Home > Research > Projects > Opensource machine learning metamodels for asse...
View graph of relations

Opensource machine learning metamodels for assessing wind turbine energy losses due to blade erosion and support turbine predictive maintenance

Project: Research

Description

The project has developed machine learning metamodels for the rapid prediction of wind turbine power degradation and energy loss due to leading edge erosion. The metamodels are based on high-fidelidy computational fluid dynamics and consider a fairly wide range of erosion damage. The models refer to the NACA64618 airfoil used on the outermost part of the NREL 5 MW turbine and can be used to support wind turbine blade predictive maintenance. The metamodels have been made opensource and they are freely available on GitHub at https://github.com/LANCASTER-CFD/Leading-Edge-Erosion. These metamodels are meant to support blade predictive maintenance with respect to leading edge maintenance costs.

Key findings

leading edge erosion, machine learning, predictive maintenance
StatusFinished
Effective start/end date1/05/2330/06/24

Facilities/Equipment

  • High End Computing cluster (HEC)

    Facility/Equipment: Facility

Research outputs