- HOME
- Activities
- Lecture Series
- The 11th GMSI International Lecture: Dr. Romney B. Duffey, Moderator: Prof. S. Sakai(Department of Mechanical Engineering)
The 11th GMSI International Lecture: Dr. Romney B. Duffey, Moderator: Prof. S. Sakai(Department of Mechanical Engineering)
2013.01.29
Lecture Room 2-31A, 3F Engineering bldg. #2, Hongo Campus
Dr. Romney B. Duffey
Ph.D, B.Sc., FASME, MemINEA
Principal and President DSM Associates Inc.
Title:Risk in Today’s World
Schedule: January 29 (Tue) -30 (Wed) , February 1(Fri) 2013
Place: Lecture Room 2-31A, 3F Engineering bldg. #2, Hongo Campus
<Summary>
How do you understand and manage your own risk? Risk exposure and risk management are key issues in today’s personal,
industrial, financial areas. Despite uncertainty, the purpose and goal must be to predict all the errors and accidents that
occur in modern technological society, and hence to reduce risk and manage safety. We explicitly recognize the dominant
contribution and involvement of humans in risk-taking in modern technological society. The rapid pace of change, open social
communication and global reach make events more visible and people fearful, and open to risk and loss. We discuss recent
major unanticipated disasters due to human failures in making key decisions include panic from global financial crises,
unexpected loss of life and property due to floods and tsunamis, perceived environmental damage from offshore oil spills and
reactor meltdowns, and the fear of terrorist attacks, all the way to the technological failures in our everyday
technological systems like bridges, trains, aircraft, computers and industrial plants. The role of regulation, emergency and
government decision-making are also examined.
In the lectures, we address multiple modern systems for which risks exist or are perceived for both modern societies and
their people. We must predict both the occurrence rate, even for rare events, and the influence of the human on the
initiation, development, decisions and learning processes, as well as recovery and emergency corrective actions. This
development links such diverse areas as cognitive psychology, statistical mechanics, information theory, sampling methods,
and physical reasoning, and includes the dynamic and dominant contribution from human decision making and learning. We also
examine the issues of risk quantification, tolerable risk, and the definition of acceptable risk in modern society. We
develop and demonstrate a general accident theory, based on the Learning Hypothesis, combining data review, physical
analysis, mathematical equations, statistical reasoning, and theoretical validation. We review and re-interpret about 200
years of available data covering millions of outcomes and one-off rare or extreme events. To form a firm predictive basis,
we must describe the indescribable, and simplify the complex. We illustrate and justify these ideas and concepts by showing
the common causes, using data and many examples of technological system failures from multiple areas of human endeavor. We
include major “named” events like Fukushima, Space Shuttle, Deepwater Horizon, Texas City, Toulouse AZT, Concorde, Great
Northeast Blackout and Superstorm Sandy, plus Missile Defense efficiency, and also the everyday outcomes like aircraft
near-misses, oil rig incidents, medical errors, railway derailments, auto accidents, bridge collapses … indeed incidents
covering the whole spectrum of modern technology. This lecture series describes in detail the new approach to risk
measurement and management, based on models and ideas that are derivable from, and testable against available world data.
<Program>
◆ January 29 2013 (Tue)
13:30 -15:00 Risk in Society- the definition and the paradoxes
Evidence is given from multiple industries and applications demonstrating the risk exposure and effects of the universal
human contribution. Supporting the Learning Hypothesis, we learn from our mistakes. Case studies are derived from power
blackouts, aircraft near misses, missiles and rocket interceptions, oil spills and multiple other industrial and engineering
systems. The definition of risk is discussed based on uncertainty, probability and consequences. Derivation and
demonstration of the Human Bathtub that exists for predicting the probability of error occurrence or observance, utilizing
the ideas from reliability engineering Demonstrated consistency with human skill acquisition and cognitive psychological
concepts for learning, such as the Ohlsson Theory of Error Correction, and with Petroski’s model of the role of failures in
Engineering. The fundamental risk paradoxes are given. Selected events illustrate the role of management, political, and
regulatory actions arising because of human failings and fallibility. Q&A.
◆ January 30 2013 (Wed)
13:30 -15:00 Risk Measures and prediction- rare events and
the universal learning curve
We examine both frequent examples, like medical errors, auto and industrial accidents, and also rare events, like global
financial crises and major floods. Derivation is given of the predictive equations for the event or system failure rate
based on the accumulated experience implied by some 200 years of data, and the implications of claims for “zero failures”
and of “never happening again”. Comparison is made with other work on probabilistic safety, human error rates, and
“unknown unknowns”. Recommended “best” values for error and event prediction are discussed. Introduction to sampling
analysis and the treatment of complexity and randomness is then explored to provide a deeper understanding of uncertainty.
Extension to human reliability, probabilistic safety analysis concepts and utilization of Bayesian reasoning and prediction.
Detailed case studies from oil spills, near misses and traffic accidents. Q&A.
◆ February 1 2013 (Fri)
10:30 -12:00 Risk Management and Safety Culture ? societal attitudes,
practical approaches and the influence of technological change
Comparisons to conventional ideas of “tolerable risk” and safety regulations Estimation of objective measures for “safety
culture” and to Safety and Risk Management Systems effectiveness, deriving applications for risk prioritization. Extension
and challenges related to perception, fear of the Unknown, and the prediction of as yet unobserved outcomes. Implications
of the learning paradox, and of future societal, personal, business, management, financial and technological risk, and the
avoidance of risks and liabilities in modern technological society. Q&A.
Attached file: The_11th_international_lecture.pdf