Artificial Intelligence-Based Control Design for Reliable Virtual Synchronous Generators

Virtual synchronous generator (VSG) is a promising solution for inertia support of the future electricity grid to deal with frequency stability issues caused by high penetration renewable generations. However, power variation in electronic interface converters VSG emulation increases stress on semiconductor devices and hence has negative impact their reliability. Unlike existing works that only consider control design, this article proposes double-artificial neural network (ANN)-based method designing parameter considering simultaneously reliability stability. First, representative profile generated extract various injection profiles under different values through detailed simulations. Next, functional relationship between (H) lifetime consumption (LC) established proposed double-ANN model: ANN t provides fast accurate modeling thermal from given operating profile; aid , xmlns:xlink="http://www.w3.org/1999/xlink">LC built estimation LC parameters next step. The approach not guideline design certain requirement, but can also be used optimal other factors (e.g., article). technique applied grid-connected system as demonstration example.