Managers of production activities need to control the risks presented by various disruptions, for example, risks to their production schedules. Dismantlement and disassembly activities are subject to similar risks. Sandia's efforts in risk management therefore include a program for risk analysis of disruptions.
Disruption has many causes, such as equipment failure, industrial accidents, contamination (e.g., PCBs), natural phenomena (e.g., earthquake), sabotage, or regulatory or political actions (e.g., requiring a permit). The techniques we have developed provide information that allows a manager to reduce the risk of a serious disruption, in a cost-effective manner, by reducing the probability of occurrence (e.g., through strategic redundancy) or the consequences of an event (e.g., by inventories of key components).
We developed a model to assess the vulnerability of a production system to a disruption of unacceptable duration. The model was tested at one production facility, modified, and then applied to the entire nuclear weapons production complex. Two major products from this research were: (a) the methodology and models for performing disruption risk analysis, and (b) the results of such an analysis of the production complex for a particular time. The results included a list of choke points and critical links throughout the complex (including seven production plants and three design laboratories). Critical links are those processes that, if lost or disabled for longer than their critical times, would cause the entire system to fail. System failure is defined as the inability to be back on schedule (in this case, of weapon deliveries) within a given length of time (in this case, one year). Critical time is the maximum length of time that a process can be down and not cause system failure. The model supports risk management by easily showing the effects of increasing the capacity of choke points, increasing recovery times by means of inventory, and providing redundancy of key facilities or equipment. The model can also be used to identify potential industrial sabotage targets and to optimize risk reduction actions. Currently, this methodology is being modified to be incorporated into a dismantlement process model.
The methodology is based on a two-phased approach. First, a network of the production processes is constructed, and process data are collected (e.g., process time, yield, capacity, average inventory). Using those data, the network analytical model then calculates a critical time for each process and also identifies choke points. Next, using fault-tree techniques, each process is analyzed to determine what could cause the process to be shut down for longer than its computed critical time. In addition, the probability of such an occurrence is estimated. The analysis identifies the critical links in the complex and determines the effect of the loss of any critical links. The accompanying figure depicts this two-phase process.
To our knowledge, no other comparable model exists that can analyze the disruption vulnerability of a complex production system. When applied to the production of nuclear weapons, managing disruption risk was closely linked to national security. When applied to nuclear weapon dismantlement, disruption risk management is closely connected with political and social issues, such as supporting treaty agreements and presidential commitments.
Future Work
The work in the immediate future is related to the Pantex process model. It would be applicable to a future consolidated nuclear weapons complex, but no firm plans have yet been made.
For further information, contact:
Edwin A. Kjeldgaard
Sandia National Laboratories, MS-0718
Albuquerque, NM 87185-0718
Phone: (505) 845-8011
e-mail: eakjeld@sandia.gov
or
Michael G. Vannoni
Sandia National Laboratories, MS-1373
Albuquerque, NM 87185-1373
Phone: (505) 844-1789
e-mail: mgvanno@sandia.gov
Submitted October 1995 Layout design by Wanda Mar.
Submitted October 1995