|
| |
Old "What's New" News
June 21st, 2003
** BUG DESCRIPTION **
There is a slight bug with the normalized sensitivity analysis
feature (the -SENSIT switch) of Kintecus. The bug causes Kintecus to
appear to take forever to calculate normalized sensitivity analysis
matrix files. This is noticeable if you add the "-show" switch on the
command line and you will see the time in the Time(s) column repeat
the same time forever. You will also get a CONC.TXT output file that
is extremely large and will continue to grow until it fills up the
harddisk.
** TEMPORARY FIX **
The current fix is to increase or decrease the number of normalized
sensitivity analysis matrix files you are creating using the -SENSIT
switch. For example, if you experience this error trying to output 10
normalized sensitivity analysis matrix files using -SENSIT:1:10
switch, then increase the 10 to 11 or decrease it to 9. This will
force Kintecus to select different timing meshes that do not cause it
to hang on the previous time. If you are using the other way to
compute normalized sensitivity analysis matrix files via
"-SENSIT:1:TIMES,n1,n2,n3,nn" switch, such as "-SENSIT:1:TIMES,1e-3,
2e-3, 2.5e-3, 3e-3, 4.9e-3" and suppose Kintecus is hanging on the
time of 2.5e-3 seconds, then change the 2.5e-3 time to 2.51e-3 seconds
or change it to 2.49e-3 seconds. Kintecus should be able then to
"latch on" and continue with the rest of the computation.
April 7th, 2003
- Kintecus V3.1 has been released. New features include:
a) Running Kintecus V3.1 on Pentium-IV machines will result in an almost DOUBLING in
speed. Your run's execution time will be almost cut in half!
b) Better Equilibrium solution finding for larger systems.
c) detection and correction of bad Time(s) in one's experimental data for
Fitting/Optimizations
d) Graphical Interface now includes the ability to store your perturbation files (such as
temperature programs, concentration profiles, piston compressions, etc), thermodynamic
databases, ChemName files, Periodic Tables as worksheets in your Excel Workbooks. This
allows for easy transference of entire models between groups/individuals. One can even
have a dynamic graph of a perturbation on the same worksheet that holds the perturbation
data (see the Gas_Desorption_off_Tungsten.xls Excel file) or see the
GRI-Mech_3.0.xls Excel file for a sample of holding a thermo. database in a worksheet.
e) Bug fixes. There are a number of bug fixes, one notable bug fix is that running
Kintecus in kinetics mode, under constant pressure and having a temperature profile or
concentration change of a species of more than 1 point causes Kintecus to hold the
pressure constant for the first point, but after the first point, the pressure is not held
constant. This is now fixed in V3.1.
January 30th, 2003
- *ALERT* If you are using the
Excel Graphical Interface and you are Fitting Parameters against Data then you should
download a corrected Excel worksheet HERE.
The corrected worksheet fixes the bug with experimental data in the last column of the
FITDATA worksheet being overwritten with the experimental data from the previous
column. If you are Fitting Data with Kintecus directly (you are NOT using the Excel
worksheet) then you should be fine. Also be sure to have the times in the Time(s)
column in the FITDATA worksheet to contain only 3 to 6 decimal places. Having times that
are 10 or more decimals places long (such as 1.234567890123) forces Kintecus to skip that
data point (and any other data points). It is simply impossible to force the integrator to
stop at that exact time! More example "good time
points": 1.2345E-31, or 4.54321E+50, BAD times: 0.0043322114758484 or
321.32234522344.
January 27th, 2003
- When using the Excel graphical interface and you are fitting
parameters against a dataset,
there is a bug in which Excel will overwrite the last column of data with the
previous column.
For example, if your FITDATA worksheet has Time(s) in the first column, E in the
second
column and S in the third column, when you click run, Excel will output the FITDATA
worksheet as Time(s), E and E, the third column containing data will be overwritten with
E. THIS ONLY HAPPENS IF YOU ARE USING THE
EXCEL GRAPHICAL INTERFACE. If you
directly enter your experimental data into the file "fitdata.txt" there
will be no problem. A
temporary fix is to duplicate the last column of data: i.e. in the FITDATA Excel
worksheet,
the columns: Time(s),E and S should have four columns of
data: Time(s), E,S and S
AGAIN. The next release of Kintecus and the Excel Graphical
Interface will fix this.
