How can you model extreme events and predict them before they happen?
One of the world’s foremost experts in probabilistic mechanics and stochastic mechanics plans to show you.
Mircea D. Grigoriu, Ph.D., F.EMI, will share findings from his latest paper, “Data-based estimates of extreme rare events in dynamical systems,” as a keynote speaker at the Engineering Mechanics Institute’s upcoming 2021 conference.
The professor of civil and environmental engineering at Cornell University has published seminal textbooks on structural vibrations and stochastic calculus and has written more than 200 papers.
A long-time active member of the engineering mechanics community, Grigoriu took time out to answer a few questions from Civil Engineering Source about himself and the paper he’ll be speaking about:
Civil Engineering Source: What is your civil engineering specialty and why are you passionate about it?
Prof. Grigoriu: My work is focused on the quantification of uncertainties in material properties and of actions on physical systems and the estimation of the performance of these systems. My interest in these types of problems is stimulated by their diversity and the multitude of tools needed for solution, for example concepts of mechanics, mathematics, probability and statistics.
In layman’s terms, how would you describe what the paper you’ll be speaking on is about?
Extreme events such as tsunamis, earthquakes, floods and violent vibration in aircrafts, can have severe consequences. Yet, there are no practical and accurate methods for modeling extreme events and assessing their consequences. Standard statistical methods cannot be used because of insufficient observations (extreme events are rare/infrequent), and Monte Carlo simulation or other numerical methods are computationally prohibitive. The paper presents a practical method for characterizing extreme events and estimating their consequences. Although the method uses specialized concepts of stochastic processes, its application is conceptually simple. In our test cases, the method provides accurate results.
Why did you feel the need to write a paper on this topic? What did you hope to achieve?
I believe that there is a need for a practical method for characterizing extreme events and their effects on physical systems.
Why is this topic well-suited to the audience of civil engineers that will be attending the EMI conference?
Structural systems are subjected to a broad range of actions/loads of which some are extremes with potential catastrophic consequences. System reliability under extreme events is a topic of great interest to the EMI community, which has been studied for many years and continue to be studied in our days.
How is your method able to reasonably predict extreme events that happen randomly and unexpectedly?
The proposed method is based on (1) physically consistent representations of actions on physical systems which are fitted to observations and (2) accurate and efficient method for estimating system reliability under large rare events.
The EMI Conference will be held virtually May 25-28. See the conference website for more details and to register.
An excellent reference on this subject is the article by E.J. Gumbel in Engineering News Record Vol. 134 pp 833-837, 1945. Floods Estimated by the Probability Method. The related pertinent equations can be readily incorporated into a computer program and may be applicable to other rare events besides flood magnitudes and recurrences.