Fast Adaptive Fault Accommodation in Floating Offshore Wind Turbines via Model-Based Fault Diagnosis and Subspace Predictive Repetitive Control

TL;DR

This paper presents a rapid, adaptive fault diagnosis and control method for floating offshore wind turbines, significantly reducing blade loads caused by pitch actuator faults and enabling continuous operation.

Contribution

It introduces a novel integration of model-based fault diagnosis with subspace predictive repetitive control for fast fault accommodation in FOWTs.

Findings

Significant reduction in blade loads due to faults

Rapid fault detection and isolation achieved

Adaptive control maintains power generation during faults

Abstract

As Floating Offshore Wind Turbines (FOWTs) operate in deep waters and are subjected to stressful wind and wave induced loads, they are more prone than onshore counterparts to experience faults and failure. In particular, the pitch system may experience Pitch Actuator Stuck (PAS) type of faults, which will result in a complete loss of control authority. In this paper, a novel fast and adaptive solution is developed by integrating a model-based Fault Diagnosis (FD) scheme and the Subspace Predictive Repetitive Control (SPRC). The FD role is to quickly detect and isolate the failed pitch actuator. Based on the fault isolation results, a pre-tuned adaptive SPRC is switched online in place of the existing one, whose initial values of the parameters has been tuned offline to match the specific faulty case. After that, SPRC employs subspace identification to continuously identify a linear model of the wind turbine over a moving time window, and thereby formulate an adaptive control law to alleviate the PAS-induced loads. Results show that the developed architecture allows to achieve a considerable reduction of the PAS-induced blade loads. More importantly, the time needed to reduce the PAS-induced loads are significantly shortened, thus avoiding further damage to other components during the adaption time and allowing continued power generation.