Supervisory control of complex propulsion subsystems
Doktorsavhandling, 2022

Modern gasoline and diesel combustion engines are equipped with several subsystems with the goal to reduce fuel consumption and pollutant exhaust emissions. Subsystem synergies could be harnessed using the supervisory control approach. Look-ahead information can be used to potentially optimise power-train control for real time implementation. This thesis delves upon modelling the exhaust emissions from a combustion engine and developing a combined equivalent objective metric to propose a supervisory controller that uses look-ahead information with the objective to reduce fuel consumed and exhaust emissions.

In the first part of the thesis, the focus is on diesel engine application control for emissions and fuel consumption reduction. Model of exhaust emissions in a diesel engine obtained from a combination of nominal engine operation and deviations are evaluated for transient drive cycles. The look ahead information as a trajectory of vehicle speed and load over time is considered. The supervisory controller considers a discrete control action set over the first segment of the trip ahead. The cost to optimise is defined and pre-computed off-line for a discrete set of operating conditions. A full factorial optimisation carried out off-line is stored on board the vehicle and applied in real-time. In a first proposal, the subsystem control of the after-treatment system comprising the lean NOx trap and the selective reduction catalyst is considered. As a next iteration, the combustion engine is added to the control problem. Simulation comparison of the controllers with the baseline controller offers a 1 % total fuel equivalent cost improvement while offering the flexibility to tailor the controller for different cost objective.

In the second part of the thesis, the focus is on cold-start emissions control for modern gasoline engines. Emissions occurring when the engine is started until the catalyst is sufficiently warm, contribute to a significant proportion of tailpipe pollutant emissions. Electrically heated catalyst (EHC) in the three way catalyst (TWC) is a promising technology to reduce cold-start emissions where the catalyst can be warmed up prior to engine start and continued after start. A simulation framework for the engine, TWC with EHC with focus on modeling the thermal and chemical interactions during cold-start was developed. An evaluation framework with a proposed equivalent emissions approach was developed considering the challenges associated with cold-start emission control. An equivalent emission optimal post-heating time for the EHC is proposed that adapts to information which is available in a real-time on-line implementation. The proposed controller falls short of just 1 % equivalent emissions compared to the optimal case.

Look-ahead control

engine control

exhaust after-treatment

supervisory control

HC2, Hörsalsvägen 14
Opponent: Prof Frank Willems, Eindhoven University of Technology

Författare

Dhinesh Vilwanathan Velmurugan

Chalmers, Elektroteknik, Signalbehandling och medicinsk teknik

Diesel Engine Emission Model Transient Cycle Validation

; Vol. 49(2016)p. 1-7

Paper i proceeding

Supervisory Controller for a Light Duty Diesel Engine with an LNT-SCR After-Treatment System

SAE Technical Papers,; Vol. 2018-September(2018)

Artikel i vetenskaplig tidskrift

Dhinesh Velmurugan, Tomas McKelvey, Jan-Ola Olsson, "Data-driven near-optimal on-line control for an electrically heated catalyst equipped gasoline engine"

Passenger cars account for more than 80 % of inland passenger transport in the EU (measured in passenger-kilometres). Cars with modern gasoline and diesel combustion engines are equipped with several subsystems with the goal to reduce fuel consumption (and thereby carbon dioxide emissions) and pollutant exhaust emissions (such as carbon monoxide, hydrocarbons, particulate matter and oxides of nitrogen). The control of these emissions are necessary to limit climate change, reduce environmental and human health impact. Regulatory bodies have been setting tougher pollutant emission limits beginning in 1990's. The compliance to these regulations have been achieved in combustion engine systems using incremental addition of several subsystems. These subsystems have different response time, performance limitations and energy impacts. Subsystem synergies could be harnessed using a holistic control approach. Look-ahead prediction information can also be used to potentially optimise power-train control for real-time implementation.

This thesis focuses on modeling the exhaust emissions from a combustion engine, formulating a combined equivalent metric and developing a supervisory control approach for simultaneous reduction of energy consumed and exhaust emissions. Modeling exhaust emissions in a diesel engine obtained from a combination of nominal engine operation and deviations are modelled and verified. An equivalent fuel cost for the diesel engine and the after-treatment system, including the lean NOx trap (LNT) and the selective reduction catalyst (SCR), is formulated. A supervisory control approach is proposed for the control of the diesel engine, LNT and SCR when look-ahead predictive information is available.

Cold-start emissions in modern gasoline engines, occurring until the three-way catalyst (TWC) is sufficiently warm, contribute to a significant proportion of tailpipe pollutant emissions. Electrically heated catalyst (EHC) in the TWC is a promising technology to reduce cold-start emissions where the catalyst can be warmed up prior to engine start and continued after start. A simulation framework for the combustion engine, TWC with EHC is developed and an evaluation framework with equivalent emissions is proposed. An equivalent emission optimal post-heating time for the EHC is proposed that adapts to information that is available in a real-time on-line implementation.

MultiMEC - Multivariabla metoder för energieffektiv motorstyrning

VINNOVA (2014-06249), 2015-03-01 -- 2018-12-31.

Styrkeområden

Transport

Ämneskategorier

Energiteknik

Farkostteknik

Reglerteknik

ISBN

978-91-7905-627-8

Doktorsavhandlingar vid Chalmers tekniska högskola. Ny serie: 5093

Utgivare

Chalmers

HC2, Hörsalsvägen 14

Online

Opponent: Prof Frank Willems, Eindhoven University of Technology

Mer information

Senast uppdaterat

2022-06-30