Potential enhancements to addressing programmatic risk in the tank waste remediation system (TWRS) program

Pacific Northwest National Laboratory (PNNL) conducted a Tank Waste Remediation System (TWRS) Risk Management methodology development task. The objective of this task was to develop risk management methodology focused on (1) the use of programmatic risk information in making TWRS architecture selection decisions and (2) the identification/evaluation/selection of TWRS risk-handling actions. Methods for incorporating programmatic risk/uncertainty estimates into trade studies are provided for engineers/analysts. Methods for identifying, evaluating, and selecting risk-handling actions are provided for managers. The guidance provided in this report is designed to help decision-makers make difficult judgments. Current approaches to architecture selection decisions and identification/evaluation/selection of risk-handling actions are summarized. Three categories of sources of programmatic risk (parametric, external, and organizational) are examined. Multiple analytical approaches are presented to enhance the current alternative generation and analysis (AGA) and risk-handling procedures. Appendix A describes some commercially available risk management software tools and Appendix B provides a brief introduction to quantification of risk attitudes. The report provides three levels of analysis for enhancing the AGA Procedure: (1) qualitative discussion coupled with estimated uncertainty ranges for scores in the alternatives-by-criteria matrix; (2) formal elicitation of probability distributions for the alternative scores; and (3) a formal, more structured, comprehensive risk analysis. A framework is also presented for using the AGA programmatic risk analysis results in making better decisions. The report also presents two levels of analysis for evaluation and selection of risk-handling actions: (1) qualitative analysis and judgmental rankings of alternative actions, and (2) Simple Multi-Attribute Rating Technique (SMART).