Real-ESSI Simulator
Introduction
Availability
Version
Documentation
Modeling Approach
Modeling and Simulation Features
Introduction
The
Real-ESSI Simulator
(
Realistic
Modeling and
Simulation of
Earthquakes, and/or
Soils, and/or
Structures and their
Interaction)
is a software, hardware and documentation system for high performance,
sequential or parallel, time domain, linear or nonlinear, elastic and
inelastic, deterministic or probabilistic, 3D/2D/1D finite element modeling and
simulation, i.e analysis of
- statics and dynamics of soil,
- statics and dynamics of rock,
- statics and dynamics of structures,
- statics of and dynamics of soil-structure systems,
- dynamics of earthquakes, and
- dynamic earthquake-soil-structure interaction.
The Real-ESSI Simulator systems is used for
(a) design
and for
(b) assessment
of static and dynamic behavior of civil engineering objects, including buildings,
bridges, roads, dams, power plants, tunnels, etc.
Design: Multiple linear elastic load cases can be
combined and design quantities, sectional forces exported for design.
Assessment:
Practical, realistic, inelastic, nonlinear load staged analysis, with
accurate modeling of elastic and inelastic, nonlinear components, and with all
the necessary simulation/algorithmic features available, as listed below, is performed to
assess design, safety margins and economy of infrastructure objects.
The Real-ESSI Simulator is developed at the University of California,
Davis, in collaboration and with partial financial support from the
US-DOE,
US-NRC,
US-NSF,
US-BR,
US-FEMA,
CalTrans,
CNSC-CCSN,
CH-ENSI/B&H AG,
UN-IAEA,
CERN,
Shimizu Corp.,
etc.
The Real-ESSI Simulator develops analysis to inform and predict.
Brief history of the Real-ESSI Development.
Real-ESSI Simulator Education and Training
Real-ESSI Simulator System Availability
-
Linux, Debian Package:
Real-ESSI Simulator Simulator system Debian Packages,
for sequential and parallel Real-ESSI Simulator, version 23.07, is
available for Linux-Ubuntu version 22.04.
HERE
.
Please see
Real-ESSI Simulator System Procurement Manual
for more information on how to install Real-ESSI debian package.
Provided Real-ESSI system debian package
is not optimized for speed, programs are compiled and linked using
generic compile and link options/flags.
This is done so that provided Real-ESSI program executables and
other programs from the package are portable
and should/will run on any Linux-Ubuntu-22.04 computer.
-
Windows, WSL:
Real-ESSI Simulator Program can be installed on MS-Windows computers
through Windows Subsystem for Linux (WSL) system.
Please see
Real-ESSI Simulator System Procurement Manual for instructions on how to obtain and
install Real ESSI Debian Package using WSL.
-
Linux Direct Build:
Real-ESSI Program can be directly built (compile, link and install)
on user/analyst remote site/computer. This
will allow us to compile and link an optimized, fast, sequential and
parallel versions of the Real-ESSI Simulator. However this involves
additional effort to maintain and update the program. This is sometimes
the only option for users/analysts with Internet use limitations, for
example security limitations for some government agencies, national labs,
and companies.
-
Cloud:
Real-ESSI Program, optimized, fast, sequential and parallel versions,
are available to users/analysts through
See
Real-ESSI Cloud Computing Manual for more information.
Use by academics is free (AWS will charge hardware fee, contact Prof.
Jeremić for details, AWS), while for commercial entities, there are
Real-ESSI AWS installations that are maintained by different companies, please consult AWS
market place and search for Real-ESSI.
-
Sources:
Real-ESSI Simulator System sources are distributed to collaborators
through a version of open source license.
Real-ESSI Simulator System sources are distributed to some users, users that
want to install Real-ESSI on large super-computers, through a special,
restrictive license.
Contact Prof. Jeremić for details,
Real-ESSI Simulator Version
Current
Real-ESSI Simulator release version is:
Global Release, 24.04, Apr2024.
Next
Real-ESSI Simulator release version will be:
Global Release, 24.10, Oct2024.
Past
Real-ESSI Simulator release version was: Global Release, 23.07, Jul2023.
Change LOGS, for the GLOBAL_RELEASE, that is, what have we added, improved and changed in last year or so...
Real-ESSI Simulator system documentation
-
Real-ESSI Simulator DSL Manual,
describes the Real-ESSI Simulator system Domain Specific Language
(Real-ESSI DSL).
-
Real-ESSI Simulator Examples Collection,
Illustrates Real-ESSI Simulator system capabilities and the Real-ESSI DSL, input language.
-
Real-ESSI Simulator Online Education and Training,
provides slides and recorded lectures on theoretical topics, applied topics and
recordings of worked out examples on computers.
