Determining Resilience Thresholds for Nuclear Power Plants
Transcription
Determining Resilience Thresholds for Nuclear Power Plants
DETERMINING RESILIENCE THRESHOLDS FOR NUCLEAR POWER PLANTS Pamela Fran Nelson UNAM 27 abril 2015 Objective Develop a robust tool for monitoring organizational resilience. Implement the tool at a nuclear power plant to measure and follow human performance events including organizational performance. Track trends in order to identify time periods where the likelihood of consequential events increases. occurrence of Scope Human and Organizational Errors Pre-initiators Maintenance Post-initiators Preventive Corrective Calibration Operations Surveillance Normal Off-Normal Typical organization process flow of work activities at an NPP. Corrective Action Program (CAP) Low level CAP (“Soft” and “hard”) High level CAP 10,000 – 15,000 per year Condition Reports SCAQ Significant Condition Adverse to Quality (includes LERs) CAQ-L1 Condition Adverse to Quality at a station level CAQ-L2 Condition Adverse to Quality at a department level CNAQ Condition Not Adverse to Quality Events per week Tool: Probability of an SCAQ depending on the number of days since the last occurrence Tool: Probability of an SCAQ depending on the number of CRs since the last occurrence Organizational factors Responsible organizations or departments CRs Actions Interdepartmental factors Actions generated for other departments Communications between departments Interdepartmental factors Departments creating CRs Departments receiving actions Number of SCAQs for departments responsible for more than one SCAQ. Number of actions from SCAQs Resilience Scheme Compensatory measure Materials Science Analogy: Stress-Strain Curve Resilience curve and threshold Resilience curves per year Conceptual Performance Indicator Time series: 2005-2015 Forecasting Summary A proposed leading performance indicator was developed to predict increased likelihoods of consequential events Organizational resilience can be modeled based on a stress-strain analogy Data represents actual plant events and operating history for consequential events Plant to plant variability would be expected and plant specific analysis is needed to determine resilience thresholds Additional work is needed to collect more data to determine the level of robustness of the approach and to begin developing organization specific consequential events and associated organizational specific thresholds Distribution fitting