An active control system for the turbulent boundary layer

This thesis presents the development process and the experimental results of a system constructed to apply real-time control to the structures of the turbulent boundary layer region in order to reduce surface shear stress. The system is composed of three main components: an array of MEMS surface shear stress, τ_w sensors; a MEMS flap actuator; and a control logic which integrates the hardware components together into a closed system. The objective of this system is to reduce the stress contained in streak-like regions of high τ_w.; The sensor array, used to image the τ_w distribution, is an extension of the thermal based τ_w sensor developed by Jiang. Numerous studies have been performed using this device, the results of which have validated its performance. For this study, a new temperature compensation methodology, based on the surface temperature of the sensor chip, was employed in order to account for possible temperature variations at the wall surface.; The actuator, a pneumatically driven flap, is developed as part of the present research. The device is, in essence, a 3 mm x 1 mm cantilever beam that sits on top of an inflatable diaphragm and is capable of actuation frequencies of over 200 Hz and amplitudes of over .11 mm. When it is oscillated in the open loop mode, the effect over one cycle of motion is an average reduction by as much as 2.5% in τ_w in the region immediately downstream.; A neural network is employed to identify the streak-like regions of interest. Results have shown that this network is successful in identifying the streak-like regions of interest. The control logic employs this network in a predictive, feed-forward scheme to determine the appropriate actuator response. Offline studies have shown that under optimal conditions, the signature of the streak-like regions can be eliminated.; Online results conform well to the offline predictions. While unable to achieve the optimal conditions, online experiments show that the system is capable of responding to the streak-like regions in real-time. More significantly, the results show that the signatures of the streak-like regions are significantly altered by the control actions.