Please enter the words you want to search for:

[Return to folder listing]

   Model-Based Condition Monitoring of Power Semiconductor Devices in Wind Turbines   [View] 
 [Download] 
 Author(s)   Alexander BRUNKO 
 Abstract   A model-based condition monitoring has been established to assess the degradation of IGBT modules in wind turbines due to power cycling (thermo-mechanical degradation) and humidity (electrochemical degradation). For the estimation of the over-all lifetime of the IGBT modules, the model extracts the power losses from the converter currents and voltages and calculates the semiconductors' junction temperature based on a Cauer-model. A similar approach is applied for the calculation of the microclimate at the chip surface based on the outside humidity and operation conditions. To parametrise the degradation models, accelerated power cycling and temperature humidity tests were performed in the laboratory. Fed with operational and environmental data, the overall model can provide the health status of thedevices and this can be done even in real-time, if implemented into the converter controller. To test the system, data from an active wind farm was gathered by several data acquisition systems. The field data covers all seasons and various weather conditions. However, the calculated useful lifetimes were well beyond 100 years even for worst-case scenarios and, thus, the investigated wind turbines should not fail within their projected lifetime. Neither the degradation due to power cycling nor the degradation due to humidity driven effects nor special events like lightning strikes should become lifetime limiting. Either the failure mechanisms investigated are not the predominant ones or the powersemiconductor devices are not the primary failure. In fact, the same power semiconductors survive themuch harsher operating conditions in traction applications and, thus, the wind converters investigatedseem to be significantly overdesigned for their operation conditions. Nevertheless, there are indeedfailures that a suspected to be humidity related and, thus, require more attention. 
 Download 
Filename:0314-epe2019-full-19265750.pdf
Filesize:1.184 MB
 Type   Members Only 
 Date   Last modified 2020-08-14 by System