A DPL Fault Tree can have multiple fault trees that interact to perform one overall task. You can implement this by creating a single node that represents an entire fault tree. This node is called a module. Modules behave in the same way as gates: it has one output and any number of inputs. The output of a module is its probability of occurrence. The operations in a module are a combination of the operators in a fault tree, hence a module is like a custom defined gate.
Incorporating a module into a decision model or another fault tree is referred to as embedding a module. There is no limit on either the number of modules you can embed or the number of times you can embed the same module. Embedded modules appear as orange hexagonal nodes in both fault trees and influence diagrams.
Embedding a Fault Tree in another Fault Tree
A fault tree can be embedded as a module into another fault tree. This serves two basic purposes:
- it allows you to define a new binary event operator (i.e., custom gate) that you can use several times, and
- it improves the readability of your tree by separating subroutines from the main tree.
- In the fault tree in which you wish to embed the module, select FAULT TREE | Node | Module Ref (
). DPL displays the Select Module dialog.
- By default other modules in the open Workspace are displayed in the dialog.
- Use the browse button to embed a module from another Workspace.
- Select the module you wish to embed.
- Place the module in the fault tree.
- If you'd like the module to initialize the probability value of a binary event in the active fault tree, connect the module to that binary event.
- Create a binary event (or value) as needed.
- Connect the node to the embedded module. DPL displays the Select Module Input dialog. The dialog lists the inputs of the fault tree represented by the module which are not yet initialized.
- Select the input to which you wish to link the node.
- Click OK.
- Repeat the above steps until all inputs to the embedded module are linked.
Embedding a Fault Tree in a Decision Model
Embedding a fault tree in a decision model allows the fault tree to provide a probability input to the decision model. If you wish, you can also have the decision model provide inputs to the fault tree.
To embed a fault tree in a decision model
- While looking at the influence diagram of the decision model in which you wish to embed a fault tree, select INFLUENCE DIAGRAM | Node | Add | Fault Tree Module. The Select Module dialog appears.
- Use the browse button to embed a module from another Workspace or choose "(local)" in the drop-down list to embed a module from the current Workspace.
- Select the module you wish to embed.
- Place the node in the influence diagram. The Node Definition dialog appears.
- By default, the node is named the same as the fault tree it references. If you wish, you may change the name.
- Click OK.
To link the output of the embedded module to the decision model
- Double-click on the node which will use the input from the embedded module. DPL displays the data tab of the Node Definition dialog.
- Click on the branch on which the input is to be used.
- Click on the variable button, then select the name of the embedded module. Click OK.
- You may wish to draw an arc to this node from the embedded module node for clarity.
To have the decision model provide inputs to the fault tree:
- Embed a module as above.
- If you have not already created nodes to serve as inputs to the embedded module, create them now.
- Draw an influence arc from one of the nodes to the embedded module. DPL displays the Select Module Input dialog. The dialog lists the inputs of the fault tree represented by the module which are not yet initialized.
- Select the input to which the decision model node will provide data.
- Click OK.
- Repeat the above steps until all inputs to the fault tree are linked to the decision model.
Versions: DPL Fault Tree
See Also