How is the input for COSIMA organised?
The input to COSIMA consists of 3 text blocks, to which data can be entered or pasted.The block "Parameter" serves to specify the simulation scenario (e.g. temperature, pressure, carrier gas, structural parameters, or lognormal parameters of the initial size distribution) and to provide information required for the controlling of the program run (e.g. numerical accuracy, output generation, etc.). Completion of this block is mandatory. However, on calling the COSIMA simulation page (Models 1  5), a full set of default values is delivered, many of which need not to be changed for typical applications. Furthermore, the tab "Input description" can be selected at any time during input for information on structure and content of the "Parameter" block. The block "Tabulated leakage function" may be used to enter a table of loss rates as function of time in order to account for aerosol losses due to leakage out of the control volume or for sampling losses during an experiment. The block "Tabulated size distribution(s)" allows to enter up to 30 arbitrary number size distributions with respect to the mobility equivalent diameter to specify the initial state of the particle system and additional source events, respectively. "Leakage" and "Size distribution(s)" block may be left empty.
Having the input completed, COSIMA may be invoked by selecting "Start simulation with current input data" and klicking the "Start" button. Choosing "Start simulation with reloaded default input data" instead completely reestablishes the default values. Please note that the latest input status is preserved unless default values are reinstalled or the browser session ist terminated.
Below, the structure of the 3 input blocks is outlined and the input is described in some detail. Each parameter has assigned a characteristic value that will appear as default input on initialising the COSIMA simulation page with Model 3 (CAST soot).
To further facilitate becoming familiar with COSIMA a spreadsheet to be used with Excel or compatible products is available for download. The first page called "annotated input" is also intended to visualise input structure and content. On moving the mouse cursor on top of a parameter name field an explanatory comment box will open. All input fields are editable. However, fields shaded in grey need not to be changed for the vast majority of realistic applications. On the second page called "input generator" the input data are compressed to a form suitable for copying and pasting into the "Parameter" block of the COSIMA simulation page.
[download spreadsheet]
The Parameter Block
Line 12  TEXT  
CAST soot default scenario 

This is text line 2  
Line 3  K2  NWRITE(1)  NWRITE(2)  NTHERM  LNOLK  LNOTE  NCARD  GRAVK  FUKS  COULOM 
200  100 
500 
1 
T  F  5 
T  T  F  
Line 4  VOL  FSED  FDIFF  FTHERM (1.. 
NTHERM) 

8.430E+07 
1.260E+05 
1.030E+06 
0.0 


Line 5  TT  P  GASIN  
295.00 
1.00 
'AIR'  
Line 6  DELD  DELT  CONS  FORM  FORMC  
0.5 
0.32 
350. 
1.0 
1.0 

Line 7  RMIN  RMAX  TIME  STIME  EPS  DT  ZLIM  GLIM  
1.00E06 
5.00E05 
28.50 
1620.00 
0.001 
0.001 
1.0E09 
1.0E19 

Line 8  PACKD  DF  RPRIM  FINTPL  RSTAR  COORDN  
0.70 
1.95 
1.50E06 
0.89 
0.85 
2.50 

Line 9  FNREAL  FNIMAG  WAVLEN  
1.0 
0.0 
450.0 

Line 10  NOUTP  
8 

Line 11  TOUTP(1..NOUTP)  
27.0 
40.8 
54.5 
87.5 
145.7 
282.9 
435.9 
780.5 

Line 12  NPHASE  NCOMP  
1 
1 

Line 13  SRATE(1,K)  FSCALE  CONTIN(1,K)  SDTAB(1,K)  RG(1,K)  SIGL(1,K)  RHOQ(1,K)  
2.921E+05 
1.0 
F  T  3.983E06 
0.458 
1.335 

Line 14  SRATE(2,K)  FSCALE  CONTIN(2,K)  SDTAB(2,K)  RG(2,K)  SIGL(2,K)  RHOQ(2,K)  TQ(K+1)  
0.000E+00 
1.0 
F  F  3.000E06 
0.400 
1.330 
3800.0 

