Add simple model to BoltzmannMFX

[bjpalmer]

Take simple model documented in docs/models/model1/model1.tex and add it to BoltzmannMFX.

[bjpalmer]

The transfer of material between the continuum fields and the particles is described in the bmx_calc_txfr_fluid (transfer from particles to continuum fields) and bmx_calc_txfr_particles (transfer from continuum fields to particles). Both these routines are located in bmx_calc_txfr.cpp located in src/des.

[bjpalmer]

The number of chemical species in the fluid is defined in bmx_fluid_parms.H and the number of chemical species inside a particle is probably defined in bmx_chem_species_parms.H (not 100% sure of this). No chemical species inside the particles seemed to be defined in the inputs file for the 1-particle test case in exec. Species are defined in the fluid.

[asalmgren]

The real (as opposed to integer) data carried by each particle is defined in des/bmx_pc.H -- currently this is struct realData { enum { // Real particle data stored in p.m_data velx, vely, velz, radius, volume, fluid_A, fluid_B, consume_A, consume_B, count }; }; The way these components are used in the current code, the fluid_A and fluid_B components are used to hold the values of X_A and X_B as interpolated from the fluid. If one wants to add chemical species actually carried by the particles, then one can add them to the enum list here. They shoudl then be initialized in des/bmx_pc_init.cpp Currently bmx_chem_species_parms.H is largely irrelevant, but could be revived as needed.

…

On Thu, Mar 11, 2021 at 11:48 AM bjpalmer ***@***.***> wrote: The number of chemical species in the fluid is defined in bmx_fluid_parms.H and the number of chemical species inside a particle is probably defined in bmx_chem_species_parms.H (not 100% sure of this). No chemical species inside the particles seemed to be defined in the inputs file for the 1-particle test case in exec. Species are defined in the fluid. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#3 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ACRE6YSAHM7HYZIA3QGMWS3TDEF75ANCNFSM4Y4CATZA> .

-- Ann Almgren Senior Scientist; CCSE Group Lead

[bjpalmer]

I'm trying to generalize these data structures to handle a more generic set of cells. I've defined two new enumerations:

/**

• Generic set of indices for variables characterizing a cell. Specific real
• properties of a cell are enumerated first, followed by the first index of an
• uncharacterized block of real variables that can be fed to other routines.
• The indices below assume that the most generic cell is an ellipsoid with axes
• a, b, c oriented in space by three Euler angles and acted upon by a linear
• force on the center of mass and a torque. The cell is also characterized by a
• volume V. The cell has linear velocity and angular velocity and the cell
• volume is also growing or shrinking with some rate. For spherical cells,
• a=b=c=radius and the Euler angles, rotation velocity and torques are all
• zero.
  */
  struct realIdx
  {
  enum { // Locations of real particle data stored in particle type
  a_size, // first ellipsoid dimension
  b_size, // second ellipsoid dimension
  c_size, // third ellipsoid dimension
  psi, // first Euler angle
  theta, // second Euler angle
  phi, // third Euler angle
  vol, // cell volume
  velx, // x-component of linear velocity
  vely, // y-component of linear velocity
  velz, // z-component of linear velocity
  wx, // x-component of angular velocity
  wy, // y-component of angular velocity
  wz, // z-component of angular velocity
  fx, // x-component of linear force
  fy, // y-component of linear force
  fz, // z-component of linear force
  taux, // x-component of torque
  tauy, // y-component of torque
  tauz, // z-component of torque
  dvdt, // rate of change of the volume
  first_data // index of first value in uncharacterized block
  };
  };

/**

• Generic set of indices for integer variables characterizing a cell. Specific
• integers are listed first followed by an uncharacterized block of integers
• that can be fed to other routines. The total number of real and integer
• variables in this block and the readIdx block are stored in this struct.
  */
  struct intIdx
  {
  enum { // Locations of integer data stored in particle type
  size_real, // Size of uncharacterized block of real data
  size_int, // Size of uncharacterized block of intger data
  first_data, // index of first value of uncharacterized block
  };
  };

The idea is that the mechanics are hardwired to describe a general ellipsoid interacting with other ellipsoids. Simpler geometric models (e.g. spheres) are subsets of this ellipsoidal model. The chemistry model is arbitrary and consists of a collections of real fields (chemical species) that are unpacked by a chemistry module. The readIdx::first_data index points to the start of the block of real data for each particle and the size of the block is stored in intIdx::size_real. The integer variables are similar. Again, additional integer variables describing the cell state are stored in a block of data that can be accessed using the intIdx::first_data variable.

[bjpalmer]

The simple ABC model has been added to BoltzmannMFX and is running for short time periods. It exhibits cell growth and division and also contains mechanical interactions between cells. The behaviors included in the simulation so far are

• Modification of chemical species due to reactions
• Cell division when cells reach a size threshold
• Motion of cells due to mechanical interactions