Reliability of Wind Turbine Power Modules using High-Resolution Wind Data Reconstruction: A Digital Twin Concept

Abstract

This study introduces a Digital Twin (DT) framework for the reliability assessment of wind turbine power modules. Its importance is demonstrated by examining the effect of wind turbulence on the electrothermal behaviour and lifetime of machine side power electronic converters and semiconductor devices of direct-drive wind turbines. To this end, an electrothermal model embedded in a turbine model is established, which tracks the changes in wind speed. Using real-world, 1-sec wind speed data, the real device junction temperature profiles and the fatigue experienced by the semiconductor devices are examined for two 10-min periods. Then, these metrics are compared with the corresponding metrics of the same 10-min periods when the wind speed is assumed constant and equal to the 10-min average value, which is often used in traditional device reliability assessment methods using SCADA data. Based on simulation results, the fatigue experienced by the semiconductor devices due to sudden fluctuations of the wind is found to be significantly higher than the fatigue estimated by traditional reliability assessment methods using the SCADA data. Two methods that attempt to reconstruct the wind spectrum (Random Walk Metropolis-Hastings algorithm) and compress the wind speed data (Discrete Wavelet Transform) are proposed. These and/or other similar methods may be integrated into the DT interface to address the issue of the large volume of data required to be stored in DTs.