Applying QMU to Nuclear Weapon Stewardship

Quantification of Margins and Uncertainty (QMU) is a principle means and metric by which nuclear weapons performance is measured [1, 2]. At Lawrence Livermore National Laboratory (LLNL), Nuclear weapon QMU is predominately a Physics assessment of the performance margin and associated uncertainties of not achieving the designed nuclear performance. To more completely apply QMU to Nuclear Weapon Stewardship, a Systems Engineering or systematic approach is being employed to identify and define requirements, allowing assessment of performance with QMU. This paper describes the basics of QMU and how it is applied below the highest level nuclear performance function down and throughout the nuclear weapon design architecture. The intended audience for this paper is a novice QMU practitioner at LLNL. The application of QMU to the entire set of nuclear weapon functions demands a broad engagement of Science and Engineering disciplines. Using a hierarchical flow-down requirements structure, performance functions and failure modes are identified for all levels of system requirements. There can be multiple failure modes per functional requirement. For each function there is a spectrum of possible outputs that can vary from outright failure to excellent performance. The possible consequences and impacts on the next higher level functions will also vary. Since performance variations exist in a weapon functional hierarchy, the assessment process needs and benefits from a system perspective. Incorporating QMU into assessments using a Systems Engineering approach to stockpile management provides a means to quantify weapon performance risks against a multitude of possible failure modes and from the highest level functions down to the lowest level component functionality. Performance risk management is a core practice of nuclear weapon management. Assessing the likelihood and consequences of the many possible failure modes is done yearly as part of Annual Assessment Reviews (AAR). When annual surveillance discovers an anomaly defect as part of weapon surveillance, investigations are launched and QMU is a key tool for quantifying performance impacts. QMU assessments are separated into two pieces, failure mode assessments and QMU analysis Background While the preponderance of QMU assessment work is classified, the basics of QMU are unclassified. This paper describes QMU, explains why it is important and how it fits into the larger scope of stockpile stewardship activities. Appendix A gives a wide range of analytic approaches for estimating confidence levels through statistical analysis, explains Sandia K factors used to select sample sizes appropriate for reliability and confidence level requirements and finally it surveys first and second order reliability analyses and a variety of computational methods. Appendix B covers the statistics used as the basis for nuclear weapon surveillance sampling quantities. The reference papers are available in the LLNL, NWE share folder . \\wci-cl2\wci\NWE\QMU Nuclear Weapon Stewardship before QMU The original nuclear weapon stockpile stewards benefitted from two key experiences lacking in modern stockpile stewardship: original design development and full scale production. The original physicists, chemists and engineers gained an understanding of the numerous design parameters that affect weapon performance through the experience of building a large nuclear weapon stockpile, and they