Add Sickle Cell Model and Related Tools to RBC3D Simulation #1

smullangi3 · 2023-07-29T04:53:50Z

This pull-request adds an input file SickleCell.dat, as well as functions in ModIO.F90 for reading the input file into a cell for simulation.

ModRBC.F90 modifications include:

another possible sickle cell model (commented). Unfortunately, this model was ill-conditioned for the simulation
The "RBC_MakeUncurvedSickleSpheroid" subroutine, which creates an uncurved prolate spheroid. To produce the "SickleCell.dat" Sickle Cell model, this uncurved prolate spheroid was then inserted into a flow: vbkg = 8 for 5000 timesteps of size 0.00014, and with mechanical properties equivalent to a healthy Biconcave RBC (celltype 1 in case/tube.F90).

Tube.F90 also contains modifications to include the Sickle cell as celltype #3, where mechanical properties such as shear and bending modulus are set to 3x and 4x higher than a conventional healthy RBC respectively.

As seen in Initcond.F90, we can create a Sickle Cell in a simulation by replacing:
call Rbc_Create(rbc, nlat0, dealias)
call RBC_MakeBiconcave(rbc, radEqv, xc)
with:
call ImportReadRbc('Input/SickleCell.dat', rbc, xc)
wherever we want a Sickle Cell instead of a Biconcave RBC.

Currently, Initcond sets up a simulation of 8 cells in single file inside a tube, where cells alternate between healthy and sickle.

This code has been tested through simulations, comprising of:

1 sickle cell
Many sickle cells
Combinations of sickle and other cells
In all cases, errors don't occur and the simulation behaves as expected (as long as the timestep size isn't too large to cause divergence - a suitable size is set in the below input file case_different_celltypes/Input/tube.in).

(code commits are a bit all-over, i recommend probably doing a squash merge once this goes in)

sbryngelson · 2023-07-29T15:04:26Z

case/Input/tube.in

-1.                      ! refRad
+3                       ! nCellTypes
+1 1 1                   ! viscRat
+1. 1. 1.                ! refRad


this doesn't seem right, shoudn't it be 1 number? what about viscRat? Regardless, please create multiple "cases" that serve different example purposes.

ModConf.F90 line 245:

read(funit, *) nCellTypes; print *, 'nCellTypes = ',nCelltypes read(funit, *) viscRatIN(1:nCellTypes) print *, 'viscRat = ', viscRatIN(1:nCellTypes) read(funit, *) refRadIN(1:nCellTypes) print *, 'refRad = ', refRadIN(1:nCellTypes)

so, it looks like you need as many viscRats and refRads as the number of cell types.

Yes, I will move this sickle+rbc simulation to a new case folder.

sbryngelson · 2023-07-29T15:04:56Z

case/Input/tube.in

-10000                   ! Nt
-0.0014                  ! Ts
+10000                    ! Nt
+0.00014                  ! Ts


this very small time step is an example of why this example case should be separate from a health rbc one, so we need more than one example

Yes, will create new case-folder for rev2 for the changes.

sbryngelson · 2023-07-29T15:05:41Z

common/ModIO.F90

+    close(cell_unit)
+  end subroutine WriteRBCPlain
+
+  subroutine ReadRBCPlain(fn, rbc, xc)


by plain do you mean healthy?

Sorry, no this was bad naming on my part. I just mean that:
WriteRBCPlain takes a cell and a filename, and writes just the cell's plain information (w/o any transformations). Then, ReadRBCPlain can reread in this cell model for a different simulation.

For next rev, I will add better comments and naming convention

sbryngelson · 2023-07-29T15:07:34Z

case/initcond.F90

+          call Rbc_Create(rbc, nlat0, dealias)
+          call RBC_MakeBiconcave(rbc, radEqv, xc)
+        else 
+          call ReadRBCPlain('Input/SickleCell.dat', rbc, xc)


so plain is sickle? also, what it seems like the input data depends on the values of nlat and nlon, or no? if so, what does one do if they want a sickle cell with different nlat and nlon?

Yes, in this case, we have a file of a Sickle Cell (previously exported) which is imported using the ReadRBCPlain subroutine. I will fix naming and explain better via comments in next revision.

The imported cell will be the same exact shape as the cell which was previously exported to the defined file. So, if we want a Sickle Cell with different nlat/nlon, we would have to recreate the sickle cell with those different measurements and then re-export it.

sbryngelson · 2023-07-29T15:08:08Z

case/tube.F90

      call RBC_ComputeGeometry(rbcRef)

      rbcRef => rbcRefs(2)
      call RBC_Create(rbcRef, nlat0)
      call RBC_MakeSphere(rbcRef, radEqv)
      call RBC_ComputeGeometry(rbcRef)

+      rbcRef => rbcRefs(3)
+      call ReadRBCPlain('Input/SickleCell.dat', rbcRef)


so plain is sickle cell? what happens if the number of points on the sickle cell nlat and nlon doesn't match the values the user inputs?

the sickle cell's nlat/nlon in this case will just be the same as the nlat/nlon of the exported cell-shape. So, even if the user changes the input nlat0, the sickle cell disregards that and uses the model's data from the Input/SickleCell.dat file.

sbryngelson · 2023-07-29T15:09:01Z

common/ModRbc.F90

@@ -175,6 +177,115 @@ subroutine RBC_MakeSphere(cell, r, xc)
    end if

  end subroutine RBC_MakeSphere
+
+  subroutine RBC_MakeSickleGraham(cell, r, xc)


please make this more descriptive than Graham, also put a reference to the paper where these ideas come from in the comments. so far the code in this PR has no comments (!!)

