The Problem:
Stealth is a major driver in military aircraft design. The need for low radar cross-section implies that stores will be carried in internal weapon bays. However, these weapons bays present a cavity like geometry to the external air flow when the bay doors are opened. For cavities that are deep, very high unsteadiness is observed, with sound pressure levels of 170 dB observed in wind tunnel measurements. This level of unsteadiness leads to concern about structural integrity and the performance of the electronics on the store.
The Challenge:
There is significant interest in using flow control to reduce the unsteadiness in the weapons bay. Concepts such as leading edge spoilers, trailing edge sloping, pulsed jets and active jets have been considered. Computer simulation of these devices will be a crucial step towards deployment because of the insights that this would provide into mechanisms, the ability to simulate at full scale conditions, and the potential to run optimisation studies. However, the cavity flow problem is one of the most difficult facing CFD. This largely arises from the range of time scales important in the problem. The deep cavity tends to resonate at distinct tones which at a laboratory scale range from 100Hz to 1000 Hz. In effect, a calculation needs to be run for a long time (determined by the lowest frequency) at small time steps determined by the highest frequency.
Use of HPCx:
There had been a long standing effort to exploit RANS for deep cavity simulations. This effort had essentially stalled since there is an insufficient spectral gap between the mean flow and turbulence frequencies. Access to HPCx allowed turbulent simulation to be considered. DES and LES calculations were run. For example, a DES calculation on the standard clean cavity with doors in the vertical up configuration was computed on 300 processors using a grid with 4 million points and 50,000 time steps. The code used was the Parallel Multiblock (PMB) code. This required 3.5 days of wall clock time.
Outcome:
The results were compared with measurements made by John Ross at DERA Bedford. The results which used turbulent simulation consistently showed a response in the same modes as the measurements, and crucially this agreement was insensitive to mesh and time refinement. The detailed comparisons of the results showed a level of agreement which makes the simulation a credible tool for the types of the studies described above.
Future:
The demonstration of the basic simulation capability is now being exploited for studies into flow control, supported by EPSRC EP/C533380/1
Contact: George Barakos
References:
- Nayyar, P, Barakos, G.N and Badcock, K.J., Numerical Study of Transonic Cavity Flows using Large-Eddy and Detached-Eddy Simulation, Aeronautical Journal, 111 (1117), March, 2007,165-174.
- Nayyar, P., Barakos, G.N. and Badcock, K.J. (2005), Analysis and Control of Weapon Bay Flows, RTO-MP-AVT-123. NATO RTO
|
