Abstract This paper presents model-based fault detection, fault isolation, and fault-tolerant control schemes focused on blade pitch systems in floating wind turbines. Fault detection, isolation, and accommodation techniques are required to achieve high power capture efficiency and structural reliability in floating wind turbines. Faults in blade pitch systems should be detected at an early stage to prevent catastrophic failures. To detect faults of the blade pitch systems, a Kalman filter is designed to estimate the blade pitch angle of the system. The fault isolation algorithm is based on inference methods and capable of determining the fault type, location, magnitude and time. The fault-tolerant controller based on a reconfiguration block with a virtual sensor and shutdown mode controls the floating wind turbine to avoid unexpected external loads. The proposed methods are demonstrated in case studies with stochastic wind and wave conditions that considering different types of faults, such as biases and fixed outputs in pitch sensors and stuck pitch actuators. The simulation results show that the proposed methods can detect and isolate multiple faults effectively at an early stage. Additionally, the effectiveness of the fault-tolerant control systems for different load cases for single and multiple fault conditions is verified by numerical simulations.

中文翻译：

---

基于模型的浮动风力涡轮机叶片变桨系统的故障检测、故障隔离和容错控制

摘要 本文提出了基于模型的故障检测、故障隔离和容错控制方案，重点关注浮动风力涡轮机中的叶片变桨系统。需要故障检测、隔离和调节技术来实现浮动风力涡轮机的高功率捕获效率和结构可靠性。应该在早期检测到叶片变桨系统中的故障，以防止发生灾难性故障。为了检测叶片桨距系统的故障，设计了卡尔曼滤波器来估计系统的叶片桨距角。故障隔离算法基于推理方法，能够确定故障类型、位置、幅度和时间。基于具有虚拟传感器和停机模式的重新配置块的容错控制器控制浮动风力涡轮机以避免意外的外部负载。所提出的方法在随机风和波浪条件的案例研究中得到了证明，这些条件考虑了不同类型的故障，例如俯仰传感器和固定俯仰执行器中的偏差和固定输出。仿真结果表明，所提出的方法可以在早期有效地检测和隔离多个故障。此外，通过数值模拟验证了针对单个和多个故障条件的不同负载情况的容错控制系统的有效性。例如俯仰传感器和固定俯仰执行器中的偏置和固定输出。仿真结果表明，所提出的方法可以在早期有效地检测和隔离多个故障。此外，通过数值模拟验证了针对单个和多个故障条件的不同负载情况的容错控制系统的有效性。例如俯仰传感器和固定俯仰执行器中的偏置和固定输出。仿真结果表明，所提出的方法可以在早期有效地检测和隔离多个故障。此外，通过数值模拟验证了针对单个和多个故障条件的不同负载情况的容错控制系统的有效性。