Also, the Periodic text file containing two columns of data: new mass/isotope
element
name, mass weights should be named Periodic.dat and NOT period.dat as stated in the
manual.
The next version of Kintecus will show a DOUBLING in SPEED using the new COMPAQ 6.6
FORTRAN compiler for Pentium IV users!
November 10th 2002
August 29th 2002
- Kintecus V3.0 has been released. The major addition to this major
upgrade are high-end equilibrium calculations without the need to purchase high-cost
software. The user wishing to do equilibrium calculations only has to specify the species
present. There is no need for the user to specify any reactions, or balance any reactions,
or even specify which species are reactants or products! Equilibrium calculations,
adiabatic (variable temperature) flame, constant pressure, multiple phases, phase
stability plots and other types of equilibrium calculations can be performed. Kintecus
fitting/optimization settings (fit initial concentrations, initial temperatures, etc,
against experimental data) , perturbation affects, system property outputs can all be
specified with the new equilibrium features. Kintecus comes with a number of sample
equilibrium runs and Excel files. Shown below is an equilibrium phase stability plot
of liquid H2SO4 slowly dispersing into various gases as the temperature is increase:
- Some minor additional features have been added to Kintecus such as
having the concentrations output in mole fractions and a more strict application to the
phases. Also a new perturbation switch has been implemented: the '-SET' switch. This
switch allows one to quickly set the concentration, temperature or volume at any time and
still have that parameter act as a variable in the model.
May 26th 2002
- Kintecus V2.80_1 has been released. The only changes were to fix
all the Excel spreadsheets so that plotting of Kintecus results by German/Russian/Polish
Excel users is correct. Yup, its that comma-dots and dots-comma Excel problem again.
Hopefully, all problems associated with those international Excel VBA sheets are now
corrected.
April 22nd 2002
- The problems some international users (Germany, Poland and Russia)
were having with the Excel Graphical Interface bombing out due to the representation of
numbers with commas has been fixed. Also, the false Kintecus bombings reported by
Excel2000 Kintecus Graphical Interface under XP or WIN95 has been corrected. Plus, now
users using the Excel Graphical Interface can monitor the progress of their Kintecus runs!
April 17th 2002
- Some international users (Germany, Poland and Russia) have noticed
that their Excel Graphical Interface bombing out. This is caused by the German, Polish,
Russian version of Excel using commas instead of points to represent numbers, ie. 3.14 is
represented by 3,14. Kintecus parses it as an invalid number. This will be fixed in the
upcoming week. Another error is Excel2000 running in WindowsXP or Windows95 reports false
Kintecus crashes. This will too be corrected in the upcoming week. For now, just ignore
the false error reported by Excel. Also, some users are complaining that they cannot
monitor the progress of their Kintecus runs in the Graphical Interface. An addition to the
Interface will be added to allow one to monitor their Kintecus runs in the Excel Graphical
Interface by next week or so...
March 10th 2002
- Kintecus for the Windows zip file now comes with an Installer. You
can directly load some of the sample graphical interface models from the Windows Start
Menu with the new Kintecus installer.
January 19th 2002
- Kintecus for LINUX has been released. Three versions are available
for Pentium I, PentiumII/PentiumIII and Pentium IV. The LINUX versions are 40-80% FASTER
than the Windows version so those with LINUX systems might be interested switching over.
December 29th 2001
- Version 2.71 of Kintecus has been released. Minor changes to
executable. The GRI-MECH sample runs have been sped up by 3-4x (300-400% increase in
speed!) due primarily to the change in the GRI-MECH's respective Accuracy field. Most
sample models for Kintecus have the Accuracy field set to a fairly conservative level
(=SLOW). The GRI-MECH sample runs and the OH-combustion runs have their Accuracy changed
to a larger value without affecting the final results. In fact, the GRI-MECH runs can be
sped up ANOTHER 2x, but with some slight changes in the final temperature &
concentrations values. Some users running Kintecus inside a large 3D diffusion/convection
grid might want to further increase the Accuracies in the GRI-MECH parm.dat files.
