EHS:Embedded Systems for Feedback Control in Fluid Flow Applications

Northeastern University, Boston MA

Investigators

Abstract

Tadmor - Abstract This research focuses on the use of feedback in unsteady fluid flow control, and on the implications born out by feedback control on tightly interconnected issues of system modeling, control & observer design, algorithm development and embedded systems implementation. The research pursues the design of distributed real-time sensing, processing and actuation through the use of distributed reconfigurable hardware (FPGA) resources. Decentralized computation arises naturally in the context of sensor and actuator arrays, addressing parameterized spatiotemporal structures. FPGA resources enables applications to be realized in real-time that up to know have not been possible. FPGA implementations will give a factor of ten or more speed up to the algorithms currently being used, and become a technology enabler by allowing flow control to consider new real-time algorithms and altogether new forms of computationally intensive actuation. Effective feedback flow & combustion regulation methods are critical enablers for improving a wide range of industrial applications that involve shear layer mixing, ranging from aircraft, automotive, semiconductor and imaging, air conditioning, the chemical process industry, furnaces and incinerators, to aeroengines & ground based gas turbines for power generation. The research conducted in this project is driven by the principal benchmark of suppressing thermoacoustic instabilities in a bluff body flame holder by using distributed sensors and actuators and embedded computations. This problem gives rise to a set of case studies of escalated complexity, used to successively isolate and remove key stumbling blocks. The project vestigates methods to make such techniques more easily applicable to future uses through the development of tools and design libraries. In conjunction with the case studies, the research is developing a generic toolbox for (hybrid) models, design algorithms and embedded systems implementations in feedback flow and combustion control.