APACM Thematic Conference & IACM Special Interest Conference
Computational Engineering and Science for Safety and Environmental Problems
April 14 th (Mon), Room A+B
Fumihiko Imamura (Tohoku University, Japan)
Improvement of the Tsunami Simulation after the 2011 Tohoku Earthquake
The March 11, 2011 Tohoku earthquake tsunami disaster devastated the Pacific coast of northeastern part of Japan. The large number of casualties more than 19,000 and several types of tsunami impact such as inundation in a large area, destructive force destroying houses, buildings, infrastructures, road, and railways, and change of topography due to the erosion and deposition are reported. Such damage situation and obtaining results from the field surveys as well as numerical simulation would suggest the issues to improve the present simulation for the tsunamis in the fields of generation, propagation and inundation. Since the tsunami simulation is still essential for evaluating risk, making hazard map and issuing warnings to mitigate damage, the several challenges including real time monitoring system and utilizing a super computer such as K are under enforcement.
April 14 th (Mon), Room A+B
Sergio R. Idelsohn (Polytechnic University of Cataluña, Spain)
Real-Time Computational Fluid Dynamics: a Challenging Demand for Safety and Environmental Problems
The main target of this lecture is to look for algorithms to simulate CFD problems as fast as possible, improving the current performance of general-purpose commercial codes. This goal does not mean yet obtaining Real Time CFD solution but this goal is on this way, with the aim of changing days of simulations for hours of simulations making feasible the present challenging demand for safety and environmental problems.
Even though implicit time integration schemes are preferred in the literature against explicit ones, the latter are in a better position attending to the present of hardware technology that is oriented to the usage of general purpose graphic processor units (GPGPU).
On the other hand it is not obvious that an Eulerian approach is better in terms of efficiency and accuracy than a Lagrangian one. At least in this presentation we will try to put into question this asseveration.
For this purpose we will present a pressure-segregation method with an explicit time integrator for the momentum equations without the CFL<1 restriction for stability. This allows large time steps independent of the spatial discretization having equal or better precision that an implicit integration. The method is based on a Lagrangian formulation that may be used with moving or fixed meshes.
Tayfun E. Tezduyar (Rice University, USA)
Bringing Them Down Safely
Special parachutes will be used for the deceleration and landing of Orion Spacecraft as it brings the astronauts back to Earth. Performance and safety are the main factors in designing these parachutes. Design knowhow, computer modeling with the actual parachute data, and drop tests with the actual parachute are the key ingredients of successful design. Because of the strong mutual dependence between the parachute aerodynamics and shape, taking into account the fluid-structure interaction (FSI) between the two is essential for accurate and realistic computer modeling. Several computational challenges are involved in this beyond those involved in a typical FSI problem. They include the parachute construction with hundreds of “ring gaps” and "sail slits,” the contact between the parachutes of a cluster, and “disreefing” from one stage to another. Furthermore, the canopy design might have some of its panels and sails removed. The flow through the “windows” and wider gaps created cannot be accurately modeled in the same way as we deal with the ring gaps and sail slits and needs to be resolved. In this joint work with Kenji Takizawa (Waseda University), we address the challenges with the core and special FSI methods we developed.
Joannes J. Westerink (University of Notre Dame, USA)
From Katrina to Sandy: High Performance Scalable Computations of Hurricane Driven Wind Waves, Storm Surge, and Flow in Integrated Ocean Basin to Shelf to Inland Floodplain Systems
Hurricane wind wave, storm surge, and current environments in the coastal ocean and adjacent coastal floodplain are characterized by their high energy and by their spatial variability. These processes impact offshore energy assets, navigation, ports and harbors, deltas, wetlands, and coastal communities. The potential for an enormous catastrophic impact in terms of loss of life and economic losses is substantial.
Computational models for wind waves and storm driven currents and surge must provide a high level of grid resolution, fully couple the energetic processes, and perform quickly for risk assessment, flood mitigation system design, and forecasting purposes. In order to accomplish this, high performance scalable codes are essential. To this end, we have developed an MPI based domain decomposed unstructured grid framework that minimizes global communications, efficiently handles localized sub-domain to sub-domain communication, applies a local inter-model paradigm with all model to model communications being kept on identical cores for sub-domains, and carefully manages output by assigning specialized cores for this purpose. Continuous Galerkin and Discontinuous Galerkin implementations are examined. Performance of explicit and implicit implementations of the wave-current coupled system on up to 32,000 cores for various platforms is evaluated.
April 14 th (Mon), Room C
René de Borst (University of Glasgow, UK)
Cohesive Fracture : Stress Triaxiality, Multiple Fields, and Numerical Approaches
The cohesive approach is probably the most versatile approach to fracture, in particular for heterogeneous materials. In this lecture, we will first present an elegant enhancement of the cohesive-surface model to include stress triaxiality, which preserves the discrete character of the cohesive approach. Another extension is the coupling of propagating cohesive cracks with non-mechanical fields such as moisture and thermal flow. This, for instance, enables the modelling of hydraulic fracturing and is relevant for earthquake analyses.