Line 15  COMPNAM(I)  XCOMPI(1,K,I) 
XCOMPI(2,K,I)  
'CAST'  1.000  1.000  
Line 16  N  
1 

Line 17  TT(K)  TE(KK,K)  
0.0 
0.0 
Input description
Lines 1  2
TEXT: 2 lines of arbitrary text (maximum 2 * 72 characters)
Line 3
K2: number of size bins in sectional representation (maximum 400, typical 100  200)
NWRITE(1): number of simulation time steps after which a small printout is produced (window "Detailed overview", file out.txt)
NWRITE(2): number of time steps after which a comprehensive printout is produced (window "Detailed overview", file out.txt)
NTHERM: number of different surface temperatures to be considered in thermophoretic removal (up to 5)
LNOLK: T (.TRUE.): sampling or leakage losses accounted for (default)
F (.FALSE.): no leakage assumed. In this case input to the leakage box is ignored
LNOTE: T: thermophoretical deposition is accounted for; F: thermophoresis is neglected (default)
NCARD: number of time steps after which tabulated output is written to both window "Tabulated results" and file res.txt,
e.g. for plotting purposes or further processing
GRAVK: T: account for gravitational coagulation (default); F: neglect it
FUKS: T: transition regime interpolation applied for calculation of coagulation rates (default)
F: continuum regime expressions applied for coagulation rates of all particle sizes
COULOM: T: account for coulomb interactions in coagulation kernel; F: neglect them (default)
Line 4
VOL: container volume [cm^{3}]
FSED: surface area available for sedimentational deposition [cm^{2}]
FDIFF: surface area available for diffusional deposition [cm^{2}]
FTHERM(1 .. NTHERM): surface areas relevant for thermophoretic removal [cm^{2}], NTHERM entries
Line 5
TT: temperature [K]
P: pressure [bar]
GASIN: type of carrier gas ('AIR', 'NITROGEN', 'ARGON', 'HELIUM', or 'CO2'), 'AIR' is default
Line 6
DELD: diffusional boundary layer thickness parameter [cm]. Following Bunz and Dlugi (1991), the thickness DELTA of the
diffusional boundary layer is calculated as DELTA = DELD*DFC^0.274, wth DFC denoting the size dependent diffusion
coefficient of the particle.
DELT: thermophoretic boundary layer thickness [cm]
CONS: ratio of thermal conductivities of particles and carrier gas
FORM: dynamic shape factor, default value is 1
FORMC: coagulation shape facor, default value is 1
Line 7
RMIN: minimum (mass equivalent) particle radius to be considered in simulation [cm]
RMAX: maximum (mass equivalent) particle radius to be considered in simulation [cm]
TIME: total simulation tme [h]
STIME: simulation startup time [s]
EPS: accuracy parameter (typically between 0.01 and 0.0001), default value is 0.001
DT: initial simulation time step [s], default value is 0.001
ZLIM: lower limit for particle number concentration in size bins [cm^{3}]; if number concentration decreases below ZLIM for
a given bin, it is set to zero (default value: 1.0E9)
GLIM: lower limit for particle mass concentration in size bins [mg/m^{3}]; if mass concentration decreases below GLIM for a
given bin, its number concentration is set to zero (default value: 1.0E19)
Please note: The default values for EPS and DT are chosen conservatively and need not to be changed for typical applications. Decreasing EPS reduces the length of the integration time steps. Thus, when decreasing EPS substantially, NWRITE(1), NWRITE(2), and NCARD (cf. line 3) should be increased accordingly in order to avoid large amounts of output.
Line 8
PACKD: packing density (≤ 1), default value is 0.7 for fractal like particles. For compact species PACKD = 1.0
DF: volume fractal dimension. Since the formalism used in COSIMA assumes a spherically symmetric distribution
of primary particles in fractal like clusters, meaningful results will only be obtained with 1.5 < Df ≤ 3.
RPRIM: radius of primary particles [cm]
FINTPL: interpolation factor between gravitational coagulation efficiency limits (0 ≤ FINTPL ≤ 1), default value is 0.89
RSTAR: reduced bond length of (overlapping) primary particles (bondlength/2 RPRIM), RSTAR ≤ 1
COORDN: average number of neighbours connected to a primary particle (1  12), default value is 2.5
Please note: For compact spherical particles (PACKD = 1.0, DF = 3.0) FINTPL and RSTAR must both be set to 1.0. In this case the values for RPRIM and COORDN can be chosen arbitrarily.
Line 9
FNREAL: real part of refractive index
FNIMAG: imaginary part of refractive index
WAVLEN: wavelength of incident light [nm]
Line 10
NOUTP: number of points in time where the actual mobility size distribution is written to window "Size distributions"
and to file dist.txt; 1  300.
Line 11
TOUTP(1 .. NOUTP): times (in min) where the actual mobility size distribution is written to window "Size distributions"
and to file dist.txt; NOUTP entries. At least one point in time must be specified, e.g. the starting point.
Please note:
Output is generated at the end of the first integration time step passing a specified TOUTP. Depending on the magnitude of the integration steps this results in a slight delay with respect to to the target times. If required, the matching can be improved be reducing EPS (cf. line 7).
Line 12
NPHASE: number of release phases of aerosol source (1  30), default value is 1
NCOMP: number of components in particle phase (1  50), default value is 1
Line 13
SRATE(1,K): source rate for release mode 1 in release phase K,
[1/cm^{3}] for instantaneous source or [1/cm^{3} s] for continuous source)
FSCALE: arbitrary scaling factor for release rate
CONTIN(1,K): T for continuous source; F for instantaneous source
SDTAB(1,K): F: lognormal size distribution assumed.
T: a tabulated mobility related number distribution is used to calculate the size distribution of the released particles;
in this case RG(1,K) and SIGL(1,K) are dummy input
RG(1,K): mean geometric (mobility equivalent) radius of lognormal size distribution [cm] for mode 1
SIGL(1,K): natural logarithm of standard deviation of lognormal size distribution for mode 1
RHOQ(1,K): bulk material density of source particles in mode 1 [g/cm^{3}]
Line 14
SRATE(2,K): source rate for release mode 2 in release phase K,
[1/cm^{3}] for instantaneous source or [1/cm^{3} s] for continuous source
FSCALE: arbitrary scaling factor for release rate
CONTIN(2,K): T for continuous source; F for instantaneous source
SDTAB(2,K): F: lognormal size distribution assumed.
T: a tabulated mobility related number distribution is used to calculate the size distribution of the released particles;
in this case RG(2,K) and SIGL(2,K) are dummy input
RG(2,K): mean geometric (mobility equivalent) radius of lognormal size distribution [cm] for mode 2
SIGL(2,K): natural logarithm of standard deviation of lognormal size distribution for mode 2
RHOQ(2,K): bulk material density of source particles in mode 2 [g/cm^{3}]
TQ(K+1): end of current and start of following release phase [s].
1st release phase starts at beginning of simulation (STIME, c.f. line 7)
Please note:
To introduce a release free time interval between two source events, just define a release phase with zero source rates or
FSCALE = 0.0 in both release modes.
Line 15
COMPNAM(I): name of Ith component (maximum length: 8 characters)
XCOMPI(1,K,I): mass fraction of component I in mode 1 of release phase K ( ≤ 1.0)
XCOMPI(2,K,I): mass fraction of component I in mode 2 of release phase K ( ≤ 1.0)
Please note:
If more than one release phase is specified in line 12 (NPH > 1), the input to lines 13  15 has to be provided sequently for all phases
Line 16
N: number of base points for normalised temperature differences ΔT/T as function of time
Line 17
TT(K): times [s] for which normalised temperature differences are available (N table entries)
TE(KK,K): normalised temperature differences at time TT(K) (NTHERM entries, cf. line 3 )
The Leakage Block
If no leakage is specified (LNOLK = F, cf. line 3), this input will be ignored. Alternatively, an empty block may be provided.Line 1 
NLEAK  
4 

