Downtime decision tree

We have seen that the simulation core generates a series of events. After each event, the simulation core will look at the corresponding system and will evaluate if it has to adjust the corresponding function level (indicating increasing or decreasing downtime). For each Part, the function level is assessed. Since a Part is a ‘child’ of a system, each ‘child’ will request its ‘parent’ to reconsider its function level. Parent objects then perform the evaluation based on the inputs from all of its children. This mechanism uses the decision tree described below. E.g. if a child reports that its function level is adjusted and is lower than the parent’s function level, the parent will copy that function level, and will record a downtime of “new function level” x “calculated downtime”. In this way, the simulation core expresses the situation that an asset as a whole will be down if at least one of its parts is down.

Down and up events

Events are the trigger to adjust the function level. An event can signal the start of an action, the end of an action, or the start of a new action, while another action is still active, or the end of an action while another action is also active. The last two situations are examples of coincidental maintenance.

Refer to the figure above. Event 1 is the start of Action 1; this is a down event, signaling the start of downtime for the system. Event 3 is the end of Action 1; this is an up event, signaling the end of downtime for the system. However, meanwhile, Action 2 causes coincidental maintenance. Event 2 is another down event, signaling downtime once again for our system. A system that is already down cannot go down again. So for Segment 2, downtime should only be counted once.

The following decision tree shows the logic the simulation core follows to adjust the function level and to make sure downtime is only recorded once in case of coincidental maintenance.

In the paragraph Example scenarios below, the decision tree will be explained further based on two example scenarios. But first, we need to discuss the scoring mechanism.