Whether a discontinuity is modelled via a continuum model, or in a discrete manner, advanced discretisation methods are needed to model the internal free boundary. The most powerful methods for a truly discrete approach are finite element methods that exploit the partition-of-unity property of the shape functions, and isogeometric analysis. Within the continuum approaches, the phase-field models are probably the most promising. We will elaborate a phase-field approach for cohesive-surface models, which, although being a continuum approach, results in a well-posed boundary value problem, and is therefore mesh-independent.
Muneo Hori (University of Tokyo, Japan)
Development of Integrated Earthquake Simulation for Earthquake Hazard and Disaster Estimation
Estimation of earthquake hazard and disaster is the first step of mitigating earthquake disasters. Integrated earthquake simulation (IES) is being developed to make this estimation by seamlessly combining various numerical simulations for earthquake wave propagation, seismic structure responses, and actions against for earthquake disasters. Analysis models are automatically constructed for underground structures and urban area consisting of buildings and structures are automatically constructed for IES with the aid of Geographical Information System. IES studies multiple earthquake scenarios and makes objective estimation of earthquake hazard and disaster as its consequence. Supercomputers such as K will be used for this computation which requires huge amount of numerical computation. It is expected that multiple earthquake scenario analysis made by IES will help us to reduce unexpected events which are induced by a large earthquake as well as to evaluate variability of possible earthquake hazards and disasters.
İlker Temizer (Bilkent University, Turkey)
Computational Contact Mechanics: Isogeometric Analysis, Mortar Methods, Mixed Formulations
In this talk, recent applications of isogeometric analysis to computational contact mechanics will be presented. There are two major aspects to the numerical treatment of contact. The first aspect concerns the discretization of the problem, both that of the bodies as well as the contact variables. The mortar method has established itself as the most robust contact discretization approach in the context of finite deformation problems with non-matching meshes. The discretization of the volume interacts naturally with the contact discretization. Here, NURBS-based isogeometric analysis will be employed. The second aspect concerns the enforcement of the contact constraints. Here, in addition to classical penalty and augmented Lagrangian methods, interior point algorithms will be applied. The overall numerical approach is derived from a mixed formulation which induces both the desired contact discretization scheme as well as the constraint enforcement method. Patch tests and surface locking studies will be presented together with local and global quality monitors of the contact interactions for thermomechanical and micromechanical problems.
April 15 th (Tue), Room A
Gui-Rong Liu (University of Cincinnati, USA)
Landslides Modeling using Smoothed Particle Hydrodynamics
Landslide is one of the frequently encountered natural disasters frequently trigged by an earthquake or heavy rain. Many landslides move very rapidly just like fluid flows and have a long run-out, which are particularity dangerous. They usually cause large damage and casualties in large surrounding areas. Numerical simulations of flow-like landslides are difficult due to the existence of free surface of different materials and large flow deformations. In this talk will focus on the discussions of using the smoothed particle hydrodynamics (SPH) method for flow-like landslides for the run-out analysis. This is because SPH is a meshfree, Lagrangian particle method, and is believed to be superior to conventional numerical methods in treating free surfaces, moving interfaces, large deformations, and hence ideal in describing the complex fluidization characteristics in flow-like landslides. Case studies on some major landslides will be examined.
Olivier Allix (L.M.T, E.N.S. de Cachan, Institut Universitaire de France)
Virtual Sizing of Composite Laminates: Progress and Challenges
In the last twenty years, there have been many advances toward a better understanding of damage mechanisms and the mechanics of laminated composites, both on the microscale and on the mesoscale coupled with the development of advanced anisotropic material models. In fact, today, some industrial programs do include more or less refined composite damage models. Industries seek confidence in models and in numerical techniques not only for static loading (calculation and optimization of joints) or quasi-static loading (low-velocity impact), but even more for dynamic loading.
A similar requirement concerns the delamination tolerance regarding initial delamination defects. One of the present industrial objectives is to replace some of these numerous these tests by numerical simulations, in other words, to perform Virtual Sizing. Some of the corresponding progress and challenges in the domain will be addressed during the presentation. They are related to the question of the confidence in the modeling of composite, the problem of the robust calculation of failure and the large scale computation of composite structures.
April 15 th (Tue), Room B
Koji Okamoto (University of Tokyo, Japan)
Nuclear Power Plant Severe Accident Analysis for Environmental Protection
Fukushima-Daiichi Nuclear Power Plant Accident caused huge huge impact on Environment and Society. Currently, several numerical simulation codes are available to predict the accident. The progress of the accident was qualitatively similar to the previous researches predicted by these codes. However, there are so many uncertainties on physical modeling, numerical modeling and especially boundary conditions. The quantitative estimation is impossible. Highly non-linear phenomena should be estimated, including phase change, material properties, radiation heat transfer, chemical reaction, eutectic, re-location, fission products release and so on. These phenomena are modeled based on several experimental data. However, the validation of these codes are very difficult.
For the recovery of Fukushima, quantitative estimation of current melted/solidified debris conditions is needed For the safety evaluation for other nuclear reactors around the world, quantitative estimation of radioactive materials release probability is needed. To improve the simulation code and reduce the uncertainties, Phenomena Identifiction and Ranking Table (PIRT) technique were applied. Several key phenomena had been extracted among many physical phenomena. World wide collaboration should be needed to improve the nuclear safety, including code verification, validation and assessment.
