Choice Between Nuclear Thermal Coupling Methods

[js-jixu]

Hi, developers.

From the paper, there are two methods of nuclear thermal coupling.

The first method corresponds to phase-1. A preliminary power distribution is firstly calculated using Neutronics Field. The power distribution is then transferred to the Thermal-Hydraulics Field. Thermal-Hydraulics Field calculates a new temperature distribution based on power and transfers it to Neutronics Field for considering negative feedback and calculating new power and neutron flux distributions. The calculation of Thermal-Hydraulics Field is a transient calculation, and the calculation of Neutronics Field is a steady-state calculation. Do you think this method is capable of simulating negative feedback effects and negative doppler fuel temperature coefficients? This is figure:

The second method firstly calculate the neutron flux and power distribution of a steady state Neutronics Field. Using this distribution as the initial input condition, the transient Neutronics Field and Thermal-Hydraulics Field are simultaneously solved in a fully coupled manner.

In terms of computing resource consumption, the first one has advantages; in terms of accuracy, the second one has advantages. For me, I just want to simulate the correct negative doppler fuel temperature coefficient and negative feedback effects. I have tried the first method, but probably due to the physical setup of the single rod model, the negative feedback effect is very, very subtle.

So just from the point of view of the coupling method, I want to know which coupling method is better and can correctly simulate the interaction between the nucleus and heat.

[js-jixu]

I read the input file of phase-2 carefully and have some doubts. I modified transient-perturb-7.i, so the following discussion is only for transient-perturb-7.i.

In the [Functions] block in transient-perturb-7.i, why is the pre1 multiplied by 7.61666e+17? From the paper, you want to make the reactor exactly critical by scaling the neutron source terms. But why does multiplying pre1 to pre6 by 7.61666e+17 make the reactor exactly critical?

When I first ran transient-perturb-7.i it didn't converge. Then I modified the settings in the Executioner block. The original seeting is:

  petsc_options_iname = '-pc_type -sub_pc_type -ksp_gmres_restart -pc_asm_overlap -sub_pc_factor_shift_type'
  petsc_options_value = 'asm     lu           200                1                NONZERO'

I replaced it with:

  petsc_options_iname = '-pc_type -pc_factor_shift_type'
  petsc_options_value = 'lu       NONZERO'

The calculation can now converge. But after some steps, it becomes difficult to converge again. I put all the files generated during the calculation here. I divided the zip into two folders, the folder called original contains the original transient-perturb-7.i file, and the one called modified contains the modified transient-perturb-7.i file. Both of them have the output files of Step1.4 and cross section files.

files.zip

Although it converges in a few steps, the calculated result looks a little strange. For example, its power and temperature have risen to an impossible value. Why is this happening?

[js-jixu]

Based on the second method mentioned above and your phase-2 input file, I wrote a new fully coupled input file.

It's a supercritical single rod, and I don't have eigenvalue_scaling set. Like the transition from phase-1 to phase-2, I first calculated the neutron distribution and power distribution with a temperature of 593K in all regions in the steady state (power is a given value in the steady state calculation), and put these as the initial value for the fully coupled input file. Then I combined the transient equations of Thermal-Hydraulics Field and Neutronics Field by setting [Nt], [Precursors] and [Kernels] as shown in the input file of phase-2. The parts of the neutron physics are as follows:

[Nt]
  var_name_base = group
  temperature = T_merged
  vacuum_boundaries = 'cool_bottom cool_top clad_bottom clad_top fuel_bottom fuel_top'
  create_temperature_var = false
  init_nts_from_file = true
  block = 'fuel cool clad'
  pre_blocks = 'fuel'
[]

[Precursors]
  [pres]
    var_name_base = pre
    temperature = T_merged
    outlet_boundaries = ''
    constant_velocity_values = true
    u_def = 0
    v_def = 0
    w_def = 0
    nt_exp_form = false
    family = MONOMIAL
    order = CONSTANT
    loop_precursors = false
    transient = true
    block = 'fuel'
  []
[]

Will the total power of the single rod change without artificially giving any change in power? Since the single rod is supercritical, will the neutron flux go up indefinitely?

[js-jixu]

The answer can be found in this discussion.