Yes, will change to better meaningful names and add explanations in the comments for rev2

sbryngelson · 2023-07-29T15:09:29Z

case/Input/SickleCell.dat

@@ -0,0 +1,4 @@
+          12          24


this looks like nlat and nlon and a dealiasing factor.

Yep, this was the nlat and nlon (12, 24) of the cell in the file. So, when I import the cell, it will always be with these dimensions (defined in the file when exported).

sbryngelson · 2023-07-29T15:09:57Z

Makefile.in

@@ -1,5 +1,5 @@
 # Directories (change this to where you put RBC3D!)
-WORK_DIR = /storage/home/hcoda1/6/sbryngelson3/p-sbryngelson3-0/test-rbc/RBC3D
+WORK_DIR = /storage/home/hcoda1/6/smullangi3/RBC3D


remove this from PR

Will do in rev2

…omments to /common changes

sbryngelson · 2023-07-31T13:33:05Z

case_different_celltypes/Input/tube.in

+8                       ! PBspln_Ewld
+
+3                       ! nCellTypes
+1 1 3                   ! viscRat (Sickle Cells experience ~3 to 4x viscosity as a healthy RBC) Reference: https://doi.org/10.1016/j.actbio.2012.07.011


make this 1. unclear if the simulation is stable for viscRat /= 1.

Will change back to all 1s in rev3

common/ModIO.F90

sbryngelson · 2023-07-31T13:36:38Z

common/ModRbc.F90

+  ! Unfortunately this shape doesn't work very well since points are ill-conditioned
+  ! and diverge for this type of simulation, so we do not proceed with this sickle-cell model
+
+  ! subroutine RBC_MakeSickle(cell, rad, xc)


uncomment this and then add an write('this configuration is not confirmed working yet' and an mpi_abort() call. do not leave stale code in the codebase.

Will do this in rev3

sbryngelson · 2023-07-31T13:36:48Z

common/ModRbc.F90

+  !     end if
+  !   end do
+
+  ! end function RBC_SolveSickleRho


uncomment this function

Sure - also this function uses a different module for evaluating quartic roots. So I'll uncomment this and include that new ModPolyRoots module in rev3

Subrahmanyam Raghav Mullangi and others added 17 commits April 28, 2023 22:06

Redefined Makefile for Phoenix Cluster and Modified Initcond for Scaling

6599e7c

Merge branch 'master' of https://github.com/smullangi3/RBC3D

fda587b

Created Python Data Collection Script

6c87465

Batch Run File Configuration

aa98b87

Set up simulations for Scalability testing

5fe6ffc

Started Flawed Spherical Coords Sickle Cell Generation

918dda1

Renamed Poly Roots module for running build

e79d331

Modified sbatch to build everything first

adfb75f

Adjusted Spherical Algorithm and SS Init

a2eb82a

Fixed SS Generation to Correct Shape

f7f9013

Made SS a modifiable radius

09b1b38

Changed configurations and checks

1e0b4b8

Sickle Cell OG model

e93e0b9

Added Sickle Cell Simulation to RBC3D

73d48d1

Set up Sickle Cell for PR

5b940f9

Fix initcond formatting

a847bd9

Reformatted and modified Input config file

4d837ee

smullangi3 requested a review from sbryngelson July 29, 2023 04:55

sbryngelson reviewed Jul 29, 2023

View reviewed changes

smullangi3 added 4 commits July 30, 2023 19:23

Migrated Sickle Cell Simulation to Separate Case; Added Explanation C…

1bbd52b

…omments to /common changes

Fixed Formatting

a0c544b

Modify T size for Stability

4a06b2a

revert Makefile.in to Original

c4e71f9

smullangi3 requested a review from sbryngelson July 31, 2023 00:12

sbryngelson requested changes Jul 31, 2023

View reviewed changes

Revise Stale Code to Abort when Used

793c37d

smullangi3 requested a review from sbryngelson August 1, 2023 01:40

sbryngelson approved these changes Aug 1, 2023

View reviewed changes

sbryngelson merged commit ea82c01 into comp-physics:master Aug 1, 2023

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Add Sickle Cell Model and Related Tools to RBC3D Simulation #1

Add Sickle Cell Model and Related Tools to RBC3D Simulation #1

smullangi3 commented Jul 29, 2023 •

edited

Loading

sbryngelson Jul 29, 2023

smullangi3 Jul 30, 2023 •

edited

Loading

sbryngelson Jul 29, 2023

smullangi3 Jul 30, 2023

sbryngelson Jul 29, 2023

smullangi3 Jul 30, 2023 •

edited

Loading

sbryngelson Jul 29, 2023

smullangi3 Jul 30, 2023

sbryngelson Jul 29, 2023

smullangi3 Jul 30, 2023

sbryngelson Jul 29, 2023

smullangi3 Jul 30, 2023 •

edited

Loading

sbryngelson Jul 29, 2023

smullangi3 Jul 30, 2023

sbryngelson Jul 29, 2023

smullangi3 Jul 30, 2023

sbryngelson Jul 31, 2023

smullangi3 Jul 31, 2023

sbryngelson Jul 31, 2023

smullangi3 Jul 31, 2023

sbryngelson Jul 31, 2023

smullangi3 Jul 31, 2023

Add Sickle Cell Model and Related Tools to RBC3D Simulation #1

Add Sickle Cell Model and Related Tools to RBC3D Simulation #1

Conversation

smullangi3 commented Jul 29, 2023 • edited Loading

Choose a reason for hiding this comment

smullangi3 Jul 30, 2023 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

smullangi3 Jul 30, 2023 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

smullangi3 Jul 30, 2023 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

smullangi3 commented Jul 29, 2023 •

edited

Loading

smullangi3 Jul 30, 2023 •

edited

Loading

smullangi3 Jul 30, 2023 •

edited

Loading

smullangi3 Jul 30, 2023 •

edited

Loading