Accueil || Parcours || Recherche || S'enregistrer || Mon Compte || Contacts || Aide || Langues
Chauvin, Jonathan (2006) Estimation et contrôle d’un moteur diesel HCCI. Estimation des systèmes périodiques. Doctorat Mathématiques et Automatique, CAS- Centre Automatique et Systèmes, ENSMP p.230.
Plein texte disponible en tant que :
|
|
Résumé
Abstract. — Homogeneous Charge Compression Ignition (HCCI) combustion is characterized by a very high rate of Exhaust Gas Recirculation (EGR). This improves mixing and dilution in the cylinders, reduces pollutant formation at the expense of combustion stability. Thus HCCI engines requires real-time control to ensure a good trade–off between performance (in terms of torque production and low pollutant emissions) and combustion stability. Such closed-loop control are based on estimation of combustion parameters that are not directly measured.
This thesis, supported by IFP (Institut Français du Pétrole), proposes some control algorithms that have been tested experimentally on a 4–cylinders HCCI engine developed by IFP. We decompose the control synthesis in three steps. We propose solutions with experimental validations for the first two steps.
The first step is air path control. The goal is to estimate and to control the masses entering in the cylinders (fresh air and burned gas). These masses are directly related to collector pressure, compositions and flow-rates. These variables are estimated via nonlinear observers using commercial cars sensors. Design and theoretical convergence proof follow linearization via output injection and Lyapunov argument. Feedforward control based on motion planning for differentially flat systems are used to derive the flow-rate set points (fresh air and EGR). This feedforward control takes explicitly physical input constraints into account. Finally, fast Proportional Integral (PI) controller are designed to track these step points using as measured values the above estimations. We describe experimental results for large torque transient and also for driving phases of the eurocycle.
The second step is cylinder balancing. The goal is to estimate and control the combustion parameters in order to guarantee that all the cylinders have the same combustion in any steady-state regime. For that, we design instantaneous torque and cylinder individual air/fuel ratio (AFR) observers using commercial car sensors. We exploit here the high frequency information contained in the measured signals (sampling of 6 degree crank angle). Experimental results are reported. These results are based on a new class on asymptotic observers of an arbitrary numbers of Fourier modes associated to an unknown periodic input entering a linear time-periodic system. These observers outperform Kalman filters in terms of computation burden. Design and convergence proof are based on averaging techniques. A gain design methodology is proposed and justified for large numbers of modes via extension to infinite dimension of the finite-dimensional convergence analysis.
The third step is the fuel path control. During large transient, the fuel path must follow the slower air path transient. We describe this still open problematic and point out its main difficulties.
| Type d'EPrint: | Thèse (Doctorat) |
|---|---|
| Directeur de Thèse: | Rouchon, Pierre |
| Date: | 29 Septembre 2006 |
| Jury de Thèse: | Sorine, Michel et Canudas-de-wit, Carlos et Guzzella, Lino et Corde, Gilles et Petit, Nicolas et Rouchon, Pierre |
| Discipline: | Mathématiques et Automatique |
| Fonds: | ENSMP |
| Institution: | ENSMP |
| Laboratoire: | CAS- Centre Automatique et Systèmes |
| Prix: | Prix Paristech 2007 |
| Sujets: | 1. Mathématiques et leurs applications |
| Mots-clés libres: | Engine Control, Hcci, Air path control, Cylinder balancing, Air fuel ratio observer, Torque observer, Motion planning, Periodic sysems, Periodic observers, Dynamic inversion |
| Code ID: | 2804 |
| Déposé par : | Jonathan CHAUVIn |
| Déposé le : | 30 Juillet 2007 |
Table des Matières
Introduction Générale - xi
General Introduction - xv
Notations and Acronyms - xix
Acronyms - xix
Engine notations - xx
Control notations - xxi
Part I. HCCI dynamics and control problems - 1
Présentation de la Partie I - 3
Presentation of Part I - 5
1. From conventional Diesel to HCCI engine - 7
1.1. Introduction - 7
1.2. Diesel engine - 8
1.3. Why HCCI combustion? - 18
1.4. HCCI combustion properties - 20
2. Diesel engine control basics - 25
2.1. Introduction - 26
2.2. Torque control structure - 26
2.3. Air path actuators impact on the combustion - 28
2.4. Fuel path actuators impact on the combustion - 33
3. Control problems - 37
3.1. Introduction - 38
3.3. Cylinder balancing issues - 42
3.4. Fuel path control issues - 47
Part II. Air path control and cylinder balancing - 51
Présentation de la Partie II - 53
Presentation of Part II - 55
4. Air path control - 57
4.1. Introduction - 58
4.2. Control problem - 60
4.3. Intake manifold modelling - 60
4.4. System properties - 63
4.5. Air path observer - 68
4.6. Air path feedforward control - 81
4.7. Air Path Control Feedback - 87
4.8. Experimental results - 89
4.9. Conclusions and Future work - 95
5. Cylinder individual Air/Fuel Ratio estimation - 97
5.1. Introduction - 97
5.2. Modelling - 99
5.3. Cylinder-individual AFR nonlinear observer - 103
5.4. Cylinder-individual AFR Kalman filter - 105
5.5. Experimental Results - 109
5.6. Observers implementation - 110
5.7. Robustness toward parametric errors - 115
5.8. Experimental closed-loop validation - 117
6. Indicated torque estimation - 119
6.1. Introduction - 119
6.2. Transmission Dynamics inversion - 121
6.3. Indicated torque estimation - 124
6.4. Conclusions - 131
Part III. Estimation for linear-time periodic systems - 133
Présentation de la Partie III - 135
Presentation of Part III - 137
7. Periodic input estimation for linear periodic systems - 139
7.1. Introduction - 140
7.2. Problem statement and observer design - 141
7.3. Convergence results in the case of full state measurement - 144
7.4. Convergence results in the case of time invariant partial state measurement . . 145
7.5. Convergence results in the case of time periodic partial state measurement . . 148
7.6. Other applications - 153
8. H1 periodic input estimation for linear bounded systems with full
measurement - 159
8.1. Introduction - 159
8.2. Problem statement and observer design - 160
8.3. Convergence proof - 162
A. Personal Publications and Patents - 167
B. Experimental setup - 171
B.1. NADITM engine configuration - 171
B.2. Experimental set up - 172
C. From simulation to control - 175
C.1. Engine Simulator Design - 175
C.2. From Reference Simulator to Real Time Simulator - 179
D. Convergence of the periodic Kalman filter - 183
D.1. Combustion Torque estimation using Kalman filtering - 183
D.2. Individual AFR estimation convergence proof - 188
Bibliography - 193
Administrateurs de l'archive uniquement : éditer cet enregistrement