In this article a virtual sensor for online load monitoring and subsequent remaining useful life (RUL) assessment of wind turbine gearbox bearings is presented. Utilizing a Digital Twin framework the virtual sensor combines data from readily available sensors of the condition monitoring (CMS) and supervisory control and data acquisition (SCADA) system with a physics-based gearbox model. Different state estimation methods including Kalman Filter, Least Square estimator and a quasi-static approach are employed for load estimation. For RUL assessment the accumulated fatigue damage is calculated with the Palmgren-Miner model. A case study using simulation measurements from a high-fidelity gearbox model is conducted to evaluate the proposed method. Estimated loads at the considered IMS and HSS bearings show moderate to high correlation (R = 0.50-0.96) to measurements, as lower frequency internal dynamics are not fully captured. The estimated fatigue damage differs by 5-15 % from measurements.

In this paper a wind turbine high-speed gear stage model is developed for the purpose of real-time virtual sensing of gear and bearing loads in a Digital Twin framework. The model requirements are: accurate representation of gear meshing and shaft dynamics, high computational efficiency and compatibility with other Digital Twin components, such as physical sensors signals and virtual sensing methods. State equations are derived analytically using the Bond Graph method and implemented in the software 20sim for simulation. As opposed to standard multi-body simulation (MBS) software, 20sim allows for higher flexibility in implementing interfaces to other Digital Twin components. The model fidelity is close to state-of-the-art MBS models considering 6 DOF body motion, however a simplified gear contact formulation is used, which assumes ideal kinematic meshing. Nonetheless, the Bond Graph model is able to accurately reproduce the inhomogeneous load distribution over the tooth flank, as well as the cyclic compression and decompression for each meshing period. The results suggest that the presented model is capable of monitoring fatigue loads in gear contacts and bearings in a Digital Twin framework.

In this article a novel approach for the estimation of wind turbine gearbox loads with the purpose of online fatigue damage monitoring is presented. The proposed method employs a Digital Twin framework and aims at continuous estimation of the dynamic states based on CMS vibration data and generator torque measurements from SCADA data. With knowledge of the dynamic states local loads at gearbox bearings are easily determined and fatigue models are be applied to track the accumulation of fatigue damage. A case study using simulation measurements from a high-fidelity gearbox model is conducted to evaluate the proposed method. Estimated loads at the considered IMS and HSS bearings show moderate to high correlation (R = 0.50–0.96) to measurements, as lower frequency internal dynamics are not fully captured. The estimated fatigue damage differs by 5–15 % from measurements.

No abstract

This paper presents a Digital Twin for virtual sensing of wind turbine aerodynamic hub loads, as well as monitoring the accumulated fatigue damage and remaining useful life in drivetrain bearings based on measurements of the Supervisory Control and Data Acquisition (SCADA) and the drivetrain condition monitoring system (CMS). The aerodynamic load estimation is realized with data-driven regression models, while the estimation of local bearing loads and damage is conducted with physics-based, analytical models. Field measurements of the DOE 1.5 research turbine are used for model training and validation. The results show low errors of 6.4% and 1.1% in the predicted damage at the main and the generator side high-speed bearing respectively.