Our society is increasingly reliant on systems and interoperating collections of systems, known as systems of systems (SoS). Our national security is built on SoS, such as Army brigades, airport security, and nuclear weapons security. These SoS are often subject to changing budgets, changing missions (e.g., nation building, arms-control treaties), changing threats (e.g., asymmetric warfare, terrorism, WMDs), and changing natural environments (e.g., climate, weather, natural disasters). Can vital SoS adapt to these changing landscapes effectively and efficiently? This paper describes research at Sandia National Laboratories to develop metrics for measuring the adaptability of SoS.Wereport thatwecouldnotfindasingle or absolute adaptability metric, in large part duetolackof general objectives orstructures of SoS. However, we do report a set of metrics that can be applied relatively, plus a method for combining the metrics into an adaptability index, a single value that can be used to compare SoS designs. We show in a test case that these metrics can distinguish good and poor performance under a variable mission space and an uncertain threat environment. The metrics are intended to support a long-range goal of creating an analytic capability to assist in the design and operation of adaptable systems and SoS.
Our society is increasingly reliant on systems and interoperating collections of systems, known as systems of systems (SoS). These SoS are often subject to changing missions (e.g., nation- building, arms-control treaties), threats (e.g., asymmetric warfare, terrorism), natural environments (e.g., climate, weather, natural disasters) and budgets. How well can SoS adapt to these types of dynamic conditions? This report details the results of a three year Laboratory Directed Research and Development (LDRD) project aimed at developing metrics and methodologies for quantifying the adaptability of systems and SoS. Work products include: derivation of a set of adaptability metrics, a method for combining the metrics into a system of systems adaptability index (SoSAI) used to compare adaptability of SoS designs, development of a prototype dynamic SoS (proto-dSoS) simulation environment which provides the ability to investigate the validity of the adaptability metric set, and two test cases that evaluate the usefulness of a subset of the adaptability metrics and SoSAI for distinguishing good from poor adaptability in a SoS. Intellectual property results include three patents pending: A Method For Quantifying Relative System Adaptability, Method for Evaluating System Performance, and A Method for Determining Systems Re-Tasking.
In a total-system performance assessment (TSPA), uncertainty in the performance measure (e.g., radiation dose) is estimated by first estimating the uncertain y in the input variables and then propagating that uncertain y through the model system by means of Monte Carlo simulation. This paper discusses uncertainty in surface infiltration, which is one of the input variables needed for performance assessments of the Yucca Mountain site. Infiltration has been represented in recent TSPA simulations by using three discrete infiltration maps (i.e., spatial distributions of infiltration) for each climate state in the calculation of unsaturated-zone flow and transport. A detailed uncertainty analysis of infiltration was carried out for two purposes: to better quantify the possible range of infiltration, and to determine what probability weights should be assigned to the three infiltration cases in a TSPA simulation. The remainder of this paper presents the approach and methodology for the uncertainty analysis, along with a discussion of the results.
The abstraction model used for seepage into emplacement drifts in recent TSPA simulations has been presented. This model contributes to the calculation of the quantity of water that might contact waste if it is emplaced at Yucca Mountain. Other important components of that calculation not discussed here include models for climate, infiltration, unsaturated-zone flow, and thermohydrology; drip-shield and waste-package degradation; and flow around and through the drip shield and waste package. The seepage abstraction model is stochastic because predictions of seepage are necessarily quite uncertain. The model provides uncertainty distributions for seepage fraction fraction of waste-package locations flow rate as functions of percolation flux. In addition, effects of intermediate-scale flow with seepage and seep channeling are included by means of a flow-focusing factor, which is also represented by an uncertainty distribution.