A Platform for Overall Monitoring and Diagnosis for Hybrid Vehicles
Compared with conventional vehicles, designing hybrid electric
vehicles includes new features, such as energy management and
monitoring of the electrical components. To be able to investigate
such issues a simulation platform of a hybrid vehicle, driver, and
diagnosis system is developed based on the CAPSim model library. The
simulation platform is component based, and is able to handle
different powertrain configurations. In this investigation a
parallel hybrid is modeled and parameterized to represent a long
haulage truck. To be able to easily change a model of a component in
the vehicle model, every model of a specific component use the same
sets of input and output signals. The vehicle model is based on
dynamic equations and in general simple models of the components,
since the interplay of the components is of major interest in this
investigation.
Three model based diagnosis systems are developed and implemented in
the platform with a twofolded purpose. The first purpose is to
demonstrate the feasibility of the platform. The second purpose is
to investigate issues when designing diagnosis systems on vehicle
level of a hybrid vehicle powertrain. New features, for example mode
switches in the system and a freedom in choosing operating points of
the components via the energy management, affect the diagnosis
system. The influence of these issues on the performance of the
diagnosis system is investigated by design and implementation of
three diagnosis systems on a vehicle level. The diagnosis systems
are based on three sensor configurations. Two of these consist of
several sensors and one system uses few sensors. In one of the
systems using information from several sensors, the sensors are
placed close to the components that are to be monitored, while the
sensors in the other system is based on a different sensor
configuration. All three diagnosis systems detect specific faults,
here specifically faults in the electrical components in a hybrid
vehicle powertrain, but the methodology is generic. It is shown that
there is a connection between the design of the energy management
and the three diagnosis systems, and that this interplay is of
special relevance when models of components are valid only in some
operating modes. The diagnosis system based on few sensors is more
complex and includes a larger part of the vehicle model than the
system based on several sensors placed close to the components to be
monitored.
Christofer Sundström, Erik Frisk and Lars Nielsen
2010
Page responsible: webmaster
Last updated: 2021-11-10
LiU startsida
