This paper discusses the impact of a nonsteady outflow condition on the compressor stator flow that is forced through a mimic in the wake of a linear low-speed cascade to simulate the conditions that would be expected in a pulsed detonation engine. 2D/3C-PIV measurements were made to describe the flow field in the passage. Detailed wake measurements provide information about static pressure rise as well as total pressure loss. The stator profile used for the investigations is highly loaded and operates with three-dimensional flow separations under design conditions and without active flow control. It is shown that sidewall actuation helps stabilize the flow field at every phase angle and extends the operating range of the compressor stator. Furthermore, the static pressure gain can be increased by 6% with a 4% loss reduction in time-averaged data.

No abstract

No abstract

This paper presents a constrained Repetitive Model Predictive Controller (RMPC) implemented as closed-loop flow controller for an experimental compressor stator cascade. The objective of the controller is to decrease the impact of periodic disturbances on the passage flow. The disturbances are generated by movable flaps that are located downstream of the trailing edges of the stator vanes. The flaps emulate the throttling effect of periodically closed combustion tubes in a pulsed detonation engine on the flow over the stator vanes. The RMPC adjusts the actuation amplitude of fluidic sidewall actuators according to the present state of the passage flow. The current flow situation is monitored by pressure sensors that are mounted flush to the surface of one of the stator vanes. This data is fed back in real-time to the RMPC which thereupon modifies the actuation amplitude. By learning from period to period, a control command trajectory is computed that reduces detrimental effects of the periodic disturbance in an optimal manner while respecting the input constraints of the physical system. Five-hole-probe measurements in the wake of the passage are utilized to compare the optimized, transient actuation trajectories to the case of constant amplitude actuation.

No abstract

No abstract

No abstract

This contribution presents the capability of iterative learning active flow control to decrease the impact of periodic disturbances in an experimental compressor stator cascade with sidewall actuation. The periodic disturbances of the individual passage flows are generated by a damper flap device that is located downstream of the trailing edges of the blades. These mimic the throttling effect of periodically closed combustion tubes in a pulsed detonation engine. For the purpose of rejecting this disturbance the passage flow is manipulated by fluidic actuators that introduce an adjustable amount of pressurized air through slots in the sidewalls of the cascade. Pressure sensors that are mounted flush to the suction surface of the middle blade provide information on the current flow situation. This data is fed back in real-time to an optimal iterative learning controller. By learning from period to period the controller modifies the actuation amplitude such that, eventually, a control command trajectory is calculated that reduces the impact of the periodic disturbance on the flow in an optimal manner.