The properties of materials which are relevant to flow simulations are:
The following table lists the properties used in PHOENICS, with their PIL-variable names and their SI units.
| PIL variable | SI units | nature |
| RHO1 | kg/m**3 | phase-1 density |
| DRH1DP | m**2/Newton | proportionate change of RHO1 with pressure |
| RHO2 | kg/m**3 | phase-2 density |
| DRH2DP | m**2/Newton | proportionate change of RHO2 with pressure |
| ENUL | m**2/s | kinematic laminar (reference) viscosity |
| ENUT | m**2/s | kinematic turbulent contribution to effective viscosity |
| PRNDTL(indvar) | dimensionless,if >0 | Prandtl or Schmidt number |
| PRNDTL(indvar) | watts/(m*degC), if <0 | and if indvar is enthalpy or temperature, thermal conductivity |
| PRNDTL(indvar) | kg/m*s, if <0 | and if indvar is not enthalpy or temperature, exchange coefficient |
| PRT(indvar) | dimensionless | turbulent contribution to the effective Prandtl or Schmidt number |
| PHINT(indvar) | according to indvar | equilibrium interface value for phase 1 |
| PHINT(indvar) | according to indvar | equilibrium interface value for phase 2 |
| TEM1 or TMP1 | degCelsius | phase-1 temperature |
| TEM2 or TMP2 | degCelsius | phase-2 temperature |
| EL1 | m | phase-1 turbulence length |
| EL2 | m | phase-2 turbulence length |
| CP1 | joule/(kg*degC) | constant-pressure specific heat of phase 1 |
| CP2 | joule/(kg*degC) | constant-pressure specific heat of phase 2 |
| DVO1DT | 1/degC | proportionate change of first-phase specific volume (i.e 1/RHO1) with temperature |
| DVO2DT | 1/degC | proportionate change of second-phase specific volume (i.e. 1/RHO2) with temperature |
| EMISS | 1/m | absorptivity = proportion of radiation which is absorbed per unit length |
| SCATT | 1/m | proportion of radiation which is scattered per unit length |
| CFIPS | Newton*s/m**4 | momentum-transfer rate from one phase to another per unit volume and per unit of velocity difference |
| CMDOT | kg*s/m**4 | mass-transfer rate per unit volume and per unit of velocity difference |
| CINT(indvar) | dimensionless | ratio of exchange coefficient to inter-phase friction factor for phase-1 side of interface |
| CINT(indvar) | dimensionless | ratio of exchange coefficient to inter-phase friction factor for phase-2 side of interface |
| CVM | dimensionless | virtual-mass coefficient for two-phase flow. |
The same properties are listed also at the top of the open-source Fortran file gxprutil.for, which lists in addition:
the indices
by which they are referred to in EARTH and GROUND coding and, the relevant cases, in the PROPS file.
Then, if PHOENICS recognises the names, it will evaluate and use
the properties which prevail at each location from the
corresponding formulae;
and also
(when the formulae make reference to them) from the prevailing:
temperature, pressure and concentrations.
Method b is the more flexible; for it allows arbitrary (combinations of) properties to be investigated, whether or not materials possessing them actually exist.
Both ways can be employed during the course of the same simulation. For example, some parts of the domain can be occupied by named materials, while other parts are occupied by materials of which the properties are set individually.
The material-property information is stored in PHOENICS in two places, namely:
Both are ASCII files which may be edited by users who wish to supply additional entries.
In the present article, attention will be focussed on what the Q1 should contain, no matter which method is used. However, a full account of the setting of properties by the menu-interactive method can be found in the appropriate section of TR 326.
SETPRPS(argument1, argument2)
where:
For example:
SETPRPS(1,0)
dictates that the properties will be those of atmospheric air, because that is the significance of IMAT=0 .
Likewise,
SETPRPS(2,67)
would set the second-phase fluid to be water with the properties pertaining to 20 degrees Celsius.
