Package 'criticality'

BN Function

Description

This function creates a Bayesian network from pre-formatted nuclear facility data.

Usage

BN(dist = "gamma", facility.data, ext.dir)

Arguments

dist	Truncated probability distribution (e.g., "gamma", "normal")
facility.data	.csv file name
ext.dir	External directory (full path)

Value

A Bayesian network that models fissile material operations (op), controls (ctrl), and parameters that affect nuclear criticality safety

Examples

ext.dir <- paste0(tempdir(), "/criticality/extdata")
dir.create(ext.dir, recursive = TRUE, showWarnings = FALSE)

extdata <- paste0(.libPaths()[1], "/criticality/extdata")
file.copy(paste0(extdata, "/facility.csv"), ext.dir, recursive = TRUE)
file.copy(paste0(extdata, "/mcnp-dataset.RData"), ext.dir, recursive = TRUE)

BN(
  facility.data = "facility.csv",
  ext.dir = ext.dir
)

---

Model Function

Description

This function builds the deep neural network metamodel architecture.

Usage

Model(
  dataset,
  layers = "8192-256-256-256-256-16",
  loss = "sse",
  opt.alg = "adamax",
  learning.rate = 0.00075,
  ext.dir
)

Arguments

dataset	Training and test data
layers	String that defines the deep neural network architecture (e.g., "64-64")
loss	Loss function
opt.alg	Optimization algorithm
learning.rate	Learning rate
ext.dir	External directory (full path)

Value

A deep neural network metamodel of Monte Carlo radiation transport code simulation data

---

NN Function

Description

This function trains an ensemble of deep neural networks to predict keff values (imports Tabulate, Scale, Model, Fit, Plot, and Test functions).

Usage

NN(
  batch.size = 8192,
  code = "mcnp",
  dataset,
  ensemble.size = 5,
  epochs = 1500,
  layers = "8192-256-256-256-256-16",
  loss = "sse",
  opt.alg = "adamax",
  learning.rate = 0.00075,
  val.split = 0.2,
  overwrite = FALSE,
  remodel = FALSE,
  replot = TRUE,
  verbose = FALSE,
  ext.dir,
  training.dir = NULL
)

Arguments

batch.size	Batch size
code	Monte Carlo radiation transport code (e.g., "cog", "mcnp")
dataset	Training and test data
ensemble.size	Number of deep neural networks in the ensemble
epochs	Number of training epochs
layers	String that defines the deep neural network architecture (e.g., "64-64")
loss	Loss function
opt.alg	Optimization algorithm
learning.rate	Learning rate
val.split	Validation split
overwrite	Boolean (TRUE/FALSE) that determines if files should be overwritten
remodel	Boolean (TRUE/FALSE) that determines if an existing metamodel should be reused
replot	Boolean (TRUE/FALSE) that determines if .png files should be replotted
verbose	Boolean (TRUE/FALSE) that determines if TensorFlow and Fit function output should be displayed
ext.dir	External directory (full path)
training.dir	Training directory (full path)

Value

A list of lists containing an ensemble of deep neural networks and weights

Examples

ext.dir <- paste0(tempdir(), "/criticality/extdata")
dir.create(ext.dir, recursive = TRUE, showWarnings = FALSE)

extdata <- paste0(.libPaths()[1], "/criticality/extdata")
file.copy(paste0(extdata, "/facility.csv"), ext.dir, recursive = TRUE)
file.copy(paste0(extdata, "/mcnp-dataset.RData"), ext.dir, recursive = TRUE)

config <- FALSE
try(config <- reticulate::py_config()$available)
try(if (config == TRUE) {
  NN(
    batch.size = 128,
    ensemble.size = 1,
    epochs = 10,
    layers = "256-256-16",
    loss = "sse",
    replot = FALSE,
    ext.dir = ext.dir
  )
})

---

Plot Function

Description

This function generates and saves plots and data.

Usage

Plot(i, history = NULL, plot.dir)

Arguments

i	Model number
history	Training history
plot.dir	Plot directory (full path)

Value

No output (generates and saves ggplot2 files and training histories)

---

Risk Function

Description

This function estimates process criticality accident risk (imports Sample function).

