Quadrature Method of Moments#

This subsection defines the parameters of the quadrature method of moments (QMOM) auxiliary physics, used to solve population balance equations. The physics is enabled with set qmom = true in the Multiphysics subsection and requires Lethe to be compiled with QMOMKit (-DLETHE_WITH_QMOMKIT=ON).

Note

For more information on the QMOMKit library, including details on the implementation and some application examples, visit the repository page.

The QMOM physics transports:

  • number of scalars regular scalars;

  • the raw moments \(M_0, \dots, M_{K-1}\) of the particle size distribution, where \(K\) is the number of moments. The quadrature has \(N = \lceil K/2 \rceil\) nodes, since \(N\) nodes are defined by \(2N\) moments.

All the fields are discretized with a discontinuous Galerkin finite elements of degree qmom degree (see Finite Element Interpolation (FEM)). The scalars are named after the scalar names and the moments are named moment_0 to moment_<K-1>.

Caution

All the parameters of the QMOM physics are in SI units.

Warning

The QMOM physics is under development. Only its parameters are available: enabling it in the Multiphysics subsection stops the simulation with an error, since its solver is not implemented yet.

Parameters#

The default values of the QMOM parameters are given in the text box below:

subsection QMOM
  set process            = generic
  set number of scalars  = 0
  set scalar names       = # empty: scalar_0, scalar_1, ...
  set number of moments  = 4
  set splitting scheme   = lie
  set enable nucleation  = false
  set enable coalescence = false
  set enable growth      = false

  subsection phase evolution
    set enable adaptive node count = false
  end

  subsection lipid nanoparticle production
    # Operating conditions
    set operating temperature         = 295
    set flow rate ratio               = 3
    set total flow rate               = 2e-07
    set mixing time scale             = 0
    set initial concentration organic = 1
    set ion concentration mixture     = 1

    # Aqueous phase
    set density aqueous                  = 997
    set molar mass aqueous               = 0.018015
    set mole fraction solubility aqueous = 2e-08
    set permittivity aqueous             = 78.2

    # Organic phase
    set density organic                  = 789
    set molar mass organic               = 0.046069
    set mole fraction solubility organic = 0.0029
    set permittivity organic             = 24.3

    # Mixture and particles
    set viscosity mixture = 0.002
    set shape factor      = 0.5235987755982988
  end

  subsection physical constants
    set boltzmann constant  = 1.380649e-23
    set avogadro number     = 6.02214076e+23
    set elementary charge   = 1.602176634e-19
    set permittivity vacuum = 8.8541878188e-12
  end
end
  • process: process simulated. The choices are:

    • generic: the scalars and the moments are given by number of scalars, scalar names and number of moments.

    • lipid_nanoparticle_production: lipid nanoparticle production, whose parameters are given in the lipid nanoparticle production subsection. The process imposes two scalars, ethanol_concentration and free_lipids_concentration, and 8 moments (4-node quadrature).

    A process other than generic ignores number of scalars, scalar names and number of moments, and stops the simulation with an error if they contradict it.

  • number of scalars: number of regular scalars transported together with the moments.

  • scalar names: comma-separated list of scalars names, used for the output files (e.g. set scalar names = ethanol_concentration, free_lipids_concentration). The names may only contain letters, digits and underscores. The default value is empty, in which case the scalars are named scalar_0, scalar_1, etc.

  • number of moments: number \(K\) of transported raw moments. The quadrature has \(N = \lceil K/2 \rceil\) nodes: when \(K\) is odd, one more node is used, so that all the moments are used by the quadrature. The default value, 4, gives a 2-node quadrature, the smallest one that represents a polydisperse distribution.

  • splitting scheme: splitting of the physical-space transport \(T\) and of the phase-space evolution \(P\) over a time step \(\Delta t\). The choices are:

    • lie: \(T(\Delta t)\) followed by \(P(\Delta t)\). This scheme is first-order accurate in time.

    • strang: \(P(\Delta t / 2)\), then \(T(\Delta t)\), then \(P(\Delta t / 2)\). This scheme is second-order accurate in time when both parts are at least second-order accurate.

  • enable nucleation, enable coalescence and enable growth: enable the corresponding phase-space mecanisms. They are disabled by default, so that the moments are only transported.

  • phase evolution: options of the phase-space evolution in QMOMKit.

    • enable adaptive node count: whether QMOMKit may reduce the number of quadrature nodes when the moments are close to the boundary of the realizable set.

