You're reading the documentation for a development version. For the latest released version, please have a look at v0.8.16.

Transform

CADET-Match has many variable transforms available. The most common ones will be covered first. There are a few additional ones available in the code that are retained for backwards compatibility that will not be covered and their usage is not recomended.

Each parameter for parameter estimation or error modeling must be listed in the parameters list and a transform selected.

{
    "parameters": [
        {
        },
        {
        }
    ],
}

Indexing

An index is used when you want to directly index into the arrays to modify, this is very rarely used and mostly comes up with the MultiState SMA rates between states. All transforms use the same method for indexing.

You either need to specific component and bound or specify index. Indexing starts at 0. A value of -1 is used to indicate a value that is not component specific (i.e. a scalar).

Scalar value:index=-1 Index at position n: index=n

Component 0 Bound state 0 would be component=0 bound=0.

Component 1 but independent of bound state would be component=0 bound=-1.

Independent of component or bound state would be component=-1 bound=-1.

Auto

Auto is the most common transform. This transform will convert from the upper and lower boundary to a range of 0 to 1 for the search algorithm. It will automatically switch between a linear and a log transform as needed for search performance.

Key	Values	Default	Required	Description
min	Float	None	Yes	minimum allowed value
max	Float	None	Yes	maximum allowed value
location	String	None	Yes	path inside the simulation to the variable
maxFactor	Float	1000	No	if max/min > maxFactor a log transform is used.
component	Integer	None	Yes*	Index of the component used (starts at 0)
bound	Integer	None	Yes*	Index of the bound state (starts at 0)
index	Integer	None	Yes*	Index into the location (starts at 0) Either component + bound or index must be specificed not both.
experiments	String	None	No	If provided this variable is only used for the named experiments

{
    "transform": "auto",
    "component": 2,
    "bound": 0,
    "location": "/input/model/unit_001/adsorption/SMA_NU",
    "min": 1,
    "max": 50
}

Auto Inverse

Auto-inverse works like auto except it transforms used 1/variable. This transform was designed to deal with coupling of film diffusion and pore diffusion and improves optimization.

Key	Values	Default	Required	Description
min	Float	None	Yes	minimum allowed value
max	Float	None	Yes	maximum allowed value
location	String	None	Yes	path inside the simulation to the variable
maxFactor	Float	1000	No	if max/min > maxFactor a log transform is used.
component	Integer	None	Yes*	Index of the component used (starts at 0)
bound	Integer	None	Yes*	Index of the bound state (starts at 0)
index	Integer	None	Yes*	Index into the location (starts at 0) Either component + bound or index must be specificed not both.
experiments	String	None	No	If provided this variable is only used for the named experiments

{
    "transform": "auto_inverse",
    "component": -1,
    "bound": -1,
    "location": "/input/model/unit_001/FILM_DIFFUSION",
    "min": 1e-9,
    "max": 1e-5
}

Auto kEQ

This transforms convert from kA and kD to kA and kEQ with all the other properties of auto. In reality kA and kD are coupled and this allows the search algorithm to see the coupling. There are also some fits where kA and kD are fast enough that a system is effectively in rapid equilibrium. Without this transform a large number of kA and kD values will be found with equally good results. With this transform kEQ will have a definite value and there will be a large range of kA values which provides more understanding for the problem.

Key	Values	Default	Required	Description
minKA	Float	None	Yes	minimum allowed value
maxKA	Float	None	Yes	maximum allowed value
minKEQ	Float	None	Yes	minimum allowed value
maxKEQ	Float	None	Yes	maximum allowed value
location	String	None	Yes	paths inside the simulation kA and kD
maxFactor	Float	1000	No	if max/min > maxFactor a log transform is used.
component	Integer	None	Yes*	Index of the component used (starts at 0)
bound	Integer	None	Yes*	Index of the bound state (starts at 0)
index	Integer	None	Yes*	Index into the location (starts at 0) Either component + bound or index must be specificed not both.
experiments	String	None	No	If provided this variable is only used for the named experiments

{
    "transform": "auto_keq",
    "component": 0,
    "bound": 0,
    "location": [
        "/input/model/unit_001/adsorption/LIN_KA",
        "/input/model/unit_001/adsorption/LIN_KD"
    ],
    "minKA": 1e-8,
    "maxKA": 1e8,
    "minKEQ": 1e-4,
    "maxKEQ": 1e4
}

Norm Add

This transform allows another parameter to be read and a fixed or variable value added to it and assigned to a second variable. For example if you are optimizing the charge nu for SMA with a few different charge variants you may not know all the charge variants but you know they are all close together and so you can estimate one and then use norm_add for the others to require they are close.

