ESFM Presentation Days: Monday December 9 - Tuesday December 10, 2019
The Presentation Days of the PhD Program in "Earth Science, Fluid-Dynamics, and Mathematics. Interactions and Methods" will take place at University of Trieste, on Monday, December 9 and on Tuesday December 10, 2019 (via Weiss 1, Palazzina C, ground floor, Marussi lecture hall). The (provisional) program is attached.
All PhD students will present the progress of their research during the last academic year.
The program also includes four plenary lectures held by members of the Faculty Board in the main fields of the ESFM PhD course: Mathematics, Fluid-Dynamics and Earth Science.
Prof. Vincenzo Armenio
Tackling Turbulence Using HPC
Turbulence is ubiquitous of a large number of fluid dynamic processes in physics and engineering. From the thousand-kilometer geophysical scales, to the flow within a pump or around a propeller, turbulence drives mixing and dissipation mechanisms. In spite of that turbulence is still one of the unsolved problems of classic physics, basically due to its own intrinsic non-linear and multiscale features. Only recently, the use of numerical algorithms implemented over high performance computers has allowed to explore the physics of the turbulent flow in a number of archetypal cases significant for practical purposes at a level not affordable using standard experimental techniques. In the seminar, first we give some fundamentals of turbulence dynamics to highlight the difficulties associated to its study and successively we show some examples where HPC is successful employed to solve complex problem of importance in engineering.
Prof. Luca Bortolussi
Machine Learning and Logic for Model-Based Design and Testing
Model based design and testing are two key tools in the modern engineering toolbox, where mathematical models are used as a proxy of the real system to try different design schemes and to test for malfunctioning. We will consider, in particular, requirements formalized in terms of logical properties, expressing e.g. safety of the system, and discuss how to use machine learning to build surrogate models of the dependency of such properties on model parameters. Crucially, these surrogate models also contain information about their uncertainty in predicting the correct value, which enables their use for several tasks: Design, where one or more optimal models are generated, synthesis, where the set of models satisfying the requirements is identified, counterexample generation of scenarios falsifying the desired properties, and real-time monitoring, where the learned model is used to warn about potential troublesome situations. We will also comment on challenges, particularly control and openness, and future research perspectives.
Prof. Carla Braitenberg
Earthquake Related Ground Movements Sensed by Tilt and Strainmeters and Sentinel 1 – Homage to Maria Zadro
Maria Zadro (1933-2018) was Professor of Physics of the Earth at the University of Trieste, academically descendant of Prof. Antonio Marussi, Geodesist, after whom the Marussi tensor is named. Maria Zadro installed the tilt and strainmeter stations of Friuli Venezia Giulia with support from the Civil Protection Agency and INGV. Today a further tiltmeter station is active at the border between the Province of Friuli Venezia Giulia and Veneto, at the Cansiglio karst plateau. The presentation introduces the tilt and strainmeter observations of many phenomena over a broad spectrum of frequencies and amplitudes, that include earth tides, ocean tidal loading, atmospheric loading, hydrologic pressure changes, pre-, co- and postseismic fault movement, silent earthquakes, earthquake precursors. The crustal movement is also observed with the Sentinel 1 satellite, that actively sends electromagnetic waves and measures the phases and amplitudes of the waves scattered from the ground. A successive passage of the satellite gives a repeat image which is analyzed in the phase changes to detect ground movement along line of sight. The satellite observations are coarser, but are global and have a complete areal coverage. The higher resolution of the tilt-strainmeter observations help identifying deformation signals in the interferometric signal that otherwise would be hidden in the noise. The talk will discuss the terrestrial and remotely sensed deformation, setting them in relation to the scientific achievements of Maria Zadro.
Prof. Fred Kucharski
The El Nino phenomenon and other tropical sea surface temperature modes have global impacts on climate anomalies. The basics of the atmospheric teleconnection mechanisms behind these influences will be reviewed in this presentation step by step. Then some examples will be presented for atmospheric teleconnections on various time scales.