Experimental Study of Linear Closed-Loop Control of Subsonic Cavity Flow

Abstract

A study is presented of the modeling and implementation of different concepts for linear feedback control of a single-mode resonance shallow cavity flow. When a physics-based linear model is used for cavity pressure oscillations, an H ∞ controller was designed and tested experimentally. It significantly reduced the main Rossiter mode for which it was designed, while leading to strong oscillations at other Rossiter modes. Other linear control methods such as Smith predictor controller and proportional integral derivative (PID) controller exhibited similar results. The ineffectiveness of using fixed linear models in the design of controllers for the cavity flows is discussed. A modification of the PID design produced a parallel-proportional with time-delay controller that remedied this problem by placing zeros at the frequencies corresponding to other resonance states. Interestingly, it was observed that introducing the same zero to the H ∞ controller can also successfully avoid the strong oscillations at other Rossiter modes otherwise observed in the single-mode-based design. The parallel-proportional with time-delay controller was compared to a very effective open-loop method for reducing cavity resonance and exhibited superior robustness with respect to departure of the Mach number from the design conditions. An interpretation is presented for the physical mechanisms by which the open-loop forcing and the parallel-proportional with time-delay controllers reduce the cavity flow noise. The results support the idea that both controls induce in the system a rapid switching between modes competing for the available energy that can be extracted from the mean flow.

Modeling and Feedback Control for Subsonic Cavity Flows: A Collaborative Approach

“…8 (b), respectively. The closed-loop SPL spectra obtained with the scaled LQ controller resemble those previously obtained by this group using a proportional control with time delay [48]. Furthermore, the two controllers present similar robustness properties for off-design conditions considered here.…”

“…Experimental results, in qualitative agreement with the analysis on the reduced-order model, show that the controller achieves a significant attenuation of the target resonance peak, exhibits good robustness for some off-design conditions, and compares favorably with tuned open-loop strategies. Although the experimental setup is the same as the one used in [12] and [48], the modeling, identification, and control design techniques are different, as the reduced-order model considered here is nonlinear, as opposed to the linear one considered in [48]. Furthermore, while the results presented here are similar to those obtained in [48], the analysis performed on the nonlinear model has revealed a fundamental limitation posed by the bounded control authority of the actuator.…”

“…In [47,34,35], a physically motivated linear model was proposed and used. For the same model, tuned on the basis of experimental data, it has been shown in [48] that H ∞ controllers are capable of reducing the dominant resonance for which they are designed, but introduce tones at other frequencies. This suggests that linear models may not be the most adequate to describe the cavity flow dynamics, as the latter exhibit a significant nonlinearity.…”

Variable Structure Model for Flow-Induced Tonal Noise Control with Plasma Actuators

“…Several closed-loop control methodologies, such as optimal control [15], adaptive control [26], and robust control [16], that use linear models have been applied successfully in attenuating flowinduced cavity tones. The linear models in the former two designs were constructed using system identification techniques.…”

“…Our previous work [13,14] demonstrated the use of plasma actuators for attenuating cavity tonal noise that is similar to landing gear bay noise. The problem has been frequently employed as a testbed in the study of flow-induced noise control [15][16][17][18]. The simplicity and absence of any mechanical moving parts, for example, pumps, make the plasma actuator a promising option for aeroacoustic applications.…”