Then the Q1EAR file, which is created at the end of a SATELLITE run records in a structured manner what information (including defaults) will be sent to the solver module, EARTH, contains the lines:
Group 9. Properties RHO1 = 1.189000E+00 RHO2 = 9.982300E+02 DVO2DT = 1.180000E-04 ;DRH2DP = 0.000000E+00 ENUL = 1.544000E-05 CP1 = 1.005000E+03 ;CP2 = 4.181800E+03and the EARDAT file, which conveys the same information in more condensed form, contains the lines:
DOMAIN PHASE_1_MAT I 0 DOMAIN PHASE_2_MAT I 67which are reflected in the RESULT file as:
Group 19. EARTH Calls To GROUND Station USEGRD = T ;USEGRX = T SPEDAT(SET,DOMAIN,PHASE_1_MAT,I,0) SPEDAT(SET,DOMAIN,PHASE_2_MAT,I,67)These lines signify that, for the whole domain of simulation, the first-phase material has the properties associated with material 0 of the PROPS file. while the second-phase material has the properties of material 67.
It should be remarked the, since version 3.4.2, Earth no longer makes specific use of the domain-fluid concept, but picks up all the information which it needs from other EARDAT items.
(b) If the newer In-Form-based selection-by-name method is employed, the following lines could be placed in the Q1-file in order to select, for example, mercury as the phase-1 fluid:
fluid_name=mercury
load(089)
Then the Q1EAR, EARDAT and RESULT files would all contain copies of the formulae which EARTH will use for the computation of density. specific heat, viscosity and thermal conductivity.
The EARDAT version is:
PROPERTY RHO1 C=POL3(TEM1&14.293&-2.68226&5.3957$ PROPERTY RHO1 CE-4&-3.16674E-7) PROPERTY ENUL C=POL3(TEM1&5.47854&-.02372&4.3529$ PROPERTY ENUL C9E-5&-2.79475E-8)/(-10110) PROPERTY CP1 C=POL3(TEM1&159.54&-.10108&1.23163$ PROPERTY CP1 CE-4&-3.60116E-8) STORED COND C=POL3(TEM1&3.90003&.01799&-8.2070$ STORED COND C1E-6&1.52734E-9) PROPERTY PRNDTL(TEM1) C=COND/-10110
Here the 'POL3' and 'TEM1' are clues indicating that third-order polynomials are to be used for each of RHO1, ENUL, CP1 and COND, these being the formulae which are to be found in the loaded input-library case 089.
In-Form's case 089 refers only to phase-1 fluids. The just-described lines will therefore dictate that mercury is the first-phase fluid.
Of course it can easily be modified to allow phase-2 fluids to be selected.
Input-Library cases which illustrate this mode of property setting are 761 and 762.
(c) The phrase "one material fills the whole domain" does not, it should be mentioned, preclude the presence of "blocked-off" regions, which the material does not occupy, provided that whatever is contained in those regions is without influence on the phenomena being simulated other than, perhaps, to impose the no-slip condition at their boundaries.
Such regions are indicated by possession of PRPS values (see below) equal to:
It frequently occurs that PHOENICS is required to simulate flow and heat transfer in circumstances in which solid bodies are present. Often these solids interact thermally with the fluids.
Moreover, different parts of the domain may be occupied by different fluids, as when a glass bottle containing hot water is cooled by contact with external air.
This requirement is met by assigning different IMAT values to the spaces which each material occupies. Specifically:-
Thus,
FIINIT(PRPS) = 67
would dictate that water (i.e. material 67) would occupy the whole of space.
patch(block,inival,nx/4-1,3*nx/4,ny/4-1,3*ny/4,nz/4-1,3*nz/4,1,1) coval(block,prps,fixval,111.0)the central part of the grid would be occupied by a block of steel; for that is the material for which IMAT = 111.
