please check the attached file for required workthank you
INEEL/EXT-03-01117 INEEL/EXT-03-01117 Novel Threat-Risk Index Using Probabilistic Risk Assessment and Human Reliability Analysis February 2004 Idaho National Engineering and Environmental Laboratory Bechtel BWXT Idaho, LLC INEEL/EXT-03-01117 Novel Threat-Risk Index Using Probabilistic Risk Assessment and Human Reliability Analysis Martin M. Plum David I. Gertman George A. Beitel Jerry H. Phillips James A. Vail Ronald L. Boring Patrick H. McCabe Kyle S. Staples Jeffrey C. Joe David H. van Haaften Robert E. Polk Garrold L. Sommers February 2004 Idaho National Engineering and Environmental Laboratory Idaho Falls, Idaho 83415 208-526-0042 Prepared for the U.S. Department of Energy Office of Nuclear Energy, Science and Technology Under DOE Idaho Operations Office Contract DE-AC07-99ID13727 Novel Threat-Risk Index Using Probabilistic INEEL/EXT-03-01117 Risk Assessment and Human Reliability Analysis Page ii of viii ABSTRACT In support of a national need to improve the current state-of-the-art in alerting decision makers to the risk of terrorist attack, a quantitative approach employing scientific and engineering concepts to develop a threat-risk index was undertaken at the Idaho National Engineering and Environmental Laboratory (INEEL). As a result of this effort, a set of models has been successfully integrated into a single comprehensive model known as Quantitative Threat-Risk Index Model (QTRIM), with the capability of computing a quantitative threat-risk index on a system level, as well as for the major components of the system. Such a threat-risk index could provide a quantitative variant or basis for either prioritizing security upgrades or updating the current qualitative national color-coded terrorist threat alert. Novel Threat-Risk Index Using Probabilistic INEEL/EXT-03-01117 Risk Assessment and Human Reliability Analysis Page iii of viii EXECUTIVE SUMMARY This document is a summary report of a research and development project conducted by the Idaho National Engineering and Environmental Laboratory (INEEL) to improve the current state-of-the-art in alerting decision makers to the risk of terrorist attack. A quantitative approach employing scientific and engineering concepts and utilizing a set of models has been successfully integrated into a single comprehensive model with the capability of computing a quantitative threat-risk index on a system level, as well as for the major components of the system. Several novel improvements are included in the INEEL approach, including emphasis on the probability of attack (perhaps the most difficult of all parameters to estimate), systems consideration, and integration of the behavior of humans with engineered systems. In contrast, most other contemporary approaches use qualitative information to populate generalized risk matrices and either apply the risk at only the system level or at the subsystem level. Furthermore, in lieu of a good estimate of probability of attack, other approaches typically use a qualitative value such as high, medium, or low. Without quantifying high, medium, or low, the safe approach is to apply high to most conditions and defend against most scenarios that could occur. That approach reduces to vulnerability assessment rather than risk assessment. The solution to vulnerabilities is to apply a fix, often very precise, but far in excess of the expected value of the potential loss. As a result, precious National resources are not focused to deal with risk where it may be the highest. The INEEL improvement is a single comprehensive model, the Quantitative Threat-Risk Index Model (QTRIM), with the capability of computing a quantitative threat- risk index on a system level, as well as for the major components of the system. The INEEL QTRIM is a unique approach in that it integrates models of human response, physical and engineering behavior, risk, and economic engineering to more precisely identify the risk associated within or between particular target classes. QTRIM consists of five individual models: 1. The targeting model, consisting of a model to estimate the probability of terrorist attack; 2. The human reliability model, consisting of a series of human response logic models; 3. The physical system model, which when applied to the hydroelectric system, as in this study, includes dam failure models, a flood inundation model, and a loss of service model; 4. The probabilistic risk analysis model, SAPHIRE, an INEEL event tree/fault tree software program; 5. The consequence/loss model, providing a socioeconomic account of the attractiveness of potential targets, or, alternatively, the potential economic damage from an attack. Data for models 1 and 2 are gathered from human factors sources of information, such as procedures, interviews, staffing policy, equipment to be operated, distance to be traversed, motivation, target selection based upon target value, opportunity, ease of access, etc. These factors are used in conjunction with traditional human engineering, human reliability, economics, and fault tree modeling. Model 3 is a set of deterministic physical system models that describe the physical response of a system to an impulse (such as an attack). Model 4 is an INEEL-developed software workstation that integrates the collective information into events, event sequences, and end states of various consequences. It is based on underlying fault tree analysis. The SAPHIRE program provides for uncertainty analysis and dependency calculation among sub-events. Model 5 is a newly developed socioeconomic response model that uses probabilistic representations to address the cost associated with certain failure scenarios and end states; it also helps to establish the attractiveness of the target to hostile entities. Thus, it strongly supports Novel Threat-Risk Index Using Probabilistic INEEL/EXT-03-01117 Risk Assessment and Human Reliability Analysis Page iv of viii model 1, as well as indicates where resources should be focused to mitigate the results of terrorist- inspired events. The major advances associated with the proposed QTRIM are a better approximation of a threat and prioritization of a specific infrastructure target, integration of human factors and human reliability information, and consideration of multiple infrastructure targets, including multiple target dependency effects that could greatly enhance the consequences of a terrorist attack. QTRIM calculates the probability of attack, the success probability for that attack, estimates dollar values for all levels of ensuing damage, and considers an entire infrastructure system, all with automatic prioritization. All parameters are integrated into the event tree/fault tree. This allows quantitative comparisons and consequently identifies and prioritizes risk reduction actions across the system under study. The probability-of-attack model, if verified and validated, could be used to predict the probability of attack on specific facilities, based on available terrorist resources and on outcomes expected by the terrorist. Preliminary application of the model showed that for the sample dam system selected in this study, that the probability of attack and failure from terrorist attacks might be so low that very few security improvements are warranted above normal best practices for an industrial facility (the reliability of the model is not, at this time, at a stage that such predictions can be acted upon). While the model is at the developmental stage, the primary difference between the development stage and final is the quality of the facility specific input data. Refinements to the model are expected to result in minor changes from current values; nevertheless it already provides estimates of probability of attack that correspond with published terrorist activity, and the physical and economic models predicted values within 10% of officially used values for the same parameters. For a proof-of-principle demonstration, the INEEL developed an application using a system of dams and hydroelectric facilities. An integrated team of systems engineers, risk analysts, human factors specialists, physical system modeling and simulation experts, economic engineers, hydro-system experts, national security experts, and transmission and grid systems personnel worked with U.S. Army Corps of Engineers and the U.S. Bureau of Reclamation personnel, who provided supplemental information, including emergency action plans and access to key personnel during the course of the study. QTRIM represents a highly successful and well-received study and proof of principle of using a software modeling system (SAPHIRE) in conjunction with an integrated systems approach to identify and quantify risk. The utility of a quantitative approach in replacing or supplementing the current qualitative approach is readily apparent. Huge cost savings can be realized by identifying, region-by-region and then state-by-state, which targets are at risk, what is their priority, and where resources should be directed. State and National Legislatures and security agencies can potentially save millions in taxpayer dollars by limiting spending to the higher priority threats identified by this risk analysis approach. Although QTRIM was demonstrated on a system of dams, the basic principles and models are independent of a specific infrastructure, and could be readily applied to many other systems. Component models, such as the targeting model, may be used independent of QTRIM. Novel Threat-Risk Index Using Probabilistic INEEL/EXT-03-01117 Risk Assessment and Human Reliability Analysis Page v of viii CONTENTS Abstract ......................................................................................................................................................... ii Executive Summary .............................................................................................