Line 2  end 
TL(K)  RL(K) 
0.0 
45.3 

... 
... 

... 
... 

94410.0 
43.6 
Input description
Line 1
NLEAK: number of points in time where leakage rate function is supplied (1  200)
Lines 2  end
TL(K): points in time for leakage rate function specification [s]
RL(K): leakage rate (volume%/day) at TL(K)
Please note:
The leakage rate is interpolated linearly between the table entries. If the leakage table consists of only one data pair (NLEAK = 1), a constant aerosol loss rate is assumed after the specified time.
The Size Distribution(s) Block
Typically, this data block is used to initialise the simulation with a measured mobility related number size distribution (e.g. from a SMPS).
If all aerosol sources are assumed to exhibit lognormal size distributions (SDTAB(1,K) = F and SDTAB(2,K) = F for all release phases, cf. lines 13 and 14 in "Parameter" block), this input will be ignored. Alternatively, an empty block may be provided.
Line 1 
NPH 
NMODE  AGAIN 
1 
1 
F  
Line 2 
N  LACCUM  
114 
F  
Line 3  end 
D(K)  RELNUM(K)  
14.10 
0.00 

... 
... 

68.50 
8803.37 

... 
... 

820.00 
3.84 
This input has to be repeated sequently for every release phase and mode for which a tabulated mobility size distribution needs to be specified (at most 30 phases comprising of 2 modes)
Input description
Line 1
NPH: index of release phaseNMODE: index of release mode (1 or 2)
AGAIN: T: number size distribution read in last will be reused for release mode definded by NPH and NMODE
F: new number size distribution will be read in
Line 2
N: number of grid points in number size distribution table ( ≤ 400), should be smaller than chosen number of size bins
in sectional representation (K2, c.f. line 3)
LACCUM: interpolation of source input to simulation size grid using number distribution (F, default) or accumulated
number distribution (T), alternatively.
Line 3  end
D(K): mobility equivalent particle diameter [nm]
RELNUM(K): number concentration accumulated in size class K (in arbitrary units).
Absolute number concentration is determined from SRATE(I,K) and FSCALE, c.f. lines 13 and 14 in "Parameter" block.