Dynamic Analysis in Design and Operation of Large Floating Offshore Wind Turbine Drivetrains
MetadataShow full item record
- Institutt for marin teknikk 
This thesis presents a versatile rotor-bearing simulation model, including elements required to model rolling-element bearings as main bearings in large floating offshore wind turbines. Furthermore, the thesis presents new knowledge on the dynamic behavior of a main bearing, based on field measurements in one of the world’s first floating offshore wind farms. Larger wind turbines are a fundamental contribution to reducing cost of energy for installations on both floating substructures and bottom-fixed foundations. Increasing turbine size and rating requires lightweight and compact main bearing and drive train solutions and challenges established layouts, components selection and analysis methods used in the design process. Floating wind turbines are necessary to access the full offshore wind power potential. Although pioneering floating wind farms exist and more are under planning, floating operation is currently not a design driver for the standardized rotor-nacelle assembly part of a turbine. It is simply a different set of conditions for which the calculated ultimate loads and the accumulated fatigue damage must not exceed the type certification basis. Considering the solution spaces for drivetrains and floating substructures in combination with the limited experiences with larger turbines and floating operation, more knowledge on drivetrain load effects is needed. Reliability problems in wind turbine gearboxes have partly been attributed to limited understanding of drivetrain dynamics. Numerous studies addressing design and analysis of gear-based wind turbine drivetrains have followed. Despite increasing interest in floating offshore wind turbines, only a few studies have assessed the effects of floating operation on the drivetrain. These studies generally point to the main bearings as the most critical components. Main bearings have received limited attention in literature, regardless of the design challenges for large wind turbines. The main bearings are an integral part of the turbine’s load-carrying structure, and main bearing replacement commonly requires the removal of both the rotor and the nacelle. Consequently, the design life for the main bearings should match or exceed the design life of the turbine. This aspect is important with respect to life extension studies, as main bearings should not be regarded as readily replaceable. Development of cost-efficient solutions for heavy maintenance is critical for floating turbines, and the value of avoiding bearing replacement is potentially higher than for bottom-fixed turbines. Modeling and simulation are essential for offshore wind power development. Global structural analyses and local drivetrain analyses are commonly performed separately, in what is termed the decoupled approach. The main bearings are effectively connecting these domains, and attention to the modeling requirements for main bearings is needed. The model presented in this thesis is developed through consistent use of the bond graph method, providing a clear model structure. Body-fixed equations of motion are derived by using Lagrange’s method and contained in field elements. An alternative cage model and a flexible outer ring are introduced, providing a more complete modeling basis. Modal representation is used for the flexible bodies, including a solution for moving loads on the outer ring. An example simulation of a generic rotor-bearing system during run-up with and without a bearing damage is included to demonstrate the capability and usefulness of the model for transient analyses. The simulation results illustrate that the behavior of the non-linear system during a transient is not intuitive or easily inferred from the characteristics of the individual parts. The thesis also presents a first-tier experimental analysis based on field measurement data from a main bearing in a 6 MW turbine on a spar-type floating substructure. Circumferential strain in the stationary bearing ring has been measured in multiple positions by using Optical Fiber Bragg Grating sensor arrays. Analyses of the data in the time domain and the frequency domain show that in-plane bending deflection occurs in the main bearing ring, largely driven by differential blade bending moments at 3P frequency caused by wind loads and with limited influence from floater motion. This is an important result for future design of floating offshore wind turbines. Furthermore, sum and difference frequencies in the measured response as the result of non-linear system behavior indicate that decoupled drivetrain analysis approaches, where global and local analysis are carried out separately, may be insufficient in some cases. The results from the modeling and simulation contribute to advancing knowledge on rotordynamic system models suitable for studying rolling-element bearing dynamics in mechatronics- and control-related applications, during transient events and fault conditions, and subject to damage. The results from the experimental analyses confirm the basis for the modeling requirements with respect to flexible ring representation and thus indirectly with respect to the importance of a cage. Furthermore, it can be inferred from the experimental analysis that the significance and implications of the results depend on the actual main bearing design and the turbine size. For some designs, it is recommended to consider performing coupled analysis and including the hub consistently as a flexible element in both drivetrain modeling and full-scale testing of the drivetrain, thus moving the load interface from the main shaft to the blade attachment.