research/cantera
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions
3127 commits 2 branches 16 tags 37 MiB
Commit graph

3127 commits

This branch
This branch
All branches
Author SHA1 Message Date
Harry Moffat
cf730962a6 Protected against sundials not being compiled in 2011-09-13 23:32:29 +00:00
Harry Moffat
c58697dbd3 Replaced mlequ with gaussj full pivoting in a few spots. 
There were no changes above numerical roundoff in the test suite.
 2011-09-13 18:32:08 +00:00
Harry Moffat
b1b6dd945c Fixed some pivoting issues in mlequ. I've encountered a stoich matrix that 
causes this routine to signal a singular matrix when it isn't.
This iteration uses a row swapping algorithm to get rid of the problem.
Next, I'll add a gauss-jordon (with full pivoting) algorithm on top
of the row swapping algorithm.
 2011-09-13 17:49:33 +00:00
Harry Moffat
d2272b3707 Initialization for trust regions firmed up. 
Same results for the second time into the function as the first.
 2011-09-12 15:59:52 +00:00
Harry Moffat
492a814e9d Fixed an error in mdp_init_ functions 2011-09-12 15:26:28 +00:00
Harry Moffat
93d5d43a19 More cleanup 2011-09-09 21:32:50 +00:00
Harry Moffat
4d2e6b65bb Updates to the internals, trying to simplify 2011-09-09 19:16:36 +00:00
Harry Moffat
2703130a7a Started working on the documentation and cleanup of the dogleg method. 
Now possible to run the algorithm without the DEBUG_DOGLEG block on.
 2011-09-08 02:07:02 +00:00
John Hewson
8de54e200d There was a disagreement between the name of 
getElectricConduct between the base and LiquidTransport 
class.  Since the LiquidTransport class uses it, 
the base is now compatible with it.
 2011-09-03 01:00:25 +00:00
Harry Moffat
6fca9189da Added back conditional printing 2011-08-30 15:56:00 +00:00
John Hewson
58a0f42ef0 Added "export WITH_REAL_GASSES" and set default to 'y' 2011-08-24 18:50:45 +00:00
Harry Moffat
36a56093d1 Buf fixes for handling boundaries. 2011-08-15 03:57:40 +00:00
Harry Moffat
d617f43013 Added a capability to turn off writing separate log files 2011-08-15 00:54:18 +00:00
Harry Moffat
efd58ec5b7 Tweaked the algorithm in several ways. The amount of unusual behavior 
should be reduced.
 2011-08-15 00:45:16 +00:00
Harry Moffat
db3f16b6c7 Made a few member functions in MixtureFugacityTP public 
fixed an error in IonsFromNeutralVPSSTP. a member function used by the equilibrium sovler
was missing.
 2011-08-10 01:37:12 +00:00
Harry Moffat
d660415067 Rebaselined. 2011-08-01 22:07:28 +00:00
Harry Moffat
78eb7b4127 More integration of IDA into Cantera. 
added cj parameter to some existing routines.
  Added correct delta_t to the parameter list of  ResidJacEval object
 2011-08-01 15:26:20 +00:00
Harry Moffat
2f97180753 Fixed an error in the IDA_Solver class. 
Added an analytical jacobian capability to IDA_Solver.
Added column pointers as a member to DenseMatrix
 2011-07-31 23:44:45 +00:00
Harry Moffat
82e194a2b0 Added column pointers as a member to this object. This is 
needed for linking in sundance more closely.
 2011-07-31 18:58:19 +00:00
Harry Moffat
d2f885acd5 added the ida solver 2011-07-30 03:17:00 +00:00
Harry Moffat
2082b57699 Worked on getting the IDA_Solver interface operable. Previously, 
it hadn't been. It now works on one test problem.
 2011-07-30 03:08:10 +00:00
Harry Moffat
9d46965e7d Added a print_lvl check to an output statement. 2011-07-25 14:44:41 +00:00
Harry Moffat
f7d87724ae Changed the definition of the reference state of pure fluids. 
The reference state for pure fluids is defined as an ideal gas at the
current temperature of the fluid at the reference pressure.
It was defined as the real fluid properties at the current temperature
and the reference pressure. However, this definition doesn't allow
for the easy calculation of fugacities. 

