A numerical study was conducted to predict the performance curve of a downscaled model of the main coolant pump for a sodium-cooled fast reactor and to reduce the head loss by the optimization of the diffuser blade.
The ANSYS CFX program was utilized to obtain flow characteristics inside the pump as well as the overall pressure rise across the pump operating on- and off-design points. Computational domain was divided into several blocks to achieve high grid quality effectively and 7.5 million nodes were used totally to resolve small leakage flows as well as the flow inside the rotating impeller.
The corresponding experiment was conducted to validate CFD computed results. The comparison between the CFD and experimental data shows excellent agreement in terms of mass flow rate and head rise on and near design operating points.
The DOE (design of experiments) and RSM (response surface method) were utilized to reduce the head loss by the diffuser blade in the pump. The diffuser blade was defined as four geometric parameters for DOE. The analysis of 25 cases was made to solve the output parameters for all design points which are defined by the DOE. RSM was fitting the output parameter as a function of the input parameters using regression analysis techniques. The optimized model increased the total pump head on the design point and the low mass flow rate point, but total pump head on 130% of operating mass flow rate was reduced than the initial model.