Register now After registration you will be able to apply for this opportunity online.
Physics-based model for aerodynamic performance of wind turbine blades
Wind turbines operate in the first few hundred metres above ground level. In this area, the wind is turbulent and gusty, generating unsteady aerodynamic forces that cause premature structural damage and reduced performance. A better understanding of unsteady turbulent flows over aerofoils would help to improve the estimation of the dynamic forces.
The MISTERY project involves an interdisciplinary team of researchers in aerodynamics, machine learning and electronics, from ETHZ and OST in Switzerland, and CentraleSupélec and Centrale Nantes in France. The team investigates the impact of turbulence on aerodynamic performance of wind turbine blades. To achieve this, we study the aerodynamics of a 1:1 scale of a section of a wind turbine blade in a large wind tunnel (4m x 5m with wind speeds up to 50m/s) in Nantes (figure 1). The blade is instrumented with over 300 pressure sensors and the flow is visualised with PIV. These measurements will create a large open database that will be used to develop models for flow control of for structural health monitoring.
Keywords: aerodynamics, potential flow, pressure measurements, PIV, wind energy
The student will contribute to the development of a physics-based aerodynamic model. The MISTERY project involves several ETH labs developing different types of models, including pure machine learning models, physics-enhanced machine learning models, and graph models. Under the supervision of IFD and OST, the student will demonstrate that a physics-based model is probably the best option for such a complex problem.
The model is based on inviscid flow and uses inputs from a few sensors detecting the dynamic positions of the stagnation point (where the flow hits the blade) and the separation point (where the flow starts to detach from the blade). A panel method is then used to estimate the dynamic pressure distribution and aerodynamic force. The accuracy of the results depends heavily on the precision of the inputs. A clear definition of the stagnation and separation points is therefore necessary.
To achieve this, the student will conduct experiments on an instrumented wing in the ETH wind tunnel (figure 2). Pressure and flow visualisation will be measured simultaneously to assess the dynamics of the stagnation and separation points for different angles of attack, wind speeds and pitching motions. The inviscid model will be tested and used to analyse the interplay between these two points.
The student will contribute to the development of a physics-based aerodynamic model. The MISTERY project involves several ETH labs developing different types of models, including pure machine learning models, physics-enhanced machine learning models, and graph models. Under the supervision of IFD and OST, the student will demonstrate that a physics-based model is probably the best option for such a complex problem. The model is based on inviscid flow and uses inputs from a few sensors detecting the dynamic positions of the stagnation point (where the flow hits the blade) and the separation point (where the flow starts to detach from the blade). A panel method is then used to estimate the dynamic pressure distribution and aerodynamic force. The accuracy of the results depends heavily on the precision of the inputs. A clear definition of the stagnation and separation points is therefore necessary. To achieve this, the student will conduct experiments on an instrumented wing in the ETH wind tunnel (figure 2). Pressure and flow visualisation will be measured simultaneously to assess the dynamics of the stagnation and separation points for different angles of attack, wind speeds and pitching motions. The inviscid model will be tested and used to analyse the interplay between these two points.
The main objectives are to:
1. get familiar with the inviscid model,
2. design and build the experimental setup,
3. conduct experiments,
4. analyse and evaluate the dynamics and interplay of the stagnation and separation points,
5. improve and validate the hybrid model.
If the student is successful and has enough time, an extended Kalman filter framework developed by the Chair of Structural Mechanics and Monitoring of ETHZ will be integrated into the model to estimate the sensitivity of the inputs on the estimated aerodynamic force.
The main objectives are to: 1. get familiar with the inviscid model, 2. design and build the experimental setup, 3. conduct experiments, 4. analyse and evaluate the dynamics and interplay of the stagnation and separation points, 5. improve and validate the hybrid model. If the student is successful and has enough time, an extended Kalman filter framework developed by the Chair of Structural Mechanics and Monitoring of ETHZ will be integrated into the model to estimate the sensitivity of the inputs on the estimated aerodynamic force.
Prof. Dr. Filippo Coletti (fcoletti@ethz.ch),
Dr. Julien Deparday (julien.deparday@ost.ch)
Prof. Dr. Filippo Coletti (fcoletti@ethz.ch), Dr. Julien Deparday (julien.deparday@ost.ch)