Example: core library case 922
This case does indeed illustrate what happens when a heated steel block is suspended in a bath of cooling water.
The following three images show:
It happens that case 922 is a variant of core library case 921, inspection of which shows that the STORE(PRPS) and FIINIT(PRPS) settings have actually been effected by the use of PIL macros.
The lines in 921.htm which set the materials, and therefore their properties, are simply:
#use_props :fluid:=water20 INIT(SOLID,PRPS,0.0,steel)in which:
Users are of course free to edit any of these macros, or create new ones, for their personal convenience.
3.1.1 Setting uniform properties
When the properties are uniform, i.e. have the same magnitudes
everywhere
(except in blocked regions), all that is needed is for the Q1 to
contain statements of the kind:
PIL_name_of_property = value of property.
Examples are:
RHO1 = 1.189
CP1 = 1000.0
ENUL = 1.E-5
CFIPS = 100.0
CMDOT = 0.01
It is of course the user's responsibility to choose the values which meet his or her needs, and to maintain consistency of units.
The values which appear in the Q1 file are echoed in the Q1EAR, in EARDAT and in RESULT.
3.1.2 Setting properties which vary in accordance
with formulae coded in GXname files
Properties which vary, for example as functions of temperature, pressure or
composition, can be specified, in many cases by:
PIL_name_of_property = GRNDx
An example is:
RHO1=GRND3
Further examples may be inspected by clicking on the RHO1 above, and on the following variable names, grouped in accordance with whether they relate to:
3.1.3 Using In-Form
The method of setting non-uniform properties by way of formulae placed in the
Q1 file is explained fully in the
PHOENICS Encyclopaedia article
on In-Form, which there is no need to recapitulate.
Two sections are of especial relevance, namely:
A library case which illustrates the use of such formulae is 763, in which a variety of nearly-equivalent formulae for the same fluid are shown to produce (of course) almost the same flow fields.
Before In-Form became available, the most convenient method of introducing property formulae which were not represented in the GXx subroutines was to make use of PLANT.
This method is still available. It is described in detail here.
3.1.5 Introducing user's-own Fortran
Before PLANT was introduced, users wishing to introduce novel material
properties would do so by introducing their own coding into the
user-accessible GROUND sub-routine.
This method is also still available, and is described on-line here.
As explained above, the presence of different materials in different locations is conveyed to the PHOENICS solver by ascribing to each cell a value of the PRPS variable.
This is done by setting initial values, via FIINIT and INIT, for steady-flow cases; and for transient flows also if the materials do not move their positions.
Wherever the PRPS value is one of those listed in the PROPS file supplied by CHAM, or in another referred to by the user, the values of the properties which are used by the solver are those in the file, as has already been described.
However, wherever the PRPS has been set equal to -1.0 , the solver uses the values of the properties which correspond to the PIL variables RHO1, ENUL, CP1, etcetera.
Often this is achieved by setting FIINIT(PRPS) = -1.0, and using INIVAL-type patches for all regions occupied by recognised materials. However, how the PRPS field is set is immaterial; it is only the value in the cell which matters.
Of course, when objects move their positions relative to the computational grid, their motion must be reflected by changes in the PRPS values of the affected cell.
Two further remarks should be made concerning recently-introduced features of PHOENICS, namely PARSOL and In-Form. They are:
For example, a solid sphere may be immersed in a stream of air; and a cartesian grid may be used for the computation. Then some of its cells have their faces intersected by the surface of the sphere.
These so-called 'cut-cells' contain two materials, not one.
When the solver detects such cells, it sets PRPS equal to the value appropriate to the fluid. This may be -1, which is the case when the PIL variables are being used to define properties. Otherwise it will be a value appearing in the PROPS file
Only the cells which lie wholly within the enclosing surface of the solid retain the PRPS value which was assigned to the object.
If therefore this is undesired (as is usual when solids are immersed in a fluid), application of the formula must be specifically excluded by use of the 'with' condition.
