|
Industrial Fluid Properties |
IFPSC |
Simulation Collective |
 |
Molecule Transferability
The objective of this section of the 2006 IFPSC is to test the
transferability of a force field between similar molecules for a given property.
A computer modeling practitioner working in an industrial environment will be
called on to predict properties for a range of different materials - in many
cases these will be different from the materials that were used to parameterize
force fields. A force field that is able to predict properties for new materials
that were not in the original "learning set" may be more valuable in this
situation than one which may provide more accurate results - but only for those
materials for which it was specifically created.
The ability of molecular simulation methods to predict properties for "new"
materials for which little or no experimental data is available is often used as
a justification for their development (in contrast to "group-additivity"-type
methods which may fail for totally new compounds). We want to
test/promote/validate this capability.
The Molecule Transferability Challenge
Compute the low-strain rate limit (Newtonian) viscosity for 5 different low
molecular weight organic fluids at three different state points (total of 15
calculations). The molecules are related and contain only carbon, hydrogen, and
oxygen.
Background
Viscosity is a fundamental transport property of primary importance in
lubrication, having a great impact on heat transfer, friction and wear
characteristics, and energy efficiency of lubricated contacts.
Fluids entrained in lubricated contacts of non-conforming surfaces typically are
compressed to very high pressures, in the range of 1 GPa or more. Experimental
data describing viscosity at such pressures are scarce and are not easily
predicted on the basis of viscosity data at atmospheric pressure.
While real lubricants are typically of larger molecular weight than the small
molecules considered in this challenge, a force field that successfully
demonstrates the properties of transferability demanded in this contest
challenge could be used to predict the properties of larger materials of
interest as potential lubricant basestock components.
Rules of the game:
Any theory/modeling/simulation method may be used as long it complies with
the rules and restrictions specified for the challenge problem. For
example, non-force-field-based methods must comply with the spirit of the
restrictions placed on force fields. This compliance will be evaluated by
the judges on a case-by-case basis.
The force field employed must be capable of describing the interactions of
each of the molecules in a consistent manner. The same force field must be used
in all calculations.
Any force field previously published in the open literature prior to the
announcement of this challenge is acceptable, as long as its development adheres
to the condition above.
Force fields may be parameterized using any physical property data (including
viscosity) describing mono-functional molecular species, including the relevant
class of molecules. No data from multi-functional species from the relevant
class of molecules may be used in the force field development.
Contest Scoring
Quantitative Accuracy (50%)
At each state point, full credit will be awarded for predictions within the
acceptable uncertainty limit of 5% of the experimental value.
A linear interpolation of partial credit will be awarded for predictions with an
absolute deviation above the minimum threshold and a maximum of 30%. No points
will be awarded for prediction above the maximum deviation.
Relative Ranking (50%)
For a given state point, the computed viscosities of each species will be
compared to the experimental values. The molecule that is in best agreement to
experiment will be used as a normalization of the remaining values to determine
relative viscosities. The relative ratios will be compared to the corresponding
normalized ratios using the experimental data.
The quantity (eta/etaref)simulation will receive full credit if it is
within 5% of the corresponding quantity (eta/etaref)from expt. As in
the quantitative accuracy section, similar partial credit will be awarded on a
sliding scale.