Advanced Page
In the Calculate window, the > page specifies how to calculate FTA results. The following table describes the advanced properties on this page.
 |
If you are using a license for the FTA Lite module, none of the options on this page are available.
|
|
Property
|
Description
|
||
|---|---|---|---|
|
Calculation method
The properties under this heading specify how you want to calculate FTA results. The first three methods use cut set-based methods to identify what combinations of events cause the top event of the tree to occur. The next method uses various exact calculation methods to determine the likelihood of occurrence of the top event. Calculations are based on the failure or repair data properties entered for lowest-level events. The final method uses Monte Carlo simulation. This process observes the dynamic behavior of the tree with varying sets of input. It does not use formulas derived from some type of mathematical analysis to compute analytical results.
|
|||
|
Cut set summation
|
When selected (default), the cut set summation method is used. It assumes that the simultaneous occurrence of multiple cut sets is rare. Therefore, the probability of a top event occurring is the sum of the probabilities of its cut set. This method is equivalent to selecting Cross product and entering 1 for Order of product terms.
|
||
|
Cross product
|
When selected, the cross product method is used. It assumes that the simultaneous occurrence of multiple cut sets is rare above a specified order. Order of product terms becomes available so that you can specify the order of product terms to take into account. The default is 1, which makes this method equivalent to the cut set summation method. However, you can specify a higher order (up to the number of cut sets) if you want to obtain more accurate results.
|
||
|
Esary Proschan
|
When selected, the Esary Proschan method is used. It assumes that all cut sets are independent. This method produces conservative results for coherent trees.
|
||
|
Exact
|
When selected, a rich set of advanced techniques is used. Depending on the type of tree, this method uses various solution techniques to obtain exact results in efficient ways.
|
||
|
Monte Carlo simulation
|
When selected, a Monte Carlo simulation engine is used to calculate unreliability and unavailability measures. This calculation method is available only for time dependent and Lambda Tau trees. When it is selected, properties under the Simulation heading become available.
|
||
|
Dynamic gate calculations
The single property under this heading is available only for time dependent trees when exact calculations are selected. For a tree with dynamic gates, parts of the tree must be converted to equivalent Markov models. For more information, see Dynamic Gate Calculations.
|
|||
|
Precision for Markov models of dynamic gates
|
The precision that the Markov calculation engine is to use to calculate results for dynamic gates. This property is available only if Exact or Perform exact calculations is selected. You can either select from the many choices or enter a value.
|
||
|
Simulation
The properties under this heading are available only if Monte Carlo simulation is selected as the calculation method for a time dependent or Lambda Tau Tree. They are also available if Uncertainly analysis is selected on the > page for any tree type. Both Monte Carlo simulation and uncertainty analysis use simulation techniques.
|
|||
|
Number of simulation iterations
|
The number of simulations to perform on the fault tree. You can enter any integer greater than 0. The default is 1000. A Monte Carlo simulation performs random tests on the specified tree to provide an analysis of overall unreliability and unavailability. The greater the number of simulation iterations requested, the greater number of random analyses performed, yielding more accurate results.
|
||
|
Specify random number seed
|
Indicates whether the random number seed for Monte Carlo simulation is specified. When this checkbox is selected (default), the field to the right becomes available so that you can enter the starting seed value. You can enter any integer other than 0. The default is 1. When this checkbox is cleared, a different starting seed value is used each time you run FTA calculations. Using the same starting seed value for each calculation run provides for duplicating results.
|
||
|
Use variance reduction
|
When selected, variance reduction is used to reach required accuracy for small probability values. This checkbox is available only when Monte Carlo simulation is selected as the calculation method for time dependent or Lambda Tau trees. Instead of solving the original problem, this analytical simulation method solves a related problem that produces results with a smaller variance. It then uses an adjustment factor to obtain results for the original problem.
|
||
|
The properties under this heading are available only if Monte Carlo simulation is selected as the calculation method for a time dependent or Lambda Tau Tree. They are also available if Uncertainly analysis is selected on the > page for any tree type. Both Monte Carlo simulation and uncertainty analysis use simulation techniques.
|
|||
|
Confidence interval type
|
The type of confidence bounds to use for calculations. Choices are:
• None. When selected (default), no bounds are calculated.
• Lower. When selected, only lower bounds are calculated.
• Upper. When selected, only upper bounds are calculated.
• Two Sided. When selected, both upper and lower bounds are calculated.
|
||
|
Confidence level
|
The confidence level percentage to use. If None is selected as the interval type, this property is unavailable. For all other interval types, the default is 95. However, you can specify any value from 0 to 100 (not inclusive). The one or more bounds specified by the interval type are computed for this level and shown in the results.
|
||
|
Use parametric method
|
Indicates whether to use the parametric method.
• When selected (default), the confidence bounds are calculated using an appropriate distribution. The parameters are obtained from the results mean and standard deviation. For example, for failure rate, the bounds are obtained using the lognormal distribution.
• When cleared, the confidence bound are obtained using the sorted values of the results obtained at different simulation iterations.
|
||
|
Other
This heading and its single property are shown only for SAE trees. They are hidden for time dependent and Lambda Tau trees.
|
|||
|
Turn off all monitors
|
Indicates whether all monitors are turned off for calculations. When cleared (default), monitors are taken into account in calculations. When selected, calculations ignore monitors. Unlike the previous checkbox, this one is not affected by the Qualitative analysis only checkbox. Additionally, your selection is saved for future calculations of SAE trees.
|
||