Maintained by ppernot
Once the simulation results are available (or a part of them), they can be analyzed to plot kinetic traces, mass-spectra or perform sensitivity analysis.
The analysis of reaction fluxes is done in the Fluxes module.
Load MC results has first to be pressed to read the results files.
A short summary is presented below the button when the operation
is completed.News
Outputs/MoleFrac_tFinal.csvA plot of the time-dependent mole fractions of the chemical species and their uncertainty bands. The plot is zoomable (select the area and double-click; double-click again to cancel the zoom).
Show error bands to display the 95% uncertainty bands
resulting from the Monte-Carlo runs.
Fixed colors to attribute a fixed color to each species,
independently of the set of species to be displayed.
Draw PPM scale to add a series of horizontal lines showing
the ppm, ppb and ppt mole fraction levels.
The following controls enable to select the classes of
species to be displayed. Classes Cx refer to the number
of heavy atoms in a species.
A plot of (pseudo) mass spectra. The MS for neutrals and ions are presented separately. For neutral species, these are not electronic impact (EI) fragmentation mass spectra, but abundances at the unfragmented species mass.
Show error bars to display the 95% uncertainty bars
resulting from the Monte-Carlo runs.
Draw PPM scale to add a series of horizontal lines showing
the ppm, ppb and ppt mole fraction levels.
The following controls enable to select the classes of
species to be displayed. Classes Cx refer to the number
of heavy atoms in a species.
Log sampling time enables to define the time-snapshot
to be used to draw the MS.
MS Amplitude Range controls the range of the y-axis
of the mass-spectra.
Adjust bar width enables to tune the width of the MS bars.
Size of names controls the peaks labels size (0 = no labels)
This is where one performs sensitivity analysis (SA) to identify key reactions. In the Monte-Carlo framework, key reactions are those which have the largest impact on the prediction uncertainty of a quantity of interest. This will run only if MC runs have been performed.
Sensitivity indices presents a choice of sensitivity index (SI)
quantifying the dependence between the concentrations and reaction
rates over the Monte-Carlo sample:
Rank correl.: Spearman correlation coefficient.
dCorr: distance correlation t-test of multivariate
and functional independence (Szekely, G.J. and Rizzo, M.L. (2013).
The distance correlation t-test of independence in high dimension.
Journal of Multivariate Analysis, Volume 117, pp. 193-213.
doi ).
dHSIC: d-variable Hilbert Schmidt independence criterion
(Gretton, A., K. Fukumizu, C. H. Teo, L. Song, B. Scholkopf
and A. J. Smola (2007). A kernel statistical test of independence.
In Advances in Neural Information Processing Systems, pp. 585-592.
doi ).
Note: the dCorr and dHSIC features are experimental…
Choose species: target species for the SA.
If no species is specified, the sensitivity indices are estimated
from the sum of the “correlation” matrix over all species.
For Rank correl., the square of the correlation matrix is used.
Warning: the computation of dCorr and dHSIC indices might be very
time consuming, especially if no target species is specified.
Run SA: start computation of sensitivity indices
Plot type:
Bar Plot shows a barplot of the Top 20 reactions
with the larges sensitivity index, whether a target species
is specified or not.
Scatterplot is active only if a target species is specified.
It draws scatterplots between the MC concentrations of the
target species and the 20 MC reaction rates with largest SI.
The value of the SI is reported in red.
This module checks in MC samples if final concentrations, photorates or reaction rates present numerical exceptions (zero, infinity, NaN…).
The following statistics are reported for each problematic variable:
Var=0: a logical indicaticating if the variance over the MC sample is null
Nzero: fraction of the MC samples with a null value
Ninf: fraction of the MC samples with infinite value
This module plots statistics about the IRKC implicit-explicit integrator (cf. https://github.com/ppernot/Reactor):
NFE : number of evaluations of F_E
NFI : number of evaluations of F_I
NACCPT : number of accepted steps
NREJCT : number of rejected steps
NFESIG : number of evaluations of F_E used
in estimating the spectral radius
MAXM : maximum number of stages used
This module plots statistics about the IRKC implicit-explicit integrator (cf. https://github.com/ppernot/Reactor):
SPRAD : spectral radius of the jacobian of F_E