Please join us on Wednesday, November 29, 2023 for The Pierce Seminar at 4 PM in Room 1-131 with Prof. Babak Moaveni. 

Abstract Title: Digital Twinning of Offshore Wind Turbines Using Vibration Measurements

Abstract: 

This presentation highlights our latest applications of physics-based and data-driven digital twins for monitoring and performance assessment of offshore wind turbines (OWTs). Physic-based digital twinning is performed through Bayesian model updating and Hierarchical Bayesian model updating where uncertain parameters of numerical models are estimated using measured data. The Bayesian inference approach provides the probability distribution of updating parameters as well as the modeling errors. Model updating was performed on two operational OWTs, one on a jacket support structure at the Block Island Wind Farm in Rhode Island and the second on a monopile in the North Sea. 

Data-driven digital twinning is performed through sparse Gaussian Processes to infer the strain and moment response of the monopile OWT at its tower-base (as measured by strain gauges) using environmental and operational conditions from SCADA as model inputs. This modeling will enable the strain estimation of offshore wind turbines that can inform the fatigue and remaining useful life of the turbine. Furthermore, a Long Short-Term Memory network, which is a type of Recurrent Neural Network is fitted using measured training data to provide the wind speed at nacelle hub when vibration measurements on the tower are available. In another study, a transfer learning strategy is evaluated for virtual sensing of an OWT using measurements on another turbine in the same farm. The transfer learning study is implemented using the experimental data collected on two instrumented OWTs located in the Coastal Virginia Offshore Windfarm (CVOW).

Bio:

Dr. Moaveni is a Professor at the Department of Civil and Environmental Engineering at Tufts University. He obtained his Ph.D. in structural engineering at University of California-San Diego in 2007. His main research interests include probabilistic system identification of dynamic systems, signal processing, Bayesian inference, model updating, verification and validation of computational models.