  • lipid nanoparticle production: process parameters of the production of lipid nanoparticles (LNP), in SI units. The default values correspond to a water/ethanol formulation in a microfluidic mixer at room temperature: a flow rate ratio of 3 and a total flow rate of 12 mL/min.

    • operating temperature: temperature of the mixture, in K.

    • flow rate ratio: ratio of the aqueous phase flow rate to the organic phase flow rate.

    • total flow rate: total flow rate of the aqueous and organic phases, in m3/s.

    • mixing time scale: time scale of the mixing of the aqueous and organic phases, in s.

    • initial concentration organic: lipid concentration in the organic phase before mixing, in mol/m3.

    • ion concentration mixture: ion concentration in the mixture, in mol/m3.

    • density aqueous and density organic: densities of the aqueous and organic phases, in kg/m3.

    • molar mass aqueous and molar mass organic: molar masses of the aqueous and organic phases, in kg/mol.

    • mole fraction solubility aqueous and mole fraction solubility organic: mole fraction solubility of the lipid in the pure aqueous and organic phases, between 0 and 1.

    • permittivity aqueous and permittivity organic: relative permittivities of the aqueous and organic phases.

    • viscosity mixture: dynamic viscosity of the mixture of the aqueous and organic phases, in Pa s.

    • shape factor: volume shape factor of the particles, which is \(\pi/6\) for spheres.

  • physical constants: physical constants, in SI units.

    • boltzmann constant: Boltzmann constant, in J/K.

    • avogadro number: Avogadro number, in 1/mol.

    • elementary charge: elementary charge, in C.

    • permittivity vacuum: vacuum permittivity, in F/m.

Functions of the QMOM fields#

The initial conditions (see Initial Conditions), the values of the dirichlet boundary conditions (see Boundary Conditions - Multiphysics), the analytical solution (see Analytical Solution) and the source terms (see Source Term) of the QMOM fields are all given in the same way. The scalars are given as a list of expressions, while the moments are not given one by one: they are the raw moments of a particle size distribution. The default values are:

# QMOM block within initial conditions, boundary conditions, and analytical solution
subsection QMOM
  subsection scalars
    set Function expression = 0
  end
  subsection distribution
    set type                = none
    set Function expression = 0
  end
end
  • scalars: values of the scalars.

    • Function expression: expressions of the scalars, as functions of space and time, separated by semicolons. There must be one expression per scalar, in the same order as the scalar names. The default value, a single expression, is therefore only valid for one scalar.

    Note

    The function expressions are ignored if number of scalars is 0.

  • distribution: particle size distribution from which the moments are computed.

    • type: family of the particle size distribution. The raw moments of each family are computed analytically by QMOMKit. The choices are:

      • none: no particles. All the moments are zero and the Function expression of the distribution is ignored.

      • normal: normal distribution, with parameters the mean and the standard deviation of the size, in m. It is only meaningful when the standard deviation is much smaller than the mean, since the normal distribution extends to negative sizes.

      • lognormal: log-normal distribution, with parameters the mean and the standard deviation of the natural logarithm of the size. The median size, in m, is the exponential of the first parameter.

      • gamma: gamma distribution, with parameters the shape and the scale, in m.

      • weibull: Weibull distribution, with parameters the shape and the scale, in m.

      • beta: beta distribution on the interval \([0, 1]\), with shape parameters \(\alpha\) and \(\beta\).

      • chi: chi distribution, with parameter the number of degrees of freedom.

      • chi_squared: chi-squared distribution, with parameter the number of degrees of freedom.

    • Function expression: number density of the particles, in particles per m3, followed by the parameters of the distribution, as functions of space and time, separated by semicolons. The number of expressions is therefore 3 for the normal, lognormal, gamma, weibull and beta distributions, and 2 for the chi and chi_squared distributions.

The moment of order \(k\) is \(M_k = n \, \mathrm{E}[L^k]\), where \(n\) is the number density and \(L\) the size of the particles. In particular, \(M_0\) is the number density. A number density of 0 gives zero moments whatever the distribution.

For example, with set number of scalars = 2, two scalars and a log-normal distribution of \(10^{18}\) particles per m3 with a median size of 100 nm, at the inlet of the domain:

subsection QMOM
  subsection scalars
    set Function expression = 1; 0
  end
  subsection distribution
    set type                = lognormal
    set Function expression = 1e18; log(1e-7); 0.3
  end
end