Advanced

Model-Based Control of Gasoline Partially Premixed Combustion

Ingesson, Gabriel LU (2015)
Abstract
Partially Premixed Combustion (PPC) is an internal combustion engine concept that aims to yield low NOx and soot emission levels together with high engine efficiency. PPC belongs to the class of low temperature combustion concepts where the ignition delay is prolonged in order to promote the air-fuel-mixture homogeneity in the combustion chamber at the start of combustion. A more homogeneous combustion process in combination with high exhaust-gas recirculation (EGR) ratios gives lower combustion temperatures and thus decreased NOx and soot formation. The ignition delay is mainly controlled by temperature, gas-mixture composition, fuel type and fuel-injection timing.



It has been shown that PPC run on gasoline fuel can... (More)
Partially Premixed Combustion (PPC) is an internal combustion engine concept that aims to yield low NOx and soot emission levels together with high engine efficiency. PPC belongs to the class of low temperature combustion concepts where the ignition delay is prolonged in order to promote the air-fuel-mixture homogeneity in the combustion chamber at the start of combustion. A more homogeneous combustion process in combination with high exhaust-gas recirculation (EGR) ratios gives lower combustion temperatures and thus decreased NOx and soot formation. The ignition delay is mainly controlled by temperature, gas-mixture composition, fuel type and fuel-injection timing.



It has been shown that PPC run on gasoline fuel can provide sufficient ignition delays in conventional compression-ignition engines. The PPC concept differs from conventional direct-injection diesel combustion because of its increased sensitivity to intake conditions, its decreased combustion-phasing controllability and its high pressure-rise rates related to premixed combustion, this puts higher demands on the engine control system. This thesis investigates model predictive control (MPC) of PPC with the use of in-cylinder pressure sensors. Online heat-release analysis is used for the detection of the combustion phasing and the ignition delay that function as combustion-feedback signals. It is shown that the heat-release analysis could be automatically calibrated using nonlinear estimation methods, the heat-release analysis is also a central part of a presented online pressure-prediction method which can be used for combustion-timing optimization.



Low-order autoignition models are studied and compared for the purpose of model-based control of the ignition-delay, the results show that simple mathematical models are sufficient when anipulating the intake-manifold conditions. The results also show that the relation between the injection timing and the ignition delay is not completely captured by these types of models when the injection timing is close to top-dead-center. Simultaneous control of the ignition delay and the combustion phasing using a dual-path EGR system, thermal management and fuel injection timings is studied and a control design is presented and evaluated experimentally.



Closed-loop control of the pressure-rise rate using a pilot fuel injection is also studied and the multiple fuel-injection properties are characterized experimentally. Experiments show that the main-fuel injection controls the combustion timing and that the pilot-injection fuel could be used to decrease the main fuel injection ignition delay and thus the pressure-rise rate. The controllability of the pressure-rise rate was shown to be higher when the pilot injection was located close to the main-fuel injection. A pressure-rise-rate controller is presented and evaluated experimentally.



All experiments presented in this thesis were conducted on a Scania D13 production engine with a modified gas-exchange system, the fuel used was a mixture of 80 % gasoline and 20 % N-heptane (by volume). (Less)
Please use this url to cite or link to this publication:
author
supervisor
organization
publishing date
type
Thesis
publication status
published
subject
keywords
Pressure-Sensor Feedback, Model Predictive Control, Model-Based Controller Design, Partially Premixed Combustion
pages
121 pages
publisher
Department of Automatic Control, Lund Institute of Technology, Lund University
language
English
LU publication?
yes
id
cee20a01-bd99-4334-a9fe-a98b45eb9174 (old id 8301285)
date added to LUP
2015-12-07 11:35:43
date last changed
2017-11-13 15:38:48
@misc{cee20a01-bd99-4334-a9fe-a98b45eb9174,
  abstract     = {Partially Premixed Combustion (PPC) is an internal combustion engine concept that aims to yield low NOx and soot emission levels together with high engine efficiency. PPC belongs to the class of low temperature combustion concepts where the ignition delay is prolonged in order to promote the air-fuel-mixture homogeneity in the combustion chamber at the start of combustion. A more homogeneous combustion process in combination with high exhaust-gas recirculation (EGR) ratios gives lower combustion temperatures and thus decreased NOx and soot formation. The ignition delay is mainly controlled by temperature, gas-mixture composition, fuel type and fuel-injection timing. <br/><br>
<br/><br>
It has been shown that PPC run on gasoline fuel can provide sufficient ignition delays in conventional compression-ignition engines. The PPC concept differs from conventional direct-injection diesel combustion because of its increased sensitivity to intake conditions, its decreased combustion-phasing controllability and its high pressure-rise rates related to premixed combustion, this puts higher demands on the engine control system. This thesis investigates model predictive control (MPC) of PPC with the use of in-cylinder pressure sensors. Online heat-release analysis is used for the detection of the combustion phasing and the ignition delay that function as combustion-feedback signals. It is shown that the heat-release analysis could be automatically calibrated using nonlinear estimation methods, the heat-release analysis is also a central part of a presented online pressure-prediction method which can be used for combustion-timing optimization.<br/><br>
<br/><br>
Low-order autoignition models are studied and compared for the purpose of model-based control of the ignition-delay, the results show that simple mathematical models are sufficient when anipulating the intake-manifold conditions. The results also show that the relation between the injection timing and the ignition delay is not completely captured by these types of models when the injection timing is close to top-dead-center. Simultaneous control of the ignition delay and the combustion phasing using a dual-path EGR system, thermal management and fuel injection timings is studied and a control design is presented and evaluated experimentally.<br/><br>
<br/><br>
Closed-loop control of the pressure-rise rate using a pilot fuel injection is also studied and the multiple fuel-injection properties are characterized experimentally. Experiments show that the main-fuel injection controls the combustion timing and that the pilot-injection fuel could be used to decrease the main fuel injection ignition delay and thus the pressure-rise rate. The controllability of the pressure-rise rate was shown to be higher when the pilot injection was located close to the main-fuel injection. A pressure-rise-rate controller is presented and evaluated experimentally.<br/><br>
<br/><br>
All experiments presented in this thesis were conducted on a Scania D13 production engine with a modified gas-exchange system, the fuel used was a mixture of 80 % gasoline and 20 % N-heptane (by volume).},
  author       = {Ingesson, Gabriel},
  keyword      = {Pressure-Sensor Feedback,Model Predictive Control,Model-Based Controller Design,Partially Premixed Combustion},
  language     = {eng},
  note         = {Licentiate Thesis},
  pages        = {121},
  publisher    = {Department of Automatic Control, Lund Institute of Technology, Lund University},
  title        = {Model-Based Control of Gasoline Partially Premixed Combustion},
  year         = {2015},
}