A platform for research: civil engineering, architecture and urbanism
A platform for research: civil engineering, architecture and urbanism
GPU-Accelerated and Machine-Learning-Based Wind Turbine Damper Optimization
https://orcid.org/http://orcid.org/0000-0002-8555-0663
https://orcid.org/http://orcid.org/0000-0003-4975-4447
https://orcid.org/http://orcid.org/0000-0002-3591-334X
Wind turbine wake flow generally has larger turbulence intensity compared with free flow, which would result in larger damage to downstream wind turbines. However, there still has no research about wind turbine vibration considering wake effect. Meanwhile, although previous research had reduced wind turbine vibration by installing dampers, there is still no systematic global optimization method for wind turbine damper. The question that whether optimized damper without consideration about wake effect can be directly applied in wind turbines affected by wake or not is still not clear. To this end, this study first investigates wake effect on downstream wind turbine vibration. Then, tuned mass damper and rotational inerter double tuned mass damper are installed in wind turbine tower to control tower vibration. Innovatively, this study proposes a global optimization method for dampers based on radial basis function neural network and genetic algorithm, which is significantly accelerated by GPU acceleration technology. As well, wake effect on wind turbine dampers is studied by comparing optimized dampers with and without consideration of wake. Optimized dampers can reduce at most 44% tower bottom equivalent fatigue load. Numerical results can provide references for choosing damper and damper optimization in real engineering.
GPU-Accelerated and Machine-Learning-Based Wind Turbine Damper Optimization
https://orcid.org/http://orcid.org/0000-0002-8555-0663
https://orcid.org/http://orcid.org/0000-0003-4975-4447
https://orcid.org/http://orcid.org/0000-0002-3591-334X
Wind turbine wake flow generally has larger turbulence intensity compared with free flow, which would result in larger damage to downstream wind turbines. However, there still has no research about wind turbine vibration considering wake effect. Meanwhile, although previous research had reduced wind turbine vibration by installing dampers, there is still no systematic global optimization method for wind turbine damper. The question that whether optimized damper without consideration about wake effect can be directly applied in wind turbines affected by wake or not is still not clear. To this end, this study first investigates wake effect on downstream wind turbine vibration. Then, tuned mass damper and rotational inerter double tuned mass damper are installed in wind turbine tower to control tower vibration. Innovatively, this study proposes a global optimization method for dampers based on radial basis function neural network and genetic algorithm, which is significantly accelerated by GPU acceleration technology. As well, wake effect on wind turbine dampers is studied by comparing optimized dampers with and without consideration of wake. Optimized dampers can reduce at most 44% tower bottom equivalent fatigue load. Numerical results can provide references for choosing damper and damper optimization in real engineering.
GPU-Accelerated and Machine-Learning-Based Wind Turbine Damper Optimization
https://orcid.org/http://orcid.org/0000-0002-8555-0663
https://orcid.org/http://orcid.org/0000-0003-4975-4447
https://orcid.org/http://orcid.org/0000-0002-3591-334X
Wind turbine wake flow generally has larger turbulence intensity compared with free flow, which would result in larger damage to downstream wind turbines. However, there still has no research about wind turbine vibration considering wake effect. Meanwhile, although previous research had reduced wind turbine vibration by installing dampers, there is still no systematic global optimization method for wind turbine damper. The question that whether optimized damper without consideration about wake effect can be directly applied in wind turbines affected by wake or not is still not clear. To this end, this study first investigates wake effect on downstream wind turbine vibration. Then, tuned mass damper and rotational inerter double tuned mass damper are installed in wind turbine tower to control tower vibration. Innovatively, this study proposes a global optimization method for dampers based on radial basis function neural network and genetic algorithm, which is significantly accelerated by GPU acceleration technology. As well, wake effect on wind turbine dampers is studied by comparing optimized dampers with and without consideration of wake. Optimized dampers can reduce at most 44% tower bottom equivalent fatigue load. Numerical results can provide references for choosing damper and damper optimization in real engineering.
GPU-Accelerated and Machine-Learning-Based Wind Turbine Damper Optimization
Lecture Notes in Civil Engineering
Guo, Wei (editor) / Qian, Kai (editor) / Liu, Shi (author) / Wang, Yize (author) / Liu, Zhenqing (author)
International Conference on Green Building, Civil Engineering and Smart City ; 2022 ; Guilin, China
2022-09-08
8 pages
Article/Chapter (Book)
Electronic Resource
English