Added some documentation, also, concerning the standard state
definition.
 2011-07-23 19:39:08 +00:00
Harry Moffat
ff7e66e5a3 This update adds in a Multicomponent Real Gas capability to Cantera. 
The Object is called RedlichKwongMFTP, and it implements a multicomponent real gas approximation.
Most of the functionality has been verified against sample problems involving the CO2-H2O system.
 2011-07-23 00:30:35 +00:00
John Hewson
1982d0c928 Changed return flags for errors to be unique to each return condition. 
Hopefully this will help track down convergence problems.
 2011-07-15 23:27:45 +00:00
Harry Moffat
68ed8ad548 Guarded against a divide by zero that occurred when the cauchy step 
size and residual were exactly zero.
 2011-05-30 22:19:39 +00:00
Harry Moffat
4d245501cf Update on the NonlinearSolver: 
Update to :calcSolnToResNormVector() to align it with documentation.
 2011-05-30 17:57:53 +00:00
Harry Moffat
8710440a1f Another incremental update 
- decideStep() has been documented and checked out
 2011-05-28 02:51:49 +00:00
Harry Moffat
425085627b Incremental work. Fixed one error in an output variable. 2011-05-28 00:18:33 +00:00
Harry Moffat
6eccfa16ed Worked on getting the Affine solve capability working. 
This is now working, so that we can switch to a Hessian with junk on the
diagonal when there is a condition number problem.
 2011-05-17 00:45:08 +00:00
Harry Moffat
e110f91a12 Added hooks for cholesky factorization 2011-05-02 16:13:59 +00:00
Harry Moffat
198ab18640 Added a needed function 2011-05-02 16:13:03 +00:00
Harry Moffat
ac660e53c0 Added cholesky decomp routines. 2011-04-29 23:07:27 +00:00
Harry Moffat
701dd1e1cd Added in the estimation of condition numbers for QR factored matrices. 2011-04-28 20:56:41 +00:00
Harry Moffat
24ac9bd133 Added in condition number routines for triangular matrices 2011-04-28 20:56:02 +00:00
Harry Moffat
7f485a0ba3 Worked on reducing the Print outs for DEBUG_DOGLEG option 2011-04-28 17:10:32 +00:00
Harry Moffat
60ae1d2c11 Small rebaseline for roundoff error 2011-04-19 16:45:04 +00:00
Harry Moffat
d60d189843 Fixed a bad error that crept into the solver from a recent commit 2011-04-19 16:41:45 +00:00
Harry Moffat
36b44fd278 Fixed a Makefile error. 2011-04-18 16:46:40 +00:00
Harry Moffat
1207d95824 Changed the algorithm slightly 2011-04-15 14:53:07 +00:00
Harry Moffat
5f73e33c9e Changed the algorithms a little bit to accomodate PhaseCombo, which 
doesn't have an ideal mixing term. This caused problems in places.
 2011-04-14 18:25:24 +00:00
Harry Moffat
59a800218f Changed getdlnActCoeffdlnN to be nonconst. 
added getdlnActCoeffdlnN_numderiv() function
 2011-04-14 18:24:13 +00:00
John Hewson
4b2b215da9 File named dlantr.o was misnamed in Makefile.in 
--This line, and those dlantrbelow, will be ignored--

M    Makefile.in
 2011-04-11 21:08:08 +00:00
Harry Moffat
b20a73b92a Fixed an error with unbalanced quotes 2011-04-11 20:52:00 +00:00
Harry Moffat
155d7604b5 Verified that the QR factorization works on one problem. 2011-04-07 16:14:00 +00:00
Harry Moffat
62af5826ea Added a lapack routine 2011-04-07 15:58:01 +00:00
Harry Moffat
7d4222eedf Fixed errors that were preventing linking. 2011-04-06 23:38:14 +00:00
Harry Moffat
68768e56e7 Another iteration to get the QR solve up to speed 2011-04-06 16:26:32 +00:00
Harry Moffat
fa30f63056 Added QR factorization algorithms to Cantera's fortran interface 
within the numerics directory.
 2011-04-06 15:40:04 +00:00
Harry Moffat
3994bb9438 Started an ifdef block for intermediate zeroed phases in kinetics objects. 2011-04-04 23:01:42 +00:00