myprojects-hg-reborn

Fork

I modified kernel-test.py. Now the code considers several differen cases
when going to different limits (continuum vs free kernel) to test the
Fuchs approximation.

I added the code air-mean-free-path.py which works out the mean-free
path of air molecules. Formulas for interpolations taken from Comsol and
literature.

I modified lognormal_binning.R. Now the code works out a similar test to
rebuild the initial state and plot it against the experimental data (as
it used to do for the final state. The plotted initial state is without
any correction to the overall particle number in order to ease the
comparison).

I polished a bit kernel-test.py and added a second (equivalent)
calculation of the continuum kernel including a correction for the slip.

I modified smoluchowski2.py by adding Fuchs interpolation formula both
including and excluding a slip correction.

I added the new code kernel-test.py which implements Fuchs interpolation
formula as in the book by Seinfeld and Pandis, but without the typo. In
the case of the calculation without including the slip, in the limits
for large and small particles the kernel tends to the continuum and free
molecular limit, respectively. I also added the slip correction to the
kernel (see expression for the kernel in the equation for D_i on page
661 of the book of Seinfield and Pandis).

I modified lognormal_binning.R so that now it calculates mean volumes
and mean diameters in each bin in a consistent way even for fractal
dimension D_f!=3.

I added to the code read_results_smoluchowsky_single_simulation.R the
capability to make a 3D plot of the number-size distribution as a
function of time and log(D_p), where D_p is the particle diameter.

I corrected the a wrong printout in
read_results_smoluchowsky_single_simulation.R.

I modified read_results_smoluchowsky_single_simulation.R. It now plots
some extra quantities and it reads directly some parameters (like the
fractal dimension or the size of the primary particles).

I fixed some bug in lognormal_binning.R. Now I correctly generate a set
of diameters in [D_min,D_max] which are log spaced.

Some modifications to smoluchowski2.py which now reads directly e.g. the
fractal dimension as saved from the R code lognormal.R.

I modified lognormal_binning.R. It now saves better the data to pass to
the Python code to solve Smoluchowsky equation (this code has been
modified as well).

I modified lognormal_binning.R. Now it properly treats the fractal
dimension D_f and correctly calculated the volume of solid in each bin.

I modified smoluchowski2.py. Now the kernel for the free molecule regime
is correct and it reads the size of the fundamental monomer (it plays a
role only in case D_f != 3).

I modified lognormal_binning.R. Now the code also calculates the volume
of the initial and final state and check that it is preserved.

I modified lognormal_binning.R by correcting a small bug leading to
saving the wrong (badly scaled) volume grid ("V_list") when a uniform
binning was chosen.

I changed the code read_results_smoluchowsky_single_simulation.R. The
code does not do anything with the value of the fractal dimension right
now, which is the reason why I removed it.

I modified the code smoluchowski2.py. I simply removed a trivial typo.

I modified again the codes read_results_smoluchowsky_single_simulation.R
and lognormal_binning.R. Now both of them should be correct and be able
to treat correctly the diamter and the volume, both in case of a
log-normal and uniform grid.

I re-modified the code R-codes/lognormal_binning.R. I think I solved its
previous problems. Now the code treats the diameters (which are
completely independent on the fractal dimension) as the truly
independent variable and everything else is built upon them. The
conversion factors (nm--->m and nm^3--->m^3) should now have been fixed.

I modified the code R-codes/lognormal_binning.R. Probably the previous
commitment I made was a mistake (somehow the code had a bug and was
misbehaving for fractal dimension D_f != 3). Now this appears to have
been fixed, but better to keep a watchful eye on this code.

I modified the code
R-codes/read_results_smoluchowsky_single_simulation.R. Now it should
work also with a fractal dimension, but I had to fix some conversion
factor (mainly, when going from nm to meters, I had to perform the
replacement: 10^9--->10^(27/D_f), where D_f is the fractal dimension of
the agglomerate).

I modified the code R-codes/lognormal_binning.R. I fixed some bugs which
appeared when I was using the fractal dimension, in particular
everything was due to some conversion factors I was using (mainly I had
to perform the replacement 10^9--->10^(27/D_f), where D_f is the fractal
dimension of the agglomerate). Now it seems to work but CHECK very
carefullty

I modified the code
R-codes/read_results_smoluchowsky_single_simulation.R. The code is now
able to deal with a fractal dimension and a small bug in a plot got
corrected. Also the code to generate the binning is able to deal with
the fractal dimension by now.

I modified the code R-codes/lognormal_binning.R. Now the code is also
able to create a set of log-normally distributed bins (this turns out
to be very convenient if we want to sample a log-normal distribution).

I modified the code Python-codes/smoluchowski2.py. Now the code is able
to solve smoluchowsky equation also for a non-uniform binning (mainly a
logarithmic binning at this point) and uses the expression for
smoluchowsky equation taken from the paper by Garrick. It is worth
pointing out that some old functions have been removed (they were no
longer needed since the optimized versions are used) and there are still
3 implementations of the splitting coefficients (only one of them [the
one with _garrick] is needed, the other two were simply a test to make
sure I could use properly the "where" function from scipy).

I modified the code
R-codes/read_results_smoluchowsky_single_simulation.R. Nothing
substantial, but now it can work out the shape function of the
self-similar distribution.

slight (unsubstantial) modification of the code
R-codes/lognormal_binning.R

I modified the code
R-codes/read_results_smoluchowsky_single_simulation.R. It now plots
automatically the psi_eta at 4 different times during the evolution and
it does not work plot psi_eta vs the diameter.