| Autor: |
Francisco A. Cubillos, Luis Henriquez-Vargas, Valeri Bubnovich |
| Rok vydání: |
2011 |
| Předmět: |
|
| DOI: |
10.1016/b978-0-444-53711-9.50137-1 |
| Popis: |
In this work we study by simulations the dynamic optimization of the burner power and positioning of the combustion front in a uniform porous media combustor by using a non linear model predictive control (NMPC) scheme. In this approach the combustion front can be fixed to a desired value by changing the velocity of the combustion wave by the manipulation of the operational variables: filtration velocity and composition of the gas mixture. The NMPC control strategy was a receding horizon control scheme. A grey-box neural type model (GNM) was used to describe the combustion wave dynamics as inner model in the NMPC strategy. The GNM is based on fundamental conservation laws associated with neural networks (NN) used to model uncertain parameters. In this case, the NN was used to estimate the gain factor in a 2 ° order prediction model. The study was carried out by means of simulations considering an experimental filtration combustion (FC) burner (CH4/air mixtures) manipulating the gas filtration velocity and fuel equivalence ratio. Controlled variable was the flame position along the burner and the firing rate. Results show that confinement of the combustion front and firing rate setpoint can be achieved by manipulating operational variables giving two operation modes for the combustion front: cyclic and setpoint tracking schemes. |
| Databáze: |
OpenAIRE |
| Externí odkaz: |
|
