Neural closure models

NeuralClorusure

These features are experimental, and require cloning IncompressibleNavierStokes from GitHub:

git clone https://github.com/agdestein/IncompressibleNavierStokes.jl
cd IncompressibleNavierStokes/lib/NeuralClosure

Large eddy simulation, a closure model is required. With IncompressibleNavierStokes, a neural closure model can be trained on filtered DNS data. The discrete DNS equations are given by

$$\begin{array}{rl}Mu& =0,\\ \frac{\mathrm{d}u}{\mathrm{d}t}& =F(u)-Gp.\end{array}$$

Applying a spatial filter $\mathrm{\Phi }$, the extracted large scale components $\overline{u}=\mathrm{\Phi }u$ are governed by the equation

$$\begin{array}{rl}M\overline{u}& =0,\\ \frac{\mathrm{d}\overline{u}}{\mathrm{d}t}& =F(\overline{u})+c-G\overline{p},\end{array}$$

where the discretizations $M$, $F$, and $G$ are adapted to the size of their inputs and $c=\overline{F(u)}-F(\overline{u})$ is a commutator error. We here assumed that $M$ and $\mathrm{\Phi }$ commute, which is the case for face averaging filters. Replacing $c$ with a parameterized closure model $m(\overline{u},\theta )\approx c$ gives the LES equations for the approximate large scale velocity $\overline{v}\approx \overline{u}$

$$\begin{array}{rl}M\overline{v}& =0,\\ \frac{\mathrm{d}\overline{v}}{\mathrm{d}t}& =F(\overline{v})+m(\overline{v},\theta )-G\overline{q}.\end{array}$$

NeuralClosure module

IncompressibleNavierStokes provides a NeuralClosure module.

NeuralClosure.NeuralClosure Module

Neural closure modelling tools.

source

NeuralClosure.collocate Method

collocate(u) -> Any

Interpolate velocity components to volume centers.

source

NeuralClosure.create_closure Method

create_closure(layers...; rng)

Create neural closure model from layers.

source

NeuralClosure.create_tensorclosure Method

create_tensorclosure(layers...; setup, rng)

Create tensor basis closure.

source

NeuralClosure.decollocate Method

decollocate(u) -> Any

Interpolate closure force from volume centers to volume faces.

source

NeuralClosure.wrappedclosure Method

wrappedclosure(
    m,
    setup
) -> NeuralClosure.var"#neuralclosure#1"

Wrap closure model and parameters so that it can be used in the solver.

source

Filters

The following filters are available:

NeuralClosure.AbstractFilter Type

abstract type AbstractFilter

Discrete DNS filter.

Subtypes ConcreteFilter should implement the in-place method:

(::ConcreteFilter)(v, u, setup_les, compression)

which filters the DNS field u and put result in LES field v. Then the out-of place method:

(::ConcreteFilter)(u, setup_les, compression)

automatically becomes available.

Fields

source

NeuralClosure.FaceAverage Type

struct FaceAverage <: NeuralClosure.AbstractFilter

Average fine grid velocity field over coarse volume face.

Fields

source

NeuralClosure.VolumeAverage Type

struct VolumeAverage <: NeuralClosure.AbstractFilter

Average fine grid velocity field over coarse volume.

Fields

source

NeuralClosure.reconstruct! Method

reconstruct!(u, v, setup_dns, setup_les, comp) -> Any

Reconstruct DNS velocity u from LES velocity v.

source

NeuralClosure.reconstruct Method

reconstruct(v, setup_dns, setup_les, comp) -> Any

Reconstruct DNS velocity field. See also reconstruct!.

source

Training

To improve the model parameters, we exploit exact filtered DNS data $\overline{u}$ and exact commutator errors $c$ obtained through DNS. The model is trained by minimizing the a priori loss function

$$L^{\text{prior}}(\theta )=\parallel m(\overline{u},\theta )-c{\parallel }^2,$$

or the a posteriori loss function

$$L^{\text{post}}(\theta )=\parallel {\overline{v}}_{\theta }-\overline{u}{\parallel }^2,$$

where ${\overline{v}}_{\theta }$ is the solution to the LES equation for the given parameters $\theta$. The prior loss is easy to evaluate and easy to differentiate, as it does not involve solving the ODE. However, minimizing $L^{\text{prior}}$ does not take into account the effect of the prediction error on the LES solution error. The posterior loss does, but has a longer computational chain involving solving the LES ODE.

NeuralClosure.create_callback Method

create_callback(
    err;
    θ,
    callbackstate,
    displayref,
    displayfig,
    displayupdates,
    figfile,
    nupdate
)

Create convergence plot for relative error between f(x, θ) and y. At each callback, plot is updated and current error is printed.

If state is nonempty, it also plots previous convergence.

