APPLICATION AND DEVELOPMENT PROSPECTS OF DIGITAL TWIN FOR FAULT DIAGNOSIS IN NUCLEAR POWER PLANTS

抄録

Accurate fault diagnosis of nuclear power plants (NPPs) is essential to the prevention of catastrophic accidents and ensures a stable operation condition. Digital twin (DT) has emerged as a data and model-based systematic approach which can generate fault state data similar to the actual system and provide a new paradigm for fault diagnosis. This paper firstly puts forward the concept of DT and its evolution progress. Then a conceptual architecture of DT based on the five components is presented due to lack of studies on the application of DT for fault diagnosis in NPPs. Given the nonlinear dynamics and uncertainty involved during the process of fault diagnosis, challenges and future development prospects to proper design and adaptability of a DT model in fault diagnosis of NPPs are also discussed.