Feedback control of combustion oscillations

Combustion is a complex process which is naturally unstable. Its instabilities are generally violent pressure oscillations in the combustion chamber. A non-linear model with extensive time-delays has been employed to express the combustion oscillations, and a linear time-invariant model was derived....

Full description

Saved in:
Bibliographic Details
Main Author: Tjioe, Hadi Widjaya.
Other Authors: Chu Yun Chung
Format: Final Year Project
Language:English
Published: 2010
Subjects:
Online Access:http://hdl.handle.net/10356/39799
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Nanyang Technological University
Language: English
Description
Summary:Combustion is a complex process which is naturally unstable. Its instabilities are generally violent pressure oscillations in the combustion chamber. A non-linear model with extensive time-delays has been employed to express the combustion oscillations, and a linear time-invariant model was derived. Model Predictive Control, a model-based control technique is applied in order to suppress the oscillations. This report is intended to explain systematic approaches to design the controller. It was started by designing controller for linear time-invariant model and for non-linear model afterwards. Subsequently, two similar controllers were implemented. MPC with terminal cost on its cost function which is referred as Constrained Infinite Horizon Linear Quadratic Regulator, CIHLQR, and original MPC without terminal cost on its cost function. Following that, parameters effects such as weighting factors, prediction horizon, etc, are analyzed to have better understanding of its contribution to the overall controller performance. Lastly, analysis of constraints which commonly represent actuator limits and sometimes unmeasurable was carried out. Subsequently, simulations are performed to show the effectiveness of MPC as well as CIHLQR. Both methods with appropriate tuning are shown to be able to stabilize the combustion process.