John Luxat, McMaster University
Department of Engineering Physics
Professor
Hamilton, Ontario
luxatj@mcmaster.ca
Office:
(905) 525-9140 ext. 24670
Expert In
Bio/Research
Best estimate models of physical processes, best estimate plant states, and most probable system configurations and failure events provide the most realistic representation of plant behaviour and consequences during accidents. Deviations from these best estimate conditions can and will occur, res...
Click to Expand >>
Click to Expand >>
Bio/Research
Best estimate models of physical processes, best estimate plant states, and most probable system configurations and failure events provide the most realistic representation of plant behaviour and consequences during accidents. Deviations from these best estimate conditions can and will occur, resulting in uncertainty in the outcome of a best estimate analysis. In order to quantify the variability and uncertainty in the outcome of an accident, it is necessary to identify and characterize the components contributing to uncertainty and evaluate their impact on safety consequences. A primary objective is to define, through the use of an integrated probabilistic approach, the ranges of key plant parameters that assure that safety limits are met at a prescribed level of probability and confidence.
Research in this area investigates approaches to propagation of time-dependent variability and uncertainty in plant response during accidents using dynamic sensitivity analysis and quasilinearization methods. Of particular interest is the representation of nonlinear bifurcation behaviour within a framework of quasilinearized sensitivity analysis since this allows time-dependent solutions for multiple-parameter variations to be generated from a limited set of detailed computer simulations obtained from best estimate computer codes.
Click to Shrink <<
Research in this area investigates approaches to propagation of time-dependent variability and uncertainty in plant response during accidents using dynamic sensitivity analysis and quasilinearization methods. Of particular interest is the representation of nonlinear bifurcation behaviour within a framework of quasilinearized sensitivity analysis since this allows time-dependent solutions for multiple-parameter variations to be generated from a limited set of detailed computer simulations obtained from best estimate computer codes.
Click to Shrink <<