Key	Values	Default	Required	Description
min	Float	None	Yes	minimum allowed value
max	Float	None	Yes	maximum allowed value
locationFrom	String	None	Yes	paths inside the simulation kA and kD
componentFrom	Integer	None	Yes*	Index of the component used (starts at 0)
boundFrom	Integer	None	Yes*	Index of the bound state (starts at 0)
indexFrom	Integer	None	Yes*	Index into the location (starts at 0) Either component + bound or index must be specificed not both.
locationTo	String	None	Yes	paths inside the simulation kA and kD
componentTo	Integer	None	Yes*	Index of the component used (starts at 0)
boundTo	Integer	None	Yes*	Index of the bound state (starts at 0)
indexTo	Integer	None	Yes*	Index into the location (starts at 0) Either component + bound or index must be specificed not both.
experiments	String	None	No	If provided this variable is only used for the named experiments

{
    "transform": "norm_add",
    "locationFrom": "/input/model/unit_001/COL_POROSITY",
    "componentFrom": -1,
    "boundFrom": -1,
    "locationTo": "/input/model/unit_001/PAR_POROSITY",
    "componentTo": -1,
    "boundTo": -1,
    "min": -0.1,
    "max": 0.1
}

Norm Mult

This transform allows another parameter to be read and a fixed or variable value multiplied to it and assigned to a second variable. For instance if you are estimating the shielding factor sigma for a monomer and also need to estimate it for a dimer you can estimated sigma for the monomer normally and then specify that the dimer is approximately twice as large.

Key	Values	Default	Required	Description
min	Float	None	Yes	minimum allowed value
max	Float	None	Yes	maximum allowed value
locationFrom	String	None	Yes	paths inside the simulation kA and kD
componentFrom	Integer	None	Yes*	Index of the component used (starts at 0)
boundFrom	Integer	None	Yes*	Index of the bound state (starts at 0)
indexFrom	Integer	None	Yes*	Index into the location (starts at 0) Either component + bound or index must be specificed not both.
locationTo	String	None	Yes	paths inside the simulation kA and kD
componentTo	Integer	None	Yes*	Index of the component used (starts at 0)
boundTo	Integer	None	Yes*	Index of the bound state (starts at 0)
indexTo	Integer	None	Yes*	Index into the location (starts at 0) Either component + bound or index must be specificed not both.
experiments	String	None	No	If provided this variable is only used for the named experiments

{
        "transform": "norm_add",
        "locationFrom": "/input/model/unit_001/COL_POROSITY",
        "componentFrom": -1,
        "boundFrom": -1,
        "locationTo": "/input/model/unit_001/PAR_POROSITY",
        "componentTo": -1,
        "boundTo": -1,
        "min": 0.8,
        "max": 1.5
}

Set Value

This transform copies a value from another estimated value. One of the common usage cases is when estimating the axial dispersion of the tubing. It can be a good assumption that the axial dispersion is the same in the tubing leading to the column and the tubing leaving it so with this one of them is estimated and the value copied to the other one so fewer values need to be estimated.

Key	Values	Default	Required	Description
locationFrom	String	None	Yes	paths inside the simulation kA and kD
componentFrom	Integer	None	Yes*	Index of the component used (starts at 0)
boundFrom	Integer	None	Yes*	Index of the bound state (starts at 0)
indexFrom	Integer	None	Yes*	Index into the location (starts at 0) Either component + bound or index must be specificed not both.
locationTo	String	None	Yes	paths inside the simulation kA and kD
componentTo	Integer	None	Yes*	Index of the component used (starts at 0)
boundTo	Integer	None	Yes*	Index of the bound state (starts at 0)
indexTo	Integer	None	Yes*	Index into the location (starts at 0) Either component + bound or index must be specificed not both.
experiments	String	None	No	If provided this variable is only used for the named experiments

{
        "transform": "set_value",
        "locationFrom": "/input/model/unit_000/sec_000/CONST_COEFF",
        "componentFrom": 0,
        "boundFrom": 0,
        "locationTo": "/input/model/unit_000/sec_000/CONST_COEFF",
        "componentTo": 1,
        "boundTo": 0
}

Sum

This transform reads two values and assigns it to a 3rd value. This was created for a situation where the volume of two CSTRs where estimated and a 3rd CSTR needed to have a volume equal to the sum of the first two.