-
Real-ESSI Simulator Pre-Processing Manual,
describes model development and pre processing tools.
-
Real-ESSI Simulator Output Format Manual,
describes format for all the results, that are saved in
Hierarchical Data Format 5 (HDF5).
-
Real-ESSI Simulator Post-Processing Manual,
describes post-processing and visualization of Real-ESSI simulation results.
-
Real-ESSI Simulator Procurement Manual,
describes how to obtain and install Real-ESSI Simulator System.
-
Real-ESSI Simulator Cloud Computing Manual,
describes how to use Real-ESSI Simulator system on Amazon Web Services (AWS) computers.
-
Real-ESSI Simulator System Build and Install Manual,
describes how to build/compile/link Real-ESSI Simulator system.
-
Real-ESSI Modeling and Simulation
of
Earthquakes, Soils, Structures and their Interaction (ESSI)
Guide BOOK, for Professional Practice,
provides practical modeling and simulation steps to perform
ESSI analysis for Nuclear Installations, Dams,
Buildings, Bridges, Tunnels and other soil-structure systems.
-
Real-ESSI Modeling and Simulation of ESSI Problems Core Functionality, for Professional Practice,
provices modelinng and simulation parameters for
ESSI analysis of
Buildings, Dams, Bridges, Nuclear Installations, Tunnels and other SSI systems.
To view Real ESSI
Simulator system documentation PDFs it is best to use Google Chrome.
All the libe links,
animations, etc. work well within Google Chrome. Simpler, portable versions of PDFs is
also provided, however links and animations do not work.
Complete documentation, including all of the previous documents and manuals is
available within Lecture Notes:
Nonlinear Finite Elements:
Modeling and Simulation of
Earthquakes, Soils, Structures and
their Interaction
Real-ESSI Simulator Modeling Approach
Practical, high, medium and low fidelity analysis, modeling and simulation is
performed in order to
control and reduce modeling, epistemic uncertainty.
Modeling uncertainty is introduced in models, and
subsequent simulation results, when
modeling simplifications, some unrealistic and some unnecessary,
are made.
While the Real-ESSI Simulator is capable of
analyzing high fidelity, high sophistication models, it is also possible to analyze
very simple models. In fact, the Real-ESSI Simulator allows analysis of a full
hierarchy of models, from very simple models to very sophisticated models.
Modeling and simulation from simple to more sophisticated models is encouraged,
in order to gain better understanding of model behavior and to gain confidence
in analysis results.
For example, a building model can start as a simple beam or a simple frame on
a fixed base. Next, analyst can add wall, plate and shell elements, and later
springs for foundation modeling, all linear elastic. Next, analyst might add solids
and more realistic foundation model, all still linear elastic.
Loading can start with simple static pushovers, simple dynamic pings/pushes at
certain locations.
Next, analyst
might change material models for parts of the soil-structure
system from linear elastic to nonlinear, inelastic.
Similar simple loads are applied, to test the model.
Along the way of these
changes, while climbing up the hierarchy of model sophistication, analyst will be able
to gain good understanding of model behavior, and sensitivity of model response to
various simplified or sophisticated modeling options,
various material behavior, various loads, etc. Finally, analyst
might use a sophisticated, realistic model, with realistic loading, using
many available Real-ESSI Simulator
modeling and simulation features.
It is highly recommended to follow this hierarchy of modeling from simpler toward
more sophisticated models, particularly for nonlinear analysis. It is also
highly recommended to use sound engineering judgement in inspecting and
understanding simulation results. It can happen, and it does happen, that
instant use of very sophisticated models, without model verification and
gradual understanding of model behavior through hierarchy of models, lead to
results that are difficult
to understand and that can sometimes be wrong due to use of wrong modeling details.
Presence of potential errors in results can also be due to the modeling
approximations and simplifications.
Influence of those approximations and simplifications needs to be
understood, using hierarchy of models, before simulation results can be used for
design and assessment.