If not using interactive GLMakie window, set displayupdates to true.

source

NeuralClosure.create_dataloader_post Method

create_dataloader_post(
    trajectories;
    ntrajectory,
    nunroll,
    device
)

Create trajectory dataloader.

source

NeuralClosure.create_dataloader_prior Method

create_dataloader_prior(
    data;
    batchsize,
    device
) -> NeuralClosure.var"#dataloader#44"{Int64}

Create dataloader that uses a batch of batchsize random samples from data at each evaluation. The batch is moved to device.

source

NeuralClosure.create_loss_post Method

create_loss_post(
;
    setup,
    method,
    psolver,
    closure,
    nsubstep
)

Create a-posteriori loss function.

source

NeuralClosure.create_loss_prior Function

create_loss_prior(
    f
) -> NeuralClosure.var"#loss_prior#57"{_A, NeuralClosure.var"#55#56"} where _A
create_loss_prior(
    f,
    normalize
) -> NeuralClosure.var"#loss_prior#57"

Return mean squared error loss for the predictor f.

source

NeuralClosure.create_relerr_post Method

create_relerr_post(
;
    data,
    setup,
    method,
    psolver,
    closure_model,
    nsubstep
)

Create a-posteriori relative error.

source

NeuralClosure.create_relerr_prior Method

create_relerr_prior(f, x, y) -> NeuralClosure.var"#58#59"

Create a-priori error.

source

NeuralClosure.create_relerr_symmetry_post Method

create_relerr_symmetry_post(
;
    u,
    setup,
    method,
    psolver,
    Δt,
    nstep,
    g
)

Create a-posteriori symmetry error.

source

NeuralClosure.create_relerr_symmetry_prior Method

create_relerr_symmetry_prior(; u, setup, g)

Create a-priori equivariance error.

source

NeuralClosure.train Method

train(
;
    dataloader,
    loss,
    trainstate,
    niter,
    callback,
    callbackstate,
    λ
)

Update parameters θ to minimize loss(dataloader(), θ) using the optimiser opt for niter iterations.

Return the a new named tuple (; opt, θ, callbackstate) with updated state and parameters.

source

NeuralClosure.trainepoch Method

trainepoch(
;
    dataloader,
    loss,
    trainstate,
    callback,
    callbackstate,
    device,
    noiselevel,
    λ
)

Update parameters θ to minimize loss(dataloader(), θ) using the optimiser opt for niter iterations.

Return the a new named tuple (; opt, θ, callbackstate) with updated state and parameters.

source

Neural architectures

We provide neural architectures: A convolutional neural network (CNN), group convolutional neural networks (G-CNN) and a Fourier neural operator (FNO).

NeuralClosure.cnn Method

cnn(
;
    setup,
    radii,
    channels,
    activations,
    use_bias,
    channel_augmenter,
    rng
)

Create CNN closure model. Return a tuple (closure, θ) where θ are the initial parameters and closure(u, θ) predicts the commutator error.

source

NeuralClosure.GroupConv2D Type

struct GroupConv2D{C} <: LuxCore.AbstractLuxLayer

Group-equivariant convolutional layer – with respect to the p4 group. The layer is equivariant to rotations and translations of the input vector field.

The kwargs are passed to the Conv layer.

The layer has three variants:

• If islifting then it lifts a vector input (u1, u2) into a rotation-state vector (v1, v2, v3, v4).

• If isprojecting, it projects a rotation-state vector (u1, u2, u3, v4) into a vector (v1, v2).

• Otherwise, it cyclically transforms the rotation-state vector (u1, u2, u3, u4) into a new rotation-state vector (v1, v2, v3, v4).

Fields

• islifting

• isprojecting

• cin

• cout

• conv

source

NeuralClosure.gcnn Method

gcnn(; setup, radii, channels, activations, use_bias, rng)

Create CNN closure model. Return a tuple (closure, θ) where θ are the initial parameters and closure(u, θ) predicts the commutator error.

source

NeuralClosure.rot2 Function

rot2(u, r)

Rotate the field u by 90 degrees counter-clockwise r - 1 times.

source

NeuralClosure.rot2 Method

rot2(u::Tuple{T, T}, r) -> Union{Nothing, Tuple{Any, Any}}

Rotate vector fields [ux;;; uy]

source

NeuralClosure.rot2stag Method

rot2stag(u, g) -> Any

Rotate staggered grid velocity field. See also rot2.

source

NeuralClosure.vecrot2 Method

vecrot2(u, r) -> Any

Rotate vector fields [ux;;; uy]

source

NeuralClosure.FourierLayer Type

struct FourierLayer{D, A, F} <: LuxCore.AbstractLuxLayer

Fourier layer operating on uniformly discretized functions.

Some important sizes:

• dimension: Spatial dimension, e.g. Dimension(2) or Dimension(3).

• (nx..., cin, nsample): Input size

• (nx..., cout, nsample): Output size

• nx = fill(n, dimension()): Number of points in each spatial dimension

• n ≥ kmax: Same number of points in each spatial dimension, must be larger than cut-off wavenumber

• kmax: Cut-off wavenumber

• nsample: Number of input samples (treated independently)

Fields

• dimension

• kmax

• cin

• cout

• σ

• init_weight

source

NeuralClosure.fno Method

fno(; setup, kmax, c, σ, ψ, rng, kwargs...)

Create FNO closure model. Return a tuple (closure, θ) where θ are the initial parameters and closure(V, θ) predicts the commutator error.

source

Data generation

NeuralClosure.create_io_arrays Method

create_io_arrays(data, setup) -> NamedTuple

Create $(\overline{u},c)$ pairs for a-priori training.

source

NeuralClosure.create_les_data Method

create_les_data(
;
    D,
    Re,
    lims,
    nles,
    ndns,
    filters,
    tburn,
    tsim,
    savefreq,
    Δt,
    method,
    create_psolver,
    backend,
    icfunc,
    processors,
    rng,
    filenames,
    kwargs...
)

Create filtered DNS data.

source

NeuralClosure.filtersaver Method

filtersaver(
    dns,
    les,
    filters,
    compression,
    psolver_dns,
    psolver_les;
    nupdate,
    filenames,
    F,
    p
)

Save filtered DNS data.

source