amrita-ebook Google Group

Re: Changing reaction rate model

2008-08-26T17:43:08Z

Ah yes, I posted that before we sorted it out.
I was able to reproduce E=10 almost identically to what was in the paper.
Right now, I'm running tests to try to find the neutral stability boundary
for different activation energies. There are small differences between the
results, but I suspect it may be a rounding issue when converting from one

Re: Changing reaction rate model

2008-08-26T09:19:28Z

Carlos, I think we sorted this one out as the issue with
not recompiling code.
Any progress on reproducing previous results?

Re: * BDY LIST EXHAUSTED *

2008-08-22T01:43:44Z

Matei,
You need to make a call to ArraySizes, as in:
ArraySizes {
NL = 10 # maximum number of grid levels
NGD = 2800 # maximum number of mesh patches
NGIJ = 70000 # maximum number of perimeter cells
NGIxJ = 200000 # maximum number of grid cells
NEX = 2000 # length of external boundaries table

* BDY LIST EXHAUSTED *

2008-08-21T22:21:54Z

in a problem where i use the internal boundary implementation using
the level set method, when cranking up the resolution and grid, i end
up with a warning
*** BDY LIST EXHAUSTED ***
right at the start during the build-up process, after which the
process appears to hang. Does anybody know what parameter need to be

Re: Changing reaction rate model

2008-08-06T19:56:42Z

Gary,
I've played around with the parameters (both the physical parameters
and the refinement parameters) and it seems that no matter how I
fiddle around with it, I can't get anything reasonable. I've uploaded
an image of the shock pressure history. This particular simulation
used E = 10, ETA = 0.5, Q = 1.7, and Ke = 0.2, but it seems that just

Re: Changing reaction rate model

2008-07-29T18:19:17Z

Thanks a lot for all the help. I'll take a look at the new solver

Re: Changing reaction rate model

2008-07-29T14:01:51Z

Carlos,
I have uploaded new 2step.mailit
which contains all the changes - you should
probably use these so we are using consistent files.
I have changed the source terms to either do
an induction step or an heat release step
during one chemical time step - I think
this is the most consistent way to do it.

Re: Changing reaction rate model

2008-07-29T12:54:10Z

OK, the error is in the ComputeZNDProfile, not in SourceTerms
- if you look at what it writes out for dimensionless U
for the ZND file, it is actually (Ds-UR)*SQRT(G),
hence UR that appears in your routine is not your
dimensionless fluid speed, but is a factor of sqrt(gamma)
out (UR is essentially the fluid speed scaled with the

Re: Changing reaction rate model

2008-07-29T11:37:11Z

I have played around with doing this differently and making sure
the above doesn't happen. I can alleviate the problem slightly, but
not cure it.
I think that numerical diffusion is playing a role here - in the model
you essentially have a discontinuity in the gradient of Ze here (and
in
the flow variables). In the numerics these large changes in the

Re: Changing reaction rate model

2008-07-28T20:01:13Z

Gary,
I can't seem to see the figures that you uploaded. Maybe there was an
error in uploading, as it shows the file sizes as 0 bytes
I notice that this happens when Zi is very close to 0. See attached ZiZe.
I think that it may be due to the 2 step integration in the SourceTerms
file. It's possible that the first step of the integration predicts that

Re: Changing reaction rate model

2008-07-26T09:52:51Z

Carlos, also, there is some differnece in the reaction zone
structure from the original ZND profile, near the end of the reaction
zone.
See figure P58 which compares the numerical structure with the ZND
profile... the numerical solution appears to have Ze tending to zero
asympotically as if ETA=1 rather than the slope discontinuity for

Re: Changing reaction rate model

2008-07-26T09:42:26Z

Carlos, OK I missed the equation set mistake. If you make the sign
correction
throughout PartialDerivatives, ReactiveEulerEquations AND SourceTerms
then your code more less works now, in that it propagates your
original
ZND profile.

Looking closely at the late time solution, there is a slight

Re: Changing reaction rate model

2008-07-25T18:18:46Z

Ah yes, that's right. I made a mistake there. I believe I made the
same mistake in the equation set

What do you mean by this? When I ran it, the state behind the reaction
zone was the same as the state that the ComputeZnd procedure produced
(P = 0.57005644 for both cases)

Carlos

Re: Changing reaction rate model

2008-07-25T16:49:02Z

Carlos, I believe you have the sign in front of the Q wrong in your
pressure routines in the solver.
You have
P=(G-1)*D*(i-Q*Z2)
giving
i=P/(G-1)/D+Q*Z2
for the internal energy
shoudln't it be
i=P/(G-1)/D-Q*Z2?
i.e. internal energy gets converted into heat
and hence i decreases as QZ2 increases?

Re: Changing reaction rate model

2008-07-25T15:05:04Z

Gary,
BCG is just a collection of code-generating routines which happen
to be arranged in a tree. When you run:
plugin amr_sol
ReactiveEulerEquations {
model = 3cb
BasicCodeGenerator {
scheme = roe'3cb
code = abc
BCG switches to amr_sol's code-generating tree then jumps
to the sub-tree used for the Reactive Euler equations. It then

amrita-ebook Google Group

Re: Changing reaction rate model

Re: Changing reaction rate model

Re: *** BDY LIST EXHAUSTED ***

*** BDY LIST EXHAUSTED ***

Re: Changing reaction rate model

Re: Changing reaction rate model

Re: Changing reaction rate model

Re: Changing reaction rate model

Re: Changing reaction rate model

Re: Changing reaction rate model

Re: Changing reaction rate model

Re: Changing reaction rate model

Re: Changing reaction rate model

Re: Changing reaction rate model

Re: Changing reaction rate model

Re: * BDY LIST EXHAUSTED *

* BDY LIST EXHAUSTED *