Real-ESSI Simulator Modeling and Simulation Features
- Modeling: Finite Elements
- - Solids
- - Dry/single phase solid bricks (8, 20, 27, variable 8-27 node bricks),
linear elastic and
nonlinear/inelastic/elastic-plastic, 3D
- - Saturated, two phase solid bricks (8 and 27 node u-p-U
bricks, liquefaction modeling, effective stress
analysis, fully saturated
soils), linear elastic and
nonlinear/inelastic/elastic-plastic,3D
- - Partially saturated, unsaturated, three phase
solids (8 and 27
node u-p-U bricks, liquefaction of partially
saturated/unsaturated soils, effective stress
analysis, partially saturated/unsaturated soils), linear
elastic and nonlinear/inelastic/elastic-plastic, 3D
- - Cosserat solid 8 node brick, with translational
and rotational DoFs, and with internal length, for
localization modeling, linear elastic and
nonlinear/inelastic/elastic-plastic, 3D
- - Truss,
- - Linear elastic, 3D
- - Nonlinear/inelastic/elastic-plastic, fiber cross section, 3D
- - Kelvin-Voigt, linear elastic and viscous, 3D
- - Beam,
- - Bernoulli beam, linear elastic, with variable DoFs per node, end
nodal offsets, 3D
- - Bernoulli beam, large displacement, corotational transformation,
small strain, linear elastic material, variable DoFs per node, end
nodal offsets, 3D
- - Bernoulli beam, linear elastic and nonlinear/inelastic/elastic-plastic,
fiber cross section, concrete and steel,
displacement/stiffness based, 3D
- - Timoshenko beam, linear elastic, 3D
- - Timoshenko beam, with directional shear correction factors,
for non-square or non-circular cross sections, linear elastic, 3D
- - Metamaterial/Metadevice, for mechanical/seismic wave deflections
- - Resonant unit cells metamaterial/metadevice, 3D
- - Negative stiffness metamaterial/metadevice, 3D
- - Inerter metamaterial/metadevice, 3D
- - Shell, Plate, Wall
- - Quad shell/plate/wall, with 6DoF per node, ANDES/Carlos_Felippa, linear elastic, 3D
- - Quad shell/plate/wall, with 6DoF per node, NLDKGQ/Xinzheng_Lu,
nonlinear/inelastic, reinforced concrete using
nonlinear concrete and steel models, 3D
- - Interface/Joint/Contact Axial, Dry, 3D, gap opening-closing for:
- - Soft, concrete/steel to concrete/soil/rock contact/interface/joint, and
- - Hard, steel to steel, had surface to hard surface
contact/interface/joint,
- - Interface/Joint/Contact, Axial, Saturated (Underwater), 3D,
gap opening-closing, with water pumping, automatic buoyant force
generation/calculation and application, for:
- - Soft, concrete/steel to concrete/soil/rock contact/interface/joint, and
- - Hard, steel to steel, had surface to hard surface
contact/interface/joint,
- - Interface/Joint/Contact Frictional Slip for Dry and
Saturated/Effective Stress conditions:
- - Elastic perfectly plastic,
- - Elastic hardening plastic,
- - Elastic hardening-softening plastic,
- - Base isolators, elastomeric, 3D
- - Base dissipators, frictional pendulum, 3D
- - Stochastic Shear Element
Modeling: Deterministic Material Models
- - Deterministic Elastic Material Models
- - Linear Elastic, 3D
- - Linear Elastic, 1D
- - Nonlinear Elastic, 3D
- - Micro-polar, Cosserat elastic material, 3D
- - Deterministic Elastic-Plastic Material Models:
(all elastic-plastic models are effective stress based,
can be used for dry, unsaturated/partially saturated and fully
saturated modeling, and can be used as perfectly plastic,
isotropic hardening/softening and/or kinematic hardening
models)
- - von Mises model for steel, concrete, soil and rock, 3D
- - Cosserat von Mises model, with internal length, 3D
- - Drucker-Prager model for concrete, soil and rock, 3D
- - Rounded, hyperbolic Drucker-Prager model for concrete, soil and rock, 3D
- - Rounded Mohr-Coulomb model for concrete, soil and rock, 3D
- - Leon Parabolic model for concrete, soil and rock, 3D
- - Modified Cam-Clay model for soil, 3D
- - SaniSand model for sand, 3D
- - SaniClay model for clay, 3D
- - Pisano vanishing elastic region model for stiffness reduction, G/Gmax modeling, for soil, 3D
- - Nested Surface von Mises model for stiffness reduction, G/Gmax modeling, for soil, 3D
- - Nested Surface Drucker-Prager for stiffness reduction, G/Gmax modeling, for soil, 3D
- - Nested Surface rounded Mohr-Coulomb model for stiffness
reduction G/Gmax modeling, for soil, 3D
- - Tsinghua soil liquefaction model, for soil, 3D
- - Concrete01 and Concrete02 fiber models, 1D, for nonlinear fiber concrete
beam and truss element in 3D
- - Faria-Oliver-Cervera concrete model for modeling of
concrete walls, plates and shell elements, 3D
- - Steel01 and Steel02 fiber models, 1D, for nonlinear
fiber steel beam and truss element in 3D, and as
discrete steel reinforcement for wall, plate and shell
elements in 3D
- - Viscous (Damping) Material Models
- - Rayleigh damping
- - Caughey damping (2nd order, 3rd order...)
