Fast Model Predictive Control for Combustion Instabilities


Fabian Jarmolowitz, Institute of Automatic Control

Title Fast Model Predictive Control for Combustion Instabilities
When 06.05.2010
Where Lecture Room Informatik 11
Abstract Homogeneous low-temperature combustion is a promising technique to reduce emissions of combustion systems and therefore a future concept for combustion engines and gas turbines. The drawback of this approach is its susceptibility to combustion instabilities like thermoacoustic instabilities. These instabilities are detrimental for performance, emissions and durability of the combustor components. One way to avoid these instabilities is to use active (closed-loop) control and consequently to enlarge the range of stable operations. In order to control a combustion system, the fundamental mechanisms have to be understood and simple dynamical models have to be derived for controller design. As control methodology we propose Model Predictive Control (MPC) which is based on formulating the control problem as an (open-loop) optimization problem which is solved iteratively in every time step. It is shown that MPC is nowadays applicable to very fast processes like combustion instabilities.
Slides download