Posters

Abstract

The big effort of transforming our entire energy system puts many challenges upon the wind energy industry. Large quantities of energy generating units must be built, operated, and maintained. Wind energy systems pose a unique combination of challenges due to their long operating time, the larger number of loosely coupled units and the modularity of the system structure. These aspects are not covered by general-purpose approaches for the setup of a digital twin, but instead require dedicated methods. We present a modular framework for the construction of a digital twin platform that is specifically tailored to the wind energy sector. It builds on the modular, ever-changing system structure of wind energy systems and translates this into a similarly evolving digital representation. The framework enables users to define service processes for their assets. By using an architecture constructed of micro-services, our approach gives a common basis to implement digital services. Simulations are often required to provide a service. We propose to use Functional Mock-Up Units (FMI Standard), which allows for the construction of modular systems and broad choice of modelling tools. The system is created by defining a System Structure Description, according to the SSP Standard, and simulated in a Co-Simulation setting. Our framework includes a Model Assembly Service to automate the construction of the overall simulation model. An ontology is incorporated that defines standardized abstractions for the interfaces for specific model types. To model the ever-changing structure of renewable energy systems, where components are exchanged, upgraded, or the environment is changed, e.g., due to the erection of new buildings or changed land use, component models and the overall system structure are valid only for a limited period. This is considered in the digital twin as well. A model management system is provided, where each single component model has a limited validity period. The entire system model is composed specifically for each point in time from the then-valid component models. Our use case for the demonstration of the platform's potential is the assessment of the remaining useful lifetime. The wind turbines' SCADA system is used to categorize the operating conditions. We use a surrogate model, which correlates the operating conditions to a damage increment in relevant components of the wind turbine. The calculation of the damage increments is based on IEC Standard 64100-1. The remaining useful lifetime is then estimated based on the integration of the damage increments. Furthermore, the digital twin platform is suited for an optimal operation planning. The provided estimation of remaining useful lifetime with additional predictions of available wind resources and estimations of electricity prices are used in a reinforcement learning algorithm to find the optimal operation strategy. With the digital twin approach, this optimal operation strategy can be directly fed back to the wind energy system. The digital twin platform provides a framework to build digital twins and adds custom services for any operator's needs. It is easily extendable and builds a basis for further applications of digital twins.