Monitoring system framework design for floating wind turbine using the deep learning technology and tower response identification considering sensor optimization

Abstract

Comparing with the fixed wind turbines, the dynamic characteristics of floating wind turbines (FWTs) are more complex and resulting in distinctions in the monitoring system for structural responses. An integrated monitoring system framework concept design for FWT is proposed combing the dynamic responses correlations of different structural parts and the deep learning technology, in which the intact and partial sensor failure operation modes are considered separately. The monitoring system consists of four vital structural components, including the rotor and nacelle, tower, floating foundation, and mooring system. In order to improve the tower response identification accuracy, a new method for the sensor optimization is proposed to utilize more structural information of the structure based on the effective independence method (EIM) and modal assurance criteria (MAC) methods. A multi-perception (MLP) neural network model based on the input-output coupling mechanism is established to identify the acceleration responses at the tower top and forces at tower root. The optimized identification model can greatly improve the identification accuracy of critical sensors by more than 50%, especially for bending moments.