Computational fluid dynamics of a large-scale steam generator using data assimilation empowered turbulence modeling.

The article focuses on improving and validating turbulence model constants in Reynolds-averaged Navier–Stokes (RANS) simulations of flow behavior within large-scale steam generator (SG) systems using an ensemble Kalman filter (EnKF)-based data assimilation (DA) approach. Two reactor coolant pump (RCP) configurations—a wet stator motor pump and a canned motor pump—were tested at three flow rates to assess the generalizability of previously optimized model constants. Results demonstrate that the DA-optimized k–ε turbulence model constants significantly enhance predictions of jet penetration, turbulent separation bubble size, and velocity profiles across different flow rates and configurations, outperforming the standard model. The study highlights that these improvements arise from better turbulence dissipation modeling, leading to more accurate flow and pressure loss predictions, and suggests the EnKF-based DA method's potential for broader engineering applications in SG system design and operation.