Presentations

Abstract

Current Structural Health Monitoring approaches, which are essentially based on the deployment of structural sensors on a subset of assets in a farm, are increasingly being complemented by analytics, i.e. advanced analysis algorithms. These analytics complement the data provided by the sensors with physical and/or data-based models in order to produce advanced indicators (fatigue life consumption, vibration behavior, structural defects) on all assets of the farm and their main components. These analytics are sometimes based on physical models, in particular aero-servo-hydro-elastic models of the assets under consideration. In addition, these models are also used in Life Time Extension studies to determine the fatigue consumption of the structures. For both applications, it is essential that the physical model represents the asset as closely as possible, which in itself represents a challenge. Indeed, there are generally strong confidentiality constraints in projects, particularly with regard to the turbine. These constraints prevent access to all the necessary design information (typically the control strategy and rotor aerodynamics of the turbines). In addition, the "as-built" and "as-operated" structures are different from the "as-designed" ones due to various ageing-related phenomena. Finally, each asset of a farm has its own specificities to be taken into account, particularly for offshore cases (different foundation designs, different geotechnical properties). In this context, it is necessary to have tools allowing the generation of representative aero-servo-hydro-elastic models on the basis of incomplete design data. To do this, the missing data can be supplemented in the first instance by generic data adapted to the asset in question. Then, the obtained aero-servo-hydro-elastic model can be calibrated from the available operational information (SCADA data, time series of a vibration sensor for example) and possibly supplemented by additional specific measurements (blade scan for example). To achieve this objective, IFP Energies Nouvelles has developed calibration modules: 1. Calibration of the structural properties from vibration time series analyzed by OMA (Operational Modal Analysis) 2. Calibration of the control strategy from SCADA data (10 min SCADA for the stationary part of the control strategy, high-frequency SCADA for its dynamic part) 3. Calibration of the rotor aerodynamics according to the expected and observed performance (Cp curve) and possibly blade profile information (plans, scan). 4. Calibration of the hydrodynamic properties of the foundation to match the waves loads. These calibration modules have been demonstrated by IFP Energies Nouvelles and its commercial subsidiary GreenWITS on various industrial cases (onshore and bottom-fixed offshore) that will be presented. These cases demonstrate the technical relevance of the calibration modules as well as the gains brought in applications such as Structural Health Monitoring and Life Time Extension.