Usage

Risk(
  bn,
  code = "mcnp",
  cores = parallel::detectCores()/2,
  dist = "gamma",
  facility.data,
  keff.cutoff = 0.9,
  metamodel,
  risk.pool = 100,
  sample.size = 1e+09,
  usl = 0.95,
  ext.dir,
  training.dir = NULL
)

Arguments

bn	Bayesian network
code	Monte Carlo radiation transport code (e.g., "cog", "mcnp")
cores	Number of CPU cores to use for generating Bayesian network samples
dist	Truncated probability distribution (e.g., "gamma", "normal")
facility.data	.csv file name
keff.cutoff	keff cutoff value (e.g., keff >= 0.9)
metamodel	List of deep neural network metamodels and weights
risk.pool	Number of times risk is calculated
sample.size	Number of samples used to calculate risk
usl	Upper subcritical limit (e.g., keff >= 0.95)
ext.dir	External directory (full path)
training.dir	Training directory (full path)

Value

A list of lists containing process criticality accident risk estimates and Bayesian network samples

Examples

ext.dir <- paste0(tempdir(), "/criticality/extdata")
dir.create(ext.dir, recursive = TRUE, showWarnings = FALSE)

extdata <- paste0(.libPaths()[1], "/criticality/extdata")
file.copy(paste0(extdata, "/facility.csv"), ext.dir, recursive = TRUE)
file.copy(paste0(extdata, "/mcnp-dataset.RData"), ext.dir, recursive = TRUE)

config <- FALSE
try(config <- reticulate::py_config()$available)
try(if (config == TRUE) {
  Risk(
    bn = BN(
      facility.data = "facility.csv",
      ext.dir = ext.dir),
    code = "mcnp",
    cores = 1,
    facility.data = "facility.csv",
    keff.cutoff = 0.5,
    metamodel = NN(
      batch.size = 128,
      ensemble.size = 1,
      epochs = 10,
      layers = "256-256-16",
      replot = FALSE,
      ext.dir = ext.dir),
    risk.pool = 10,
    sample.size = 1e+04,
    ext.dir = ext.dir,
    training.dir = NULL
  )
})

---

Sample Function

Description

This function samples the Bayesian network and generates keff predictions using a deep neural network metamodel.

Usage

Sample(
  bn,
  code = "mcnp",
  cores = parallel::detectCores()/2,
  keff.cutoff = 0.9,
  metamodel,
  sample.size = 1e+09,
  ext.dir,
  risk.dir = NULL
)

Arguments

bn	Bayesian network object
code	Monte Carlo radiation transport code (e.g., "cog", "mcnp")
cores	Number of CPU cores to use for generating Bayesian network samples
keff.cutoff	keff cutoff value (e.g., 0.9)
metamodel	List of deep neural network metamodels and weights
sample.size	Number of samples used to calculate risk
ext.dir	External directory (full path)
risk.dir	Risk directory

Value

A list of Bayesian network samples with predicted keff values

---

Scale Function

Description

This function centers, scales, and one-hot encodes variables.

Usage

Scale(code = "mcnp", dataset = NULL, output, ext.dir)

Arguments

code	Monte Carlo radiation transport code (e.g., "cog", "mcnp")
dataset	Training and test data
output	Processed output from Monte Carlo radiation transport code simulations
ext.dir	External directory (full path)

Value

A list of centered, scaled, and one-hot-encoded training and test data

---

Tabulate Function

Description

This function loads/saves training and test data (imports Scale function).

Usage

Tabulate(code = "mcnp", ext.dir)

Arguments

code	Monte Carlo radiation transport code (e.g., "cog", "mcnp")
ext.dir	External directory (full path)

Value

A list of centered, scaled, and one-hot-encoded training and test data

Examples

ext.dir <- paste0(tempdir(), "/criticality/extdata")
dir.create(ext.dir, recursive = TRUE, showWarnings = FALSE)

extdata <- paste0(.libPaths()[1], "/criticality/extdata")
file.copy(paste0(extdata, "/facility.csv"), ext.dir, recursive = TRUE)
file.copy(paste0(extdata, "/mcnp-dataset.RData"), ext.dir, recursive = TRUE)

Tabulate(
  ext.dir = ext.dir
)

---

Test Function

Description

This function calculates deep neural network metamodel weights and generates keff predictions for all training and test data.

Usage

Test(dataset, ensemble.size = 5, loss = "sse", ext.dir, training.dir)

Arguments

dataset	Training and test data
ensemble.size	Number of deep neural networks in the ensemble
loss	Loss function
ext.dir	External directory (full path)
training.dir	Training directory (full path)

Value

A list of deep neural network weights

---