Restore©: Modeling Interdependent Repair/Restoration Processes

Sample output from Restore© showing a completion time distribution and its corresponding cumulative probability function, simulation statistics, and input areas for controlling the appearance of the graph.

Modeling Complex Systems
Restore, developed by Argonne National Laboratory, models complex sets of steps required to accomplish a goal, such as repairing a ruptured natural gas pipeline, when the time required to complete a step may be uncertain. For example, external conditions (i.e., the time of day, weather, availability of crew) may affect one or more of the steps required to accomplish a goal. Therefore, “nature” can influence which steps are taken and the time needed to complete each step.

Restore allows a user to estimate the time and cost (which may be uncertain) needed to achieve an intermediate stage of completion, as well as overall completion. Further, Restore can model workarounds and a simultaneous complete repair to obtain a distribution for the earliest time until service can be restored. Restore provides either a temporary solution via the workaround or complete repair, whichever comes first.

The tool also identifies the “most active path” through the network of tasks. This extremely important information helps a user gain insight into the most effective ways to speed up (or slow down) progress.

Finally, unlike other project planning and risk analysis tools, Restore provides an intuitive, graphical, and object-oriented environment for structuring a model, setting its parameters, and viewing results.

Using Restore
Argonne staff members have used Restore to analyze many failure scenarios. Restore’s data library of applications was developed in collaboration with subject-matter experts on natural gas, water, electricity, and telecommunications infrastructures.

A new user with as little as 4–6 hours of training can use the tool to modify an existing data file in the library to better represent an ongoing failure condition. With experience and in collaboration with subject-matter experts, users can build new data files for new failure scenarios.

Using Restore is simple. The challenge, however, is to identify required activities and specify their completion time distributions. Argonne staff always relies on the experience of infrastructure professionals to answer this challenge. Argonne is uniquely skilled and experienced in obtaining the “right” information.

Restore runs in a few seconds or less, even when restoration processes comprise many activities and complex relationships. Output graphs summarize probability distributions for overall and intermediate completion times. A wide variety of sensitivity analysis features help the analyst develop insights into the restoration process and how to improve it.

Typical Restore© dialog box for specifying an activity. A user specifies a rule for starting the activity and a probability distribution for the completion time.

Applications of Restore
The Department of Homeland Security has used Restore to estimate times needed to re-establish service after disruptions to natural gas supplies due to transmission pipeline ruptures, compressor station failures, city gate failures, and gas-water separators at underground storage facilities.

Work is also being performed for the US Environmental Protection Agency to estimate service restoration times following damage to drinking water distribution systems.

For more information, contact:

Michael Samsa (msamsa@anl.gov) or Ronald Whitfield (rgwhitfield@anl.gov)
Decision and Information Sciences Division
Decision Support and Risk Management Group
Argonne National Laboratory
9700 S. Cass Avenue, Bldg. 900
Argonne, IL 60439, USA