count_alarms_hetero_faster: Count and plot alarms using a spline using heterogeneous...

Description Usage Arguments Value Examples

View source: R/count_alarms_hetero_faster.R

Description

Function that divides load_training into quantiles using bins as the number of quantiles. For each quantile the mean and standard deviation are calculated. A spline is fitted to the means and two splines are fitted to three sigma above and below the means, respectively. The function generates alarms when the test data fall outside the band of three sigma. The function generates a plot by default, and can also show the summary of the quantiles if show_summary = TRUE.

Usage

1
2
3
count_alarms_hetero_faster(load_training, vib_training, knots, load_test,
  vib_test, show_figure = TRUE, show_summary = FALSE, bins = 160,
  load_min = 0)

Arguments

load_training

A vector containing the training data of the load variables such as power or generator speed.

vib_training

A vector containing the training data of the vibration signal.

knots

A vector containing the locations of the knots.

load_test

A vector with the test data of the load variable.

vib_test

A vector with the test data of the vibration signal.

show_figure

A logical vector indicating if a plot should be made.

show_summary

A logical vector stating whether the quantile summaries should be shown or not.

bins

The number of bins used.

load_min

The minimum load to be used.

Value

A list of the count of the number of alarms, alarm rate, and residuals. Furthermore a plot is generated by default, set by show_figure.

count

A count of the number of alarms in the test set.

alarm_rate

The alarm rate in the test set.

residuals

The residuals in the test set.

Examples

1
2
count_alarms_hetero_var_faster(ltrain, vtrain, knots = c(20,25,27),
                                 ltest, vtest, load_min = 17)

hannea/monitor documentation built on May 14, 2017, 10:26 p.m.