The influence of rotor-stator frame change models and applying sweep method to blade profiles on axial fan performance

In this study, the influence of rotor-stator frame change (mixing plane) models and the influence of sweep on rotor blade profiles of axial flow fan are examined by CFD analysis. There are three different frame change methods defined for the passages between rotor and stator stages in CFD analysis of turbomachines which are frozen rotor, mixing plane (stage) and transient rotor-stator. In the first part of the thesis, the effect of these three methods on the performance of the reference axial fan model is analyzed. Frozen rotor and mixing plane (stage) methods are used in steady analysis and time transformation method from transient blade row model is used in unsteady analysis. First of all, analyzes are solved by using frozen rotor and mixing plane models from the steady rotor-stator frame change models at the specified revolutions per minute. Then, using the mixing plane analyzes results as initial conditions, the reference axial fan model is analyzed with time transformation method from the transient blade row models. Results are obtained over six operating points to get more accurate performance curves. After the end of analyzes, results are compared with different performance parameters. The performance of reference axial flow fan model are compared according to the results three different frame change models. Consequently, the pressure and efficiency curves obtained by mixing plane and frozen rotor models give similar results. However, it has been seen that the time transformation method produces more static pressure rise and higher efficiency compared to these two models, especially at low mass flow rates. Furthermore, the results of CFD analysis by using these three methods are examined in different blade sections for different parameters, such as pressure distributions on the rotor blades and also streamlines, pressure contours and velocity vectors which show flow characteristics on the rotor and stator blades. In the second section of the thesis, the sweep method which shifts blade profiles in the direction perpendicular to the chord is applied on reference axial fan rotor blades. In open literature, there are many studies about the sweep method for turbomachines but as a result of the studies, there are some contradictory explanations when evaluating the effects of the method on the turbomachine performance. In this study, forward and backward sweep with 30° sweep angle is applied to the rotor blades and two different rotor blade models are obtained as a result of this process. First, these forward and backward sweep fan models are analyzed by using mixing plane (stage) model. Thus, the aerodynamic performance curves and flow fields of these fans are calculated. When the swept models are compared to the results of reference fan model analysis, static pressure rise and efficiency curves are observed different from each other. Forward sweep increases the static pressure rise and the efficiency of the fan. However, forward sweep expands the fan's operating range, but shifts to lower mass flow rates. On the contrary, backward sweep decreases the static pressure rise and efficiency of the fan. Nevertheless, operating range of the fan is shifted to the higher mass flow rates, besides, this operating range gets narrowed . In addition to these results, pressure distributions on the rotor blades with the dimensionless pressure coefficient in the different blade sections are examined. Consequently, forward sweep model reduces the aerodynamic load sligthly with respect to the reference model and the backward sweep model increases the aerodynamic blade load compared to other models. Finally, in addition to the mixing plane (stage) analyzes, sweep models are also analyzed by means of the frozen rotor and time transformation methods for the highest efficiency points. Computational Fluid Dynamics (CFD) analyzes are performed using ANSYS CFX 17.2 analysis program. First, existing reference axial fan CAD geometry is imported into ANSYS DesignModeler, then blade geometries of the reference fan model are prepared by using ANSYS DesignModeler module. Then, the CFD meshes of rotor and stator blades are generated using ANSYS TurboGrid. Three different sequences of meshes (rough, medium and frequent) are generated to assess the mesh dependence of the simulations. As a result of this process, mesh dependency and y+ values of boundary layers are checked and the analyses are carried out with the medium frequency mesh structures. The number of mesh elements for stator and rotor are around 0.55 million and 1.15 million respectively. Then, model setup and numerical simulations are carried out with ANSYS CFX. Model setup consists of defined fluid domains, flow conditions, periodic interfaces, frame change models, turbulance model and solution parameters. Fluid domain involves four different parts; inlet, stator, rotor and outlet. Of these domains, inlet, stator and outlet parts are stationary whereas rotor composes the rotational part. For numerical simulations, to avoid convergence problems , inlet and outlet volumes of the flow domains are added to reference axial fan rotor and stator domains. Furthermore, in order to avoid the time problem depending on the number of mesh, rotational periodicity is used for analyzes. However, transient CFD calculations of multistage turbomachines, which are already in the design process, are limited due to computational power and the computational domain which is needed to enforce spatial and temporal periodicity because of unequal blade counts and pitches. Thus, one main issue is to maintain the periodicity by reducing the number of blade passages in each blade row. Therefore, the fluid domains are modeled as 3 rotor blades and output volume with 5 stator blades and input volume to provide the pitch interval needed to implement the time transformation approach. SST k-w turbulence model is used in all simulations. The k-w based Shear Stress Transport (SST) model was designed to give highly accurate predictions of the onset and the amount of the flow separation under adverse pressure gradients by the inclusion of transport effects into the formulation of the eddy viscosity. This results in a major improvement in terms of the flow separation predictions. Morever, a target is set for all normalised RMS residuals to be below 10-5. While it is important to drive the equation residuals to negligible iteration error, it is also important to monitor overall quantities such as mass conservation, pressure ratio and efficiency as a function of iteration. Finally, calculations are generated via the CFX-Solver module and the analysis results are displayed with CFD-Post module. As a result of this study, rotor-stator frame change (mixing plane) methods and the sweep on rotor blade profiles of axial flow fan are examined and consequently axial fan performance and flow characteristics are investigated with several performance parameters. Furthermore, valuable results are obtained at the end of the study for both subjects which would contribute to the open literature.