The prediction of the Severe Accident is very challenging issue for the safety operation of Nuclear Power Plant around the world, resulting in the environmental protection.
Andrew YT Leung (City University of Hong Kong, China / Huddersfield University, UK)
Technologies for Reducing Air Pollution for Densely Populated Cities and Their Simulation
Air pollution produced by vehicles and other factors in densely populated cities provokes respiratory and lung disorder and encourages the propagation of infectious diseases. Often, bad city planning and architecture to maximize view and value of buildings make the air pollution situation irrevocable. In order to clean street air pollutants, the paper describes a method of using the momentum induced by the piston effect of elevator lifts in high rise buildings to suck in street air into the lift shaft and to push out the air after treatment. We propose two treatment methods: the wet and the dry methods. The wet method can clean almost 100% of pollutants including virus and germs. It was used during the 2003 Severe Acute Respiratory Syndrome (SARS) period in Hong Kong. The dry method uses the principle of cyclones to reduce pollutants 30~50% at a time. The combined method is proven to be effective by experiments and simulations.
April 16 th (Wed), Room A
Jiun-Shyan Chen (University of California, San Diego, USA)
An Oscillation Limiting and Flux Conserving Meshfree Formulation for Shock Modeling
High rate impulsive loading can generate hydrodynamic processes where shocks give rise to state and field variable discontinuities. The essential shock physics include the Rankine-Hugoniot (R-H) jump condition and the second law of thermodynamics for entropy production. Oscillatory instabilities in the form of Gibbs phenomenon occur at the jump when higher-order methods are used to approximate the discontinuous solution. In this work a stable (oscillation limiting) and flux conserved formulation under the reproducing kernel particle method framework is developed for shock modeling. The Rankine-Hugoniot condition is naturally satisfied and correct shock speed is obtained under the appropriate weak form. A Riemann-embedded flux divergence operator, formulated under the framework of stabilized conforming nodal integration, guarantees correct entropy production and flux conservation and thus produces the correct shock propagation, while a Godunov-type flux-corrected velocity limits shock oscillation. The Godunov velocity correction for oscillation limiting is constrained to the shock region using a detection algorithm based on the reproducing kernel spectral decomposition property, so that higher-order accuracy is maintained elsewhere in the domain.
Antonio Huerta (Polytechnic University of Cataluña, Spain)
Real-time wave monitoring in harbors
A reduced order model approach is used to evaluate surface elevation (agitation) in harbors. The full elliptic harbor model with real geometry and bathymetry is accounted for. The strategy combines an off-line stage in which a generalized solution is computed, and an on-line phase in which, even on deployed, handheld, platforms such as smartphones or tablets, real-time response is obtained at any position and for any value of the design parameters. The efficiency of the on-line stage stems from the fact that no extra resolutions are required.
Obtaining the generalized solution is the major computational burden. This off-line stage re- quires a robust and efficient solver for the Helmholtz equation in a non-homogenous unbounded domain. Several ingredients are required. One is the implementation of efficient artificial boundary conditions, which must absorb outgoing waves, reproduce the real geometry and may be implemented as close as possible to the area of interest. Another is the use of effective nu- merical strategies high-order approaches, hybridizable Discontinuous Galerkin approaches and p-adaptivity are discussed and tested. Challenging engineering applications are shown.
April 16 th (Wed), Room B
Sung-Kie Youn (KAIST, Korea)
Isogeometric Analysis for Structural Optimization and Meshfree Method
Isogeometric analysis is a NURBS (Non-Uniform Rational B-Splines) based spline finite element method. NURBS is currently the most commonly used mathematical language in CAD and computer graphics for creating geometrical models. In the Isogeometric Analysis, NURBS is used both for the discretization of the analysis models and for the approximation of field variables.
In this presentation, a few examples of Isogeometric Analysis will be given. Then it will be shown how the features of Isogeometric Analysis combined with TSA (Trimming Surface Analysis) can be exploited for the shape/topology optimization of structures. Also a meshfree method contrived by employing TSA in the spline finite element method will be explained. Some applications of the meshfree method will be presented through numerical examples.
Zhuo Zhuang (Tsinghua University, China)
Dislocation Starvation by Computation/Experiment in Crystal Submicron Pillars
The size effect at submicron scales is a crucially important issue on plasticity of materials. There are two main mechanisms that govern the plastic behavior at the submicron scales as the dimensions variation of the compressed pillars. The first is dislocation starvation induced by the annihilation of dislocations from the free surfaces. The second is forest dislocation hardening induced by the multiplication of a large amount of dislocations. The compression tests of Cu and Al single crystal micro pillars are simulated and experimented in this paper with a range of 300~1000 nm side lengths of the cross-section containing initial dislocation networks. The mechanical behaviors are investigated during compression process including dislocation pattern transitions, size-dependent flow stress evolution and strain hardening mechanisms. Two main factors affecting the dislocation motion are taken into account, such as free surface interaction and dislocation junction.