58744 - Binary Repetitive Model Predictive Active Flow Control Applied to an Annular Compressor Stator Cascade With Periodic Disturbances 

Novel pressure gain combustion concepts invoke periodic flow disturbances in the last stator rows of a gas turbine. The paper presents studies of mitigation efforts on the effects of these periodic disturbances on an annular compressor stator rig. For this purpose, investigations were carried out using a low-speed wind tunnel featuring a throttling disc downstream the compressor stator. The stator passages were equipped with pneumatic active flow control (AFC) by means of hub-sidewall actuation, influencing the suction side of the stator blade. The actuators were connected to a closed-loop controlled pressure tank via binary fast switching solenoid valves. Thus, enabling a variation of the actuation amplitude and timing of the AFC.

In a first step, a steady blowing case is investigated, in which the passages around a traversable blade with 30 static pressure taps are actuated with five different momentum coefficients c_μ. Presenting the two-dimensional c_p-distribution on the blade it is shown that with increasing values of c_μ the extension of the corner vortex can be lowered effectively. These results could be confirmed through the evaluation of corresponding oil flow visualizations.

However, the impact of the periodic disturbances induced by the throttling disc could not be attacked specifically enough. Due to the constant actuation, the effect of the disturbances on the cp-distribution is lowered in the pressure rising phase of the disturbance, but also increased in its lowered pressure phase.

Additionally, the inflow and outflow planes of one stator passage have been investigated using a traversable five-hole probe. It is shown that although the effects of the disturbances could not be addressed specifically the static pressure rise coefficient C_p of a passage can be increased with steady blowing actuation. Though, if one considers an adapted total pressure loss coefficients ζ^*, accounting for the total pressure of the actuation, it can be shown that the efficiency of the stator row is decreasing with higher actuation amplitudes due to increasing costs of actuation mass flow.

Consequently, a closed-loop approach is presented, to address the effects of the disturbances with more purpose and lower the costs of actuation. For that, a blade instrumented with high dynamic pressure sensors is used and a surrogate control variable with the usage of the first principle component of the effects of the disturbances is defined. Finally, a single-input single-output model for the control system could be identified so that an optimization-based control approach, specifically a repetitive model predictive control (RMPC) could be applied. Latter takes advantage of the periodic nature of the induced disturbances to increase the control performance. To address the binary character of the solenoid valves the control sequence during online optimization is defined in the binary domain and a Branch & Bound algorithm combined with a quadratic program solver is used for real-time calculation of the control output. With respect to the rotational symmetry of the test rig the calculated control trajectory is transferred to all other passages considering the phase of the disturbances on each passage, so that the AFC for the whole stator row is closed-loop.

Finally, it is shown that in comparison to the actuation with steady blowing, the reduction of the corner vortex is lower, but the effect of the periodic disturbance can be addressed significantly better. As a result, C_p is raised in a similar magnitude but with less than 50% of the actuation mass flow, resulting in a much lower ζ^* for similar values of c_μ.