference, such as papers, presentations, lectures or proceedings.
Condition monitoring of wind turbines with only operational data has received more attention in the last decade due to the advantage of freely available data without extra equipment needed. Although the operational data recorded by the Supervisory Control And Data Acquisition (SCADA) system are intended for performance monitoring and typically stored only every 10 minutes, information on the turbine’s health can be extracted. A major focus is here on the temperature signals of mechanical parts such as drivetrain bearings. Despite the fact that absolute temperatures rise very late in the case of a failure, the temperature behaviour might change well in advance. Model-based monitoring is a tool to detect these small changes in the temperature signal affected by varying load and operation. Data-driven models are trained in a period where the turbine can be assumed to be healthy and represent the normal operation thereafter. Degradation and imminent failures can be detected by analysing the residual of modelled and measured temperatures. However, detecting failures in the residual is not always straightforward due to possibly unrepresentative training data and limited capabilities of this approach. A different way of using SCADA data lies in the estimation of damage accumulation with performance parameters based on the Physics of Failure methodology. A combination of model-based monitoring with damage calculation based on a Physics of Failure approach is proposed to strengthen the failure detection capabilities. The monitoring performance is evaluated in a case study with SCADA data from a wind farm.
This project has received funding from the European Union's Horizon 2020 research and innovation programme under the Marie Sklodowska -Curie grant agreement No 642108 (project AWESOME, http://awesome-h2020.eu/)
- Mechanical, Electrical and Manufacturing Engineering