The CFD analysis of the store separation subjected to the cavity flow at transonic speed

In this thesis, numerical simulations of the internal store separation used in new generation fighter aircraft by solving the Computational Fluid Dynamics (CFD) and six-degree of freedom equations together. First of all, in order to verify the solution of the equations of motion, an example of external store separation called EGLIN, experimental study has been repeated numerically. Later, the experimental cavity flow known as M219 in the literature was verified by numerical method. After the verification studies, simulations of store separation from the clean cavity and from the cavity which passive control methods were applied were carried out. Firstly the effect of the store holder used in the experimental setup on the store trajectory was examined in the verification of the store separation from the outside named EGLIN. For this purpose, numerical simulations of store separation from under the wing with and without store holder with a inviscid flow approach for different mesh densities was performed and the results were compared. It was revealed that the store holder used in the experimental setup does not affect the store trajectory in the store separation, and the store holder was not used in subsequent studies. In order to examine the effects of mesh density, inviscid numerical solutions with 600 thousand, 1 million, 1.7 million, 3.8 million and 5.4 million mesh densities were performed and compared. . It was found that 5.4 million mesh density is the limit value in terms of mesh density, the result becomes independent of the mesh density after this value. In order to examine the viscous effects, the numerical simulation of the store separation was repeated with the SST k-ω turbulence model for 2.8 million and 5.9 million mesh densities. In this part of the study, although it has been demonstrated that the inviscid flow approach can be used in external store separation, the turbulent flow model has been shown to be more compatible with the experimental results. Finally, the effects of the solver algorithm were examined in the EGLIN validation study. In this context, two numerical simulations which are density-based and pressure-based were performed with SST k-ω turbulence model for 5.8 million mesh density. Density-based solution algorithm does not allow high-order discretization due to its mesh sensitivity. Based on high-order discretization can be performed in the pressure-based solver for the same mesh structure, the pressure-based solution algorithm more compatible with the experimental results. As a result, the consistence of the results of the numerical solution with the results of the EGLIN experimental study in terms of both translation and rotation values validated the solution of the six-degree of freedom equations of motion of the numerical solution. The experimental study called M219 in the literature at 0.85 Mach free-stream flow with a length-to-depth (L / D) ratio of 5 was validated by CFD for the cavity flow verification study. In the cavity flow verification study, two clean cavity studies which are without a door and with a door with 90 degrees door angle were performed. Detached Eddy Simulation (DES) turbulence model was used in both numerical studies. Flow parameters such as mean longitudinal velocity, mean transverse velocity, turbulent kinetic energy and Reynolds stress obtained from the numerical solution were compared with the reference Large Eddy Simulation (LES) values and the Overall Sound Pressure Level (OASPL) values obtained from the cavity floor were compared with the experimental values. Both the flow parameters and the OASPL obtained from the numerical solution have been shown to be in agreement with the experimental and reference LES study. In the study of internal store separation, the cavity defined in the M219 experimental study was used as cavity and the generic store used in the EGLIN experimental study was used as store. DES turbulence model has been used in all numerical simulations, both with and without motion. The generic store, which was placed in the clean cavity with a door with 90 degree door angle, was first released with the same launch forces in two different launch times when the gravity force was maximum and minimum. Although it is seen that the store trajectory at 0.85 Mach speed from the clean cavity with an L / D ratio of 5 is quantitatively dependent on the launch time, basically the same trajectory was obtained in two different launch times. It was observed that the same flap on the back of the generic store hit the cavity door at the same point while the generic store which is tightly placed in the clean cavity with a door with 90 degree door angle and an L / D ratio of 5 leave the cavity in two different launch times. The cavity door was bent 30 degrees outward in order to prevent the store from hitting the cavity door, and the store separation from the clean cavity with a door with 60 degree door angle was simulated. The store was able to leave the cavity from the clean cavity with a door with 60 degree door angle without touching any part of cavity, but an unacceptable trajectory was achieved due to the angle of attack increased up to 20 degrees and the angular speed increased up to 80 degrees / second. This trajectory followed by the store was aimed to be improved by applying three different passive flow control methods. It has been demonstrated by experimental and several numerical studies that the inclination of the rear wall and the placement of a rectangular spoiler in front of the cavity decreases the OASPL inside the cavity in the M219 experimental study. In this study, together with these two passive methods the third passive control method which uses these two methods together were used in internal store separation. It has been determined that only tilting the trailing edge of the cavity at an angle of 45 degrees increases the maximum OASPL on the store by 2 dB, but improves the trajectory of store slightly in terms of angle of attack and angular velocity. The rectangular spoiler placed in front of the cavity reduced the maximum OASPL on the store by 4 dB and showed the best improvement in the trajectory of store at all angles and angular velocities. The combined passive control method, which a spoiler is placed in front of the cavity and tilting the trailing edge of the cavity, has been the best method in terms of acoustics by reducing the maximum OASPL on the store by 7 dB. This passive method was not more effective in terms of the trajectory of store than the passive control method which only has a spoiler in front. Finally, the effect of cavity width on store separation was investigated by performing a clean cavity with a width-depth (W / D) ratio of 1.25 and a door angle of 90 degrees. Increasing the cavity width both increases the OASPL created on the store and distract the trajectory of store from the desired one. It has been shown that bending the cavity doors outwards instead of increasing the cavity width gives better results both in terms of acoustics and the trajectory of store. Keywords: CFD, Dynamic Mesh, Store Separation, Kavity Flow, Passive Flow Control Methods