Reduced-Order Modeling and Active Control of Dry-Low-Emission Combustion

Show full item record


Title: Reduced-Order Modeling and Active Control of Dry-Low-Emission Combustion
Author: Yi, Tongxun
Description: This dissertation is a complementary experimental and theoretical investigation of combustion instability and lean blowout (LBO) in dry-low-emission (DLE) gas turbine engines, aiming to understand the fundamental mechanisms and shed light on active combustion control. Combustion instability involves complicated physicochemical processes, and many of the underlying mechanisms remain unknown, despite extensive research in the past several decades. A practical control system must be able to achieve satisfactory control performances in the presence of large uncertainties, large variations, and even unknown system dynamics. Toward this goal, an observer-based controller, capable of attenuating multiple unstable modes with unknown characteristics, is developed. A mechanism suitable for online prediction of the safety margin to the onset of combustion instability is presented, which does not require knowing the unstable frequencies. The shortage of a reliable, high-frequency, proportional fuel actuator is a major technical challenge for active combustion control. A complementary theoretical and experimental study is performed on a pump-style, high-frequency, magnetostrictive fuel actuator. Improvements to the fuel setup have been made according to the model predictions, which have been experimentally shown to be beneficial to combustion instability control. The second part of this dissertation is about modeling, prediction, and control of lean blowout. The experimentally observed, “intensified”, low frequency, near-LBO combustion oscillations have been used as incipient LBO precursors, and are characterized as low-dimension chaotic behavior in the present study. The normalized chemiluminescence RMS and the normalized cumulative duration of LBO precursor events are recommended for LBO prediction in generic gas turbine engines. Linear stability analysis shows that, with decreasing equivalence ratios, a complex conjugate pair of eigenvalues emerges from three negative real ones, moves left toward the right half phase plane, and finally crosses the imaginary axis. Model predictions qualitatively and even quantitatively match the experiments. Simulation of the nonlinear WSR models shows the “triggered instability” which is similar to that in rocket motors. It is numerically demonstrated that zero-mean small-amplitude fuel modulations based on modern feedback control principles, can be very effective in strengthening the flame’s robustness to external disturbances without exacerbating the overall emissions. Experimental demonstrations are suggested for future research.
Permanent Link: http://rave.ohiolink.edu/etdc/view?acc_num=ucin1172865804
http://hdl.handle.net/2374.OX/105942
Date: 2007

Files in this item

Files Size Format View

There are no files associated with this item.

This item appears in the following Collection(s)

Show full item record