44
That is, a simple collection of point models that are tied together to give overall system
behavior and to trade off the effects of variations of design specifications and see how
these alter total drive-cycle economics. At the same time, it was realized that many of the
important effects cannot be treated in this way vehicle aerodynamics and underhood
cooling being obvious examples. It was noted, however, that the heavy-vehicle industry
(at least in the U.S.) is most comfortable with simulation at the PC level rather than using
more computer-intensive CFD calculations such as those common in the analysis of
automotive systems.
behavior and to trade off the effects of variations of design specifications and see how
these alter total drive-cycle economics. At the same time, it was realized that many of the
important effects cannot be treated in this way vehicle aerodynamics and underhood
cooling being obvious examples. It was noted, however, that the heavy-vehicle industry
(at least in the U.S.) is most comfortable with simulation at the PC level rather than using
more computer-intensive CFD calculations such as those common in the analysis of
automotive systems.
There were a number of unanswered questions concerning the level of simulation
detail needed. Following are some examples:
·
Are issues of durability to be treated? That is, whether special models
are needed to address metal temperatures in engine systems, the effect
of temperature cycles on durability and failure rates, and how one
would link thermal models to finite-element codes that could analyze
structural questions.
are needed to address metal temperatures in engine systems, the effect
of temperature cycles on durability and failure rates, and how one
would link thermal models to finite-element codes that could analyze
structural questions.
·
There was no clear way to link the type of thermal analyses being
discussed to economic analysis or how these connections may create
constraints on alternate system configurations.
discussed to economic analysis or how these connections may create
constraints on alternate system configurations.
·
Are we talking about codes that would be used early in a design or as
part of the analysis of a completed design? This is of major importance
because early in the design, a simple system code would probably be
sufficient, while in evaluating a near-final design, a multidimensional
treatment would be needed (e.g., to evaluate alternate underhood
system configurations and the placement of components).
part of the analysis of a completed design? This is of major importance
because early in the design, a simple system code would probably be
sufficient, while in evaluating a near-final design, a multidimensional
treatment would be needed (e.g., to evaluate alternate underhood
system configurations and the placement of components).
·
It would be desirable to have all codes or systems of codes use some
consistent interface specification so data transfer between codes or
code modules would not be a problem. Who would decide such issues
and define standards that all OEMs and suppliers could live with?
consistent interface specification so data transfer between codes or
code modules would not be a problem. Who would decide such issues
and define standards that all OEMs and suppliers could live with?
The discussion shifted to what would be a reasonable first effort by DOE. That is,
what would produce tools most useful to the industry? It was suggested that DOE could
sponsor work on a nonproprietary code system that all could use. Prior to this, it would be
useful, even at the zero-dimensional level, to learn what is available in the open literature
and now in common use, what submodels exist at what level of detail, what data are
available for code validation, and what models can be borrowed directly from the auto
industry. This would show where the holes are and where model development could have
the greatest benefit to the industry as a whole. Of particular interest at all levels of detail
would be standard experimental data for code validation. Too often, individual designs
sponsor work on a nonproprietary code system that all could use. Prior to this, it would be
useful, even at the zero-dimensional level, to learn what is available in the open literature
and now in common use, what submodels exist at what level of detail, what data are
available for code validation, and what models can be borrowed directly from the auto
industry. This would show where the holes are and where model development could have
the greatest benefit to the industry as a whole. Of particular interest at all levels of detail
would be standard experimental data for code validation. Too often, individual designs