An identification method of floating wind turbine tower responses using deep learning technology in the monitoring system

Abstract

The acceleration at the tower top and the force on the tower root are the critical indicators for the structural safety of floating wind turbine (FOWT) towers. However, the corresponding monitoring sensors are prone to failure due to the harsh marine environment. Based on the coupling relationship between the tower force and the floating foundation motion responses, a method for identifying the tower top acceleration and the tower root force is proposed based on deep learning technology. Firstly, a 10 MW FOWT model is numerically simulated to obtain the corresponding responses through OpenFAST. Secondly, a multi-layer perception (MLP) model is constructed and trained. The process of determining the optimal sampling frequency is presented. Finally, the single-state and multi-state cases are studied to verify the feasibility of the proposed method. The results show that the proposed method has shown excellent performance in identifying the tower top acceleration and tower root force, which demonstrates its great promise in the field of smart health monitoring technology.