Data Assimilation of Steam Flow Through a Control Valve Using Ensemble Kalman Filter

Abstract

Abstract The present work concentrates on the simulation enhancement of steam flow through a control valve using the data assimilation (DA) approach based on ensemble Kalman filter (EnKF). The k-ω shear stress transport (SST) model is used as the predictive model in which the model constants are optimized by DA. The selected measurement data at different operating conditions are used as observation, while the rest data are involved for validation. Before DA, four flow patterns which arise on their respective operating conditions are identified and analyzed to illustrate the basic characteristics of flow in the control valve. Then DA is performed based on the sample computation by perturbing the model constants and the EnKF process to determine the optimal model constants. These optimized constants are subsequently used for the precomputation of the valve flow with significant improvement on the flow rate prediction. The velocity and turbulent kinetic energy fields with default and DA-optimized model constants are also compared. The results show that the DA enhanced model constants can significantly reduce the predicted volume flow rate error at all opening ratios presently concerned. With the optimized model constants, the velocity and turbulent kinetic energy distributions are greatly modified in the valve seat between main valve and control valve.