Information Technology Reference
In-Depth Information
related acceleration and speed to energy consumption per distance travelled. Non-
propulsion losses (accessory loads) were handled separately based on travel time.
The total energy consumption for each vehicle in a scenario under a given set of
conditions is then analyzed using statistical regression to provide a predictive model
for the scenario under the given conditions.
2.1
Propulsion Simulation
Propulsion system simulation is a desktop computer method for directly simulating
vehicle drive cycles using a complete model of a vehicle propulsion system that
represents key interactions between driver, environment, vehicle hardware and ve-
hicle controls. There are a number of examples of tools, methods and applications of
propulsion system simulation programs in the literature that use dynamic systems
modeling to estimate energy consumption given a drive cycle. The primary purpose of
these applications is to identify key design elements that influence performance, test
control algorithm alternatives and determine the effect specific propulsion system
features have on drivability. One significant example is the PSAT (Powertrain System
Analysis Toolkit) (Argonne National Laboratory) tool developed in 1999 as part of a
collaborative effort with U.S. OEM's (Ford, GM, and Chrysler).
Many automotive OEMs and Tier 1 suppliers have proprietary propulsion system
modeling and simulation tools with a team of developers and simulators capable of com-
puting energy consumption from drive cycles. Typically the drive cycles are from labora-
tory tests specified by the government regulations, obtained from driving studies or
created by simulation. Generally these tools are supported by databases of proprietary
hardware component information and controls strategies and calibration data specifically
representative of the manufacturer's products. They are configured to support investiga-
tion of systems that design changes can be modeled through simulation. One such model-
ing application is Ford Motor Company's Corporate Vehicle Simulation Program
(CVSP) that was used in the modeling effort described in this paper [3].
CVSP is a critical tool used mainly for projection of fuel economy capability of
vehicles with internal combustion engines. These projections are used to make critical
hardware and technology decisions that determine vehicle program content and ulti-
mately impact vehicle program cost. Results from the CVSP simulations are also used
to cascade targets to key subsystems and components (e.g. battery, power electronics
and electric machines for HEV's).
Within Ford, a significant amount of time and effort has been invested in verifying
the accuracy of CVSP simulations. This is critical for development of high confidence
fuel economy roadmaps and subsystem/component targets for vehicle programs. With
good system model accuracy, targets can be specified with much higher precision,
thus avoiding over-design of components to deliver aggressive fuel economy targets.
In the later stages of a program, an accurate system model can support vehicle testing
for fuel economy attribute development. The model can be used to assess selected
propulsion system control strategy and calibration changes which can help refine
vehicle test plans and improve efficiency of vehicle test efforts.
The vehicle system model integral to CVSP is implemented in the
Matlab/Simulink® environment using the Vehicle Model Architecture (VMA)
standard [2]. Models of each VMA subsystem are stored in libraries and inserted into
Search WWH ::
Custom Search