December 2nd 2001
- Version 2.69 of Kintecus has been released. Small bugs associated
with some error reporting routines were corrected. See bugs
page for further details.
November 10th 2001
- New Kintecus_V26_2.zip file released.
- The Chemkin to Kintecus translator program, CK2KIN.EXE,
has a serious bug in it. Depending on the conditions, it will sometimes drop a zero from a
number containing an exponent, i.e. 1.5E+30 will become 1.5E+03 or 7E+10 will become
7E+01. The Kintecus GRI-MECH-3.0 model and the ethanol combustion model (previous Chemkin
models translated into Kintecus format by CK2KIN.EXE) contain a few wrong
rates and TROE factors. Interestingly, re-running the ethanol.bat runs with the NEW
CORRECT model shows almost no difference in the ignition delay results, BUT the
GRI-MECH-3.0 model (a highly condensed kinetics mechanism) shows concentration profiles
off from 10%-40%. Although, the FINAL concentration and temperature values for the
GRI-MECH-3.0 runs are only off a few percent, primarily due to most of the reactions have
equilibrium constraints. The Kintecus V2.6 program is fine.
So, what should you do? If you do NOT use the Kintecus GRI-MECH-3.0 or
ethanol combustion models or you do NOT use the CK2KIN.EXE translator,
you can safely ignore this message. Again, Kintecus V2.6 is fine; there are
currently no known bugs in V2.6.
If you only use the ethanol combustion model, you can probably
ignore this message, but it would be wise to download the new
Kintecus Zip file (Kintecus_V26_2.ZIP). The new ZIP file contains the updated CK2KIN.EXE
program and the correct ethanol combustion model and GRI-MECH models as well as two
more additional GRI-MECH-3.0 test runs (grict2.bat and griv29.bat) to
further test the CK2KIN.EXE program against published, refereed experimental results. In
addition, the documentation has been slightly updated and a tiny .BAT file is
missing to complete the Kintecus Excel GRAPHICAL interface.
If you use the GRI-MECH model or have used the Chemkin to
Kintecus translator program CK2KIN.EXE, you should download
the new Kintecus ZIP file (Kintecus_V26_2.zip).
November 2001
- Kintecus Version 2.60 has been released
- Included in the release is a graphical interface
that runs through Excel 97/2000/XP. Click here (Graphical Interface)
to see some screen shots!
- Kintecus 2.60 allows users to run other programs to monitor the
execution of Kintecus, Fatal Errors and Warnings and also allows one to alter the output.
August 2001
- Kintecus Version 2.31 has been released
- The main new feature in Kintecus V2.3 is the ability to
fit/optimize
rate constants, initial concentrations, Lindemann/Troe/SRI/LT
parameters, enhanced third body factors, initial temperature,
residence time, energy of activation and many other parameters
against your dataset(s). Note that Kintecus will actually fit the
parameters at EXACTLY the time your data was measured. Unlike other
programs, Kintecus DOES NOT interpolate a function against your data
and then fit the values against this interpolation. There is no need
to “clean” your data, suggest interpolation methods nor specify
timing meshes against your experimental data since
Kintecus calculates values at exactly the times you specify in your
experimental datafile.
- Some other changes and additions have been
added such as the ability to view the delta-S,H,G,Kp, kr of the
reactions, and the H, S, G, Cp, Gas Density and other physical
properties of system as the simulation runs. See the .pdf file
for full descriptions.
- In addition, some users were complaining
that the Freeform spreadsheet output of Chemkin thermodynamic
databases didn't included Mw, so this has also been changed.
A new MAP keyword, CMW, has been added to convert the empirical
Mw data field into Kintecus Mw. For examples, see the ethanol
combustion, O2-OH combustion, Ozone decomposition and GRI-MECH
samples which all include this keyword. Again, please refer to the
.pdf documentation for a full description of these new features
in V2.3 and many other additions!
|
