Page 39 - ASME DSCC 2015 Program

Technical Program

Adaptive Air fuel Ratio Controls in Presence of oxygen Sensor faults A low Complexity Gain Scheduling Strategy for Explicit Model
Contributed regular paper. DSCC2015-9858 Predictive Control of a Diesel Air Path
Hassene Jammoussi, Imad Makki, Ford Motor Company, Dearborn, MI, Contributed regular paper. DSCC2015-9754
United States Mike Huang, Ilya Kolmanovsky, University of Michigan, Ann Arbor, Ann
Arbor, MI, United States, Ken Butts, Toyota Technical Center, Ann Arbor, MI,
Fault monitoring of the upstream universal exhaust gas oxygen (UEGO)
United States
sensor, as mandated by the California air resources board (CARB), is a
necessary action to maintain the performance of the operation of the A Low Complexity Gain Scheduling Strategy For Explicit Model Predictive
air-fuel ratio (AFR) control system and indicate the need for maintenance Control of A Diesel Air Path
when a fault is present which could potentially lead to exceeding the Towards ECu-Executable Control-oriented Models of a Three-Way
emissions limits. Catalytic Converter
When the UEGO sensor fault is accurately diagnosed, i.e. fault is detected, Contributed regular paper. DSCC2015-9653
direction is identified and magnitude is estimated, tuning of the controller
Mario Santillo, Steve Magner, Mike uhrich, Mrdjan Jankovic, Ford Motor
gains can be performed accurately with minimal calibration efforts. Pre-
Company, Dearborn, MI, United States
sented in this paper is a control strategy that utilizes the type, direction and
magnitude of fault detected to adapt the gains of the controller and update The nonlinear dynamics of an automotive three-way catalyst (TWC) present
the parameters of the Smith predictor (SP) in order to maintain the stability of a challenge to developing simple control-oriented models that are both use-
AFR control loop. The proposed approach has been validated on a vehicle ful for control and/or diagnostics and real-time executable within a vehicle
(Mustang V6 3.7L) equipped with ATI No-Hooks rapid prototyping system. engine-control unit (ECU). As such, we begin by developing a first-principles
Different fault types and magnitudes were tested and the tailpipe emissions control-oriented TWC model and then proceed to apply simplifications. The
were assessed on federal test procedure (FTP) cycles. TWC models are spatially discretized along the catalyst length to better
understand and exploit the oxygen-storage dynamics. The TWC models also
effective component Tuning in a diesel engine Model using Sensitivity
include the oxidation reaction of ceria by H O, which is considered import-
Analysis 2
ant since it represents the production of H2 within the catalyst. We present
Contributed regular paper. DSCC2015-9729
automated optimization routines to calibrate the TWC model along with a
Rasoul Salehi, Anna G. Stefanopoulou, University of Michigan, Ann Arbor, heated exhaust-gas oxygen (HEGO) sensor model using measured vehicle
MI, United States and emissions data. Finally, we demonstrate the combined models’ ability
Error propagation and accumulation is a common problem for system level to accurately reproduce the measured HEGO voltage using engine feedgas
engine modeling at which individually modeled components are connected constituent inputs, which is necessary for designing a robust model-based
to form a complete engine model. Engines with exhaust gas recirculation feedback controller.
(EGR) and turbocharging have components connected in a feedback
ConTRIBuTED SESSIon
configuration (the exhaust conditions affect the intake and the intake, con-
1-6-1 WP7 fuels Cells/Energy Storage
sequently, affects the exhaust), thus they have a challenging model tuning
Elijah Pierce A 4:00pm–6:00pm
process. This paper presents a systematic procedure for effective tuning of
an engine air-charge path model to improve accuracy at the system level
Session Chair: Scott Moura, University of California, Berkeley
as well as reducing the computational complexity of tuning a large set of
Session Co-Chair: Qian Wang, Pennsylvania State University
components. Based on using sensitivity analysis, the presented procedure is
used to inspect which component influences more a set of selected outputs A framework for Control oriented Modeling of PEM fuel Cells
in a model with high degree of freedom caused by many parameters of Contributed regular paper. DSCC2015-9735
different components. After selecting the influential component, which is
Benjamin l. Pence, Jixin Chen, Ford Motor Company, Dearborn, MI, United
the turbocharger in this study, further tuning is applied to parameters in the States
component to increase the overall accuracy of the complete engine model.
This paper develops a framework for along-the-channel and through-the-
The corrections applied to the aircharge path model of a 6 cylinder 13L
membrane control oriented modeling of polymer electrolyte membrane
heavy duty diesel engine with EGR and twin-scroll turbocharger was shown
(PEM) fuel cells. The initial modeling framework is spatially one-dimensional
to effectively improve the model accuracy.
by one-dimensional (1+1D) and is described by unsteady partial differential
equations
(PDEs). Numerical techniques convert the PDEs to ordinary differential and
algebraic equations that are convenient for state-space modeling. The mod-
eling framework includes two-phase, thermal, and other transient effects.
The generality of the modeling framework and its ability to be represented
in state-space form facilitate complexity reduction and control-oriented
application.

39