The second computer exercise covers hybrid approaches and application is prediction of active flow control for drag reduction of vehicles. Partially-Averaged Navier-Stokes (PANS) technique is chosen for the exercise but we will relate the different steps in the preprocessing, simulation and analysis to those in other hybrid approaches such as Detached Eddy simulation (DES) .
The flow that will be studied in this exercise is that around simplified vehicle body shown in Fig. 1.
Figure 1. Simplified vehicle body with active flow control.
We will apply periodic blowing and suction through the actuators at the rear end of the body with the perpose of delaying the shear layer instability as shown in Fig. 2.
Figure 2. Shear layers in the wake behind the body.
This flow has been extensively studied in our previous publications using both LES [1] and PANS [2]. We will use this example as it is a real case and it displays all the important steps in performing time-dependent simulation for flow around ground vehicle with flow control. The flow control method is to use a synthetic jet (i.e. alternating blowing and suction) at the rear end of the body.
The first step in our exercise will be to learn how to prescribe the actuation in our simulations. We will discuss the influence of the actuation on the computational grid and spatial and temporal resolution. Afterwards we will go through the setting for PANS simulation step by step. Besides the steps that are common with LES simulations we will discuss following steps important of a hybrid simulation and simulation of flow control:
- How to prescribe inlet boundary condition in a hybrid method such as PANS?
- How to prescribe the boundary conditions at the actuators?
Both results of the PANS and the LES will be available for the participants and will be used for analysis of the simulations. For that purpose we have prepared several MATLAB files that the participants will use and modify for the purpose of analyzing of the results.
Examples of the quantity that we will compute and analyze are:
- The ratio of the turbulent and the kinematic viscosity in the natural and the controlled flows. We will use this result to investigate what happens in PANS when the computational grid is refined between different simulations. We will discuss the meaning of the value of this ratio?
- The resolved and unresolved turbulent kinetic energy.
- Energy spectra will be computed for both PANS and LES and compared. The participants will learn what happens with the grid refinement in PANS and learn how to judge the accuracy and obtained results.
- The unresolved length and time scales will be computed and compared with smallest resolved length scales. The participants will learn how to determine the smallest length scales in PANS and from that we will be able to judge the quality of the simulation and obtained results.
Finally we will go through the results our PANS of vehicle flow and learn how to analyze the obtained results. We will compare the results of this hybrid approach with our LES results of the same flow. Thus, the participants will get from this exercise important experience in the quality of the results of a hybrid simulation of vehicle flow. This will prepare the participants for the trade-offs between the expected results and computational effort that they will later need to make in their daily work .
References
[1] Krajnovic S. and Fernandes J., 2011, “Numerical Simulation of the Flow Around a Simplified Vehicle Model with Active Flow Control,” International Journal of Heat and Fluid Flow, Vol 32, No 1, pp. 192-200.
[2] Han X. and Krajnovic S. and Basara B., 2012, Study of Active Flow Control for a Simplified Vehicle Model Using PANS Turbulence Model, Progress i Hybrid RANS-LES Modelling, Notes on Numerical Fluid Mechanics and Multidisciplinary Design Volume 117, pp 325-334.