Key	Values	Default	Required	Description
location1	String	None	Yes	paths inside the simulation kA and kD
component1	Integer	None	Yes*	Index of the component used (starts at 0)
bound1	Integer	None	Yes*	Index of the bound state (starts at 0)
index1	Integer	None	Yes*	Index into the location (starts at 0) Either component + bound or index must be specificed not both.
location2	String	None	Yes	paths inside the simulation kA and kD
component2	Integer	None	Yes*	Index of the component used (starts at 0)
bound2	Integer	None	Yes*	Index of the bound state (starts at 0)
index2	Integer	None	Yes*	Index into the location (starts at 0) Either component + bound or index must be specificed not both.
locationSum	String	None	Yes	paths inside the simulation kA and kD
componentSum	Integer	None	Yes*	Index of the component used (starts at 0)
boundSum	Integer	None	Yes*	Index of the bound state (starts at 0)
indexSum	Integer	None	Yes*	Index into the location (starts at 0) Either component + bound or index must be specificed not both.
experiments	String	None	No	If provided this variable is only used for the named experiments

{
        "transform": "sum",
        "location1": "/input/model/unit_001/INIT_VOLUME",
        "component1": -1,
        "bound1": -1,
        "location2": "/input/model/unit_003/INIT_VOLUME",
        "component2": -1,
        "bound2": -1,
        "locationSum": "/input/model/unit_004/INIT_VOLUME",
        "componentSum": -1,
        "boundSum": -1
}

Norm Diameter

CADET uses the cross sectional area of the column and tubing and measuring this precisely can be difficult. It is often much simpler to measure the diameter and provide a small search range and then allow this transform to convert that to the area. This assumes circular tubing and uses Area = pi*d^2/4.

Key	Values	Default	Required	Description
min	Float	None	Yes	minimum allowed value
max	Float	None	Yes	maximum allowed value
location	String	None	Yes	paths inside the simulation kA and kD
component	Integer	None	Yes*	Index of the component used (starts at 0)
bound	Integer	None	Yes*	Index of the bound state (starts at 0)
index	Integer	None	Yes*	Index into the location (starts at 0) Either component + bound or index must be specificed not both.
experiments	String	None	No	If provided this variable is only used for the named experiments

{
        "transform": "norm_diameter",
        "location": "/input/model/unit_001/CROSS_SECTION_AREA",
        "min": 0.001,
        "max": 0.1,
        "component": -1,
        "bound": -1
}

Norm Volume Length

When estimating the size of a Disperive Plug Flow Reactor needed to model a piece of tubing it is normal to estimate the dispersion, area and length. This works but can be problematic to estimate and get a realistic estimate due to the degrees of freedom. Finding the volume of the tube and the length of the tube is much easier to do accurately and this makes it a much better transform to work with.

Key	Values	Default	Required	Description
area_location	String	None	Yes	path to cross_section_area
length_location	String	None	Yes	path to col_length
component	Integer	None	Yes*	Index of the component used (starts at 0)
bound	Integer	None	Yes*	Index of the bound state (starts at 0)
index	Integer	None	Yes*	Index into the location (starts at 0) Either component + bound or index must be specificed not both.
minVolume	Float	None	Yes	minimum allowed value
maxVolume	Float	None	Yes	minimum allowed value
minLength	Float	None	Yes	minimum allowed value
maxLength	Float	None	Yes	minimum allowed value
experiments	String	None	No	If provided this variable is only used for the named experiments

{
        "transform": "norm_volume_length",
        "area_location": "/input/model/unit_001/CROSS_SECTION_AREA",
        "length_location": "/input/model/unit_001/COL_LENGTH",
        "minVolume": 1e-06,
        "maxVolume": 0.0001,
        "minLength": 0.1,
        "maxLength": 0.3,
        "component": -1,
        "bound": -1
}

Norm Volume Area

This transform works like the volume length transform except it uses volume an area and should only be used if it is easier to estimate the cross sectional area of the tubing than its length.

Key	Values	Default	Required	Description
area_location	String	None	Yes	path to cross_section_area
length_location	String	None	Yes	path to col_length
component	Integer	None	Yes*	Index of the component used (starts at 0)
bound	Integer	None	Yes*	Index of the bound state (starts at 0)
index	Integer	None	Yes*	Index into the location (starts at 0) Either component + bound or index must be specificed not both.
minVolume	Float	None	Yes	minimum allowed value
maxVolume	Float	None	Yes	minimum allowed value
minArea	Float	None	Yes	minimum allowed value
maxArea	Float	None	Yes	minimum allowed value
experiments	String	None	No	If provided this variable is only used for the named experiments

{
        "transform": "norm_volume_area",
        "area_location": "/input/model/unit_001/CROSS_SECTION_AREA",
        "length_location": "/input/model/unit_001/COL_LENGTH",
        "minVolume": 1e-06,
        "maxVolume": 0.0001,
        "minArea": 1e-05,
        "maxArea": 0.001,
        "component": -1,
        "bound": -1
}