QUANTITATIVE RISK ASSESSMENT

By James R. Lynch

The enlightened businessman prepares plans to deal with emergencies. He knows that being ready to deal with disasters--both natural and human-made--can make the critical difference to the bottom line should disaster strike. Detailed emergency planning is particularly important for those businesses that:
•Are just-in-time suppliers, in particular those that are liable to incur large penalty charges when they can not deliver on time due to a production interruption.
•Are supplied by just-in-time vendors, when a vendor's failure to deliver can result in costly production stoppage.
•Operate on tight margins where significant downtime can be crippling.
Serious emergency planning requires: first, specifying the critical aspects of the business--those that really require protection and rapid restoration; second, defining the threats to those critical aspects; and third, developing detailed action plans for prevention, protection and mitigation.
Good emergency planners follow this prescription, but generally do so in a qualitative or deterministic way.
More advanced, comprehensive plans go a step further and apply probabilistic methodologies in such a way as to quantify risks.
Quantitative Risk Assessment (QRA) techniques are routinely used today in situations where the effects of a disaster might be of such magnitude that a higher level of planning detail is required. Examples of operations where QRA techniques are routinely applied included commercial nuclear electric power stations and the Department of Energy's nuclear weapon production facilities.
With the manufacturing sector's recent shift towards just-in-time stocking practices, interruption of production flows can have far reaching and serious consequences.
The failure of a small supplier might cause downstream effects, ones that ripple through an industry where operations occur at facilities sequentially.
The microelectronics industry provides a good example. Consider a chip maker that provides custom microprocessors on a just-in-time basis to a manufacturer of automated machine tools.
A natural disaster that interrupts the microprocessor production line will shut down the machine tool line in turn. QRA's of such operations can help the manager to understand the risks involved, and where to focus capital in an effort to minimize risk.
In order to conduct a QRA of an industrial operation, a considerable amount of research is required. The production operation itself must, of course, be well understood. Environmental factors specific to the manufacturing site must be researched. Here weather history, seismic activity, flooding vulnerability (including seiche and tsunami) and other similar phenomenology must be understood in detail.
Site susceptibility to hazards released from or generated by nearby industrial operations and transportation systems must also be included in the analysis. Accordingly, QRA methodology can be applied to an industrial operation in a six-step process.
The analysis requires: The identification of the critical paths in the production operation. Here a thorough knowledge of processes, supply stock levels, normal and emergency energy supplies, etc., is required.
•The identification of the utilities, supplies and services that are critical in supporting the production operation.
•The identification of the threats, natural and human-made, that can interrupt production directly or interrupt the flow of supporting supplies and services. Development of frequency distributions for these threats is also required.
•The determination of the time required to restore interrupted operations, taking into account emergency plans and capabilities of the company, various levels of government, and individual supply/service providers. Frequency distributions are required since time-to-restore is generally a function of severity of threat. In these distributions, the impact of each level of threat on each operation/supply/service is developed.
•The rank-ordering of the threats to production. Mathematical methods for convoluting the probability matrices are used in this process.
•And, following careful review of the analytic results, preparation of recommendations focused on consequence mitigation. Here both preventive measures and planned emergency actions are considered. Recommendations for prevention and preparedness actions can be ranked in order of cost-effectiveness.
To illustrate this process, an example is useful. Emergency planning managers at an electronics manufacturing firm engaged in just-in-time supply of custom microprocessors were in the process of updating emergency response plans.
In the course of conducting a qualitative review of their operations, they identified a large number of threats with the potential to interrupt an almost equally large number of suppliers and services, all critical to the operation.
The planners determined the following threats and critical supplies and services were of sufficient concern to warrant application of QRA techniques:
Threats: aircraft crash, earthquake, flooding, nearby industrial accident, lightning strike, snow/ice storm, tornado/wind, transportation accident.
Critical Supplies/Services: bottled compressed gasses, bulk chemicals, electric power, natural gas, water, sewer service, site access.
In conducting the analysis, the probability of each of the threats was developed for several levels of threat severity. The effect the various threat levels had on each of the critical supplies/services was assessed. The ability of emergency response groups to restore supplies/services was determined, and last, the threats were rank ordered.
As can be seen, the loss of water supply is the most likely production-interrupting event, followed by loss of site access.
The analysis also provided a rank-order of threats within each supply/service category. As a result of the QRA analysis, many actions were proposed to reduce risk to production.
The two proposed mitigating actions ranked highest in terms of cost-benefit were:

•Construction of an on-site water holding tank with sufficient capacity to supply production for one day. Having such a tank would significantly reduce the overall risk of production stoppage.
• Requesting a neighboring industrial facility to move or reduce quantities of stored chlorine gas. Required evacuation (loss of access) due to an inadvertent toxic gas release at this neighboring facility was found by QRA analysis to be the second most likely cause of loss of production capability. The potential impact of this threat had not been understood prior to the conduct of the QRA analysis.
Generally, in conducting QRA analyses, specific local data can be found that characterize threat levels for a facility. Where local specific data is not available, point estimates can be made based on Federal Emergency Management Agency (FEMA) and other federal, state and local agency generic data.
In assessing the threats, computer programs have been developed to operate on the probability distributions of threats, threat severity, and emergency response capacity. This allows a rapid calculation of rank-orders, and varied ways of quickly displaying the results.
The conduct of a QRA analysis requires the use of professional emergency preparedness and planning analysts.
The development of a QRA for a typical mid-sized manufacturing facility might involve two or three months of analytic effort. However, this small investment in detailed analysis of threats to production will have high payoff in mitigating the effects of potential disasters.

James R. Lynch is a Senior Engineer with Science & Engineering Associates, Inc., of Albuquerque, N.M.

Disaster Recovery World© 1997, and Disaster Recovery Journal© 1997, are copyrighted by Systems Support, Inc. All rights reserved. Reproduction in whole or part is prohibited without the express written permission form Systems Support, Inc.