Flow Control Simulation Framework

During my fellowship, I collaborated with prof. Ole Morten Aamo and his PhD student Milan Milovanovic. Our project dealt with finding strategies for vortex shedding control around cylinders in 2D space domain. My research and responsibility was to make software analysis and implementation of optimal software codes for the solved problems (by means of performance, maintainability and extendibility) to allow us to simulate and verify using Computational Fluid Dynamics (CFD) solvers our control strategy proposals.

Research report for ERCIM (European Research Consortium for Informatics and Mathematics):
https://fellowship.ercim.eu/.../087_Gayer_FP_2008_2009_SR_NTNU.pdf

Detailed description of the project

I made the reengineering, analysis and development of a software framework and application dedicated for simulation and control of incompressible fluid flows. The simulation is based on the Navier-Stokes model with possible control actuation from a customizable control module. Components of the simulation framework are: Computational Fluid Dynamics (CFD) code VISTA featuring a Navier-Stokes solver, Utility library with shared functionality for customizable Flow control modules and the configuration system, which allows defining separate simulation cases. Flow control modules are responsible for performing control calculations and actuation, reading values in flow fields, storing results for post-processing in each time step. By creating eventual new Flow control modules, we can simulate different flow control behaviour and strategies. I realized the simulation solution and it’s components in C++ with using VISTA and Diffpack API. The mesh is generated using GRIDDLER by SINTEF ICT Applied Mathematics.

I also developed codes for a case study based on this framework, where the objective is to suppress vortex shedding around cylinders in the 2D space domain by feedback control based module with kernel coefficients pre-calculated in MATLAB using Ginzburg-Landau model.

I have simulated a case study based on the software simulation environment allowing control of vortex shedding in flows around cylinders with a state feedback controller, removing vortex shedding for the Reynolds number = 60 ( see Fig. 1 , 2 and 3 ). The simulations can be potentially repeated for higher Reynolds numbers, under the condition of providing recomputed kernels for state feedback controller. This contribution resulted in timely submission and acceptance of our article “Model-based Stabilization of Vortex Shedding with CFD Verification” to a prestigious control conference, the 48th IEEE International Conference on Decision and Control.

The characteristics from the computed flow field in whole time and space domain are stored in the VTF format using GLview Express Writer API, embedded in VISTA and also some files are produced by the Utility library in plain text files, which can be imported in MATLAB. We can post-process these results by using both the MATLAB and GLview Inova. During and after finishing of the simulation, we can monitor in the MATLAB various parameters: applied control values on the slots ( Fig. 4 ), force drag and lift, velocities and pressures in the flow field ( Fig. 5 ), pressures on the surface of the cylinder ( Fig. 6 and 7 ) and many other parameters for 2D and 3D plots. I developed MATLAB scripts for processing the stored data sets and monitoring the results. My proposal and implementation of file storage system allow performance optimal visualization and monitoring of these results as well as managing and storing of many data sets for both the developers and users.

To run the simulations, our primary focus are installations on remote multiprocessor Linux servers, so that we can eventually compute general, usually complex 3D simulations flow control cases and in future with parallel processing. While this option offers advantages over local installations, namely in terms of performance, I also found this option on the other hand often impractical namely due to lack of development tools on our server, where the simulations were performed, and certain problems with queuing system and performance of transferring simulated data sets over the network. Therefore I had proposed a portable version, which is able to run on local computers with various operating systems. To overcome the portability limitation, as well as difficult installation of Diffpack and VISTA CFD code (which currently runs only on Linux), I have created a solution based on using VmWare virtual machines with Debian Etch Linux as the guest operating system as the base for our local installations.

Pictures of results

Fig. 1 Transversal velocity map at t = 0	Fig. 2 Transversal velocity map at t = 100	Fig. 3 Transversal velocity map at t = 20
Fig. 4 Monitoring of simulation control related parameters	Fig. 5 Suppression of velocities in distances up to 6 m from the cylinder in time from 0 to 50 seconds	Fig. 6 3D plot of pressures around the cylinder boundary (180 degrees)
Fig. 7 3D plot of pressures around the cylinder boundary (180 degrees) with substraction of static pressure values

Video Suppression of vortex shedding around 2D cylinder for Reynolds number 60

Zoom around the 2D cylinder