Modeling: Probabilistic Modeling
- - Probabilistic Material Models
- - Linear Elastic, 1D
- - Nonlinear, Elastic-Plastic:
Armstrong-Frederick PDF nonlinear hardening model, 1D
- Probabilistic Forcing
- - Probabilistic Forcing: Characterization of earthquake motions
- Probabilistic Sensitivities
- - Sensitivity of probabilistic results to uncertain input parameters
(uncertain material parameters and uncertain input motions)
Modeling: Solid/Structure-Fluid Interaction, Coupling with free
surface computational fluid dynamics
for reservoirs, tanks and pools, using OpenFOAM
Modeling: Energy Calculations for static and/or dynamic, monotonic and/or cyclic analysis:
- - Energy input into finite element model, computations, static and/or
dynamic, monotonic and/or cyclic, incremental (per
loading/time step) and cumulative (per loading stage and
total) external/seismic energy input
- - Elastic strain energy computations, storage, static and/or
dynamic, monotonic and/or cyclic, incremental (per
loading/time step) and cumulative (per loading stage and
total)
- - Plastic free energy (storage) computations, static and/or
dynamic, monotonic and/or cyclic, incremental (per
loading/time step) and cumulative (per loading stage and
total)
- - Energy dissipation (losses/conversion, for each
finite element) computations, static and/or dynamic,
monotonic and/or cyclic, incremental (per loading/time step)
and cumulative (per loading stage and total)
Modeling: Loading
- - Nodal loading,
- - Linear, constant rate of load application
- - Path, use of arbitrary, load path defined by user in external file
- - Self weight, for all elements
- - Pressure/surface load for solid bricks (8, 20, 27 node bricks)
Modeling: Seismic Input
- - Analytic input of seismic motion wave field, including body
waves (P, SH, SV) and surface waves (Rayleigh, Love, etc.),
with
analytic radiation damping, using Domain Reduction Method (Bielak et
al.), for:
- - One component seismic motions (1C),
- - two component seismic motions (2C),
- - three component seismic motions (3C),
- - two independent component seismic motions (2 X 1C),
- - and three independent component seismic motions (3 X 1C),
- Automatic generation of input motions from user defined seismic wave field
- - 1D/1C Deconvolution: analytic development and input of seismic
motions, automatic calculation, from deconvolution of given
surface motions, in 1C and 3x1C, using 1C wave propagation theory
- - 3D/3C Deconvolution/Inverse Modeling: analytic
development and input of seismic motions, automatic
calculation, from inverse modeling of a 3C/3D wave field from
sparse surface and/or downhole motions, in full 3D/3C
- - Analytic input of seismic motions, automatic calculation, from
deconvolution of given surface motions, in 1C and 3x1C, using 1C
wave propagation theory
- - Analytic input of seismic motions, automatic
calculation, development using analytic, 3C plane wave
propagation, inclined wave field (P and SV, Rayleigh waves,
Love waves, Stoneley waves, etc.) (Thomson-Haskel solution)
- - Analytic input of seismic motions, automatic calculation,
from convolution of given in depth motions, in 1C and 3x1C, using 1C
wave propagation theory
- - Analytic input of seismic motions from regional scale
geophysical modeling, for example automatic generation of motions from
SW4 program or from Real-ESSI, in 3C, 3x1C and 1C
- - Applied motions at the base of the model, or any other set of
nodes
Simulation: Static
- - Load control
- - Displacement control
- - Arc Length control
- - Hyper-Spherical constraint control
Simulation: Dynamic
- - Newmark algorithm, with constant or variable time stepping
- - Hilber Hughes Taylor, Alpha method, algorithm, with constant or
variable time stepping
Simulation: Restart, simulation continuation, options
- - Restart after any loading step, loading stage, so that
loading steps and loading stages can be staggered and branched
out
- - Restart after convergence failure: after convergence failure, if
it happens, with all the numerical modeling
precautions that are available, from the previously converged
step, with possible changes in simulation algorithm and
simulation parameters.
Design: Combination of linear elastic load cases
- - Perform multiple linear elastic analysis for different load cases.
- - Combine, superimpose results from different load cases, multiplied
with a safety factor to get design quantities, sectional forces
- - Export design quantities, sectional forces in formats used by cross
section dimensioning programs.
Model building, pre-processing, mesh generation
- Models are described/defined/input using Real-ESSI Domain
Specific Language (DSL), described in
Real-ESSI DSL manual
- Models developed using other mesh generators can be translated into
Real-ESSI DSL input format
Results visualization and post processing
- - Output is in HDF5 format, compatible with major post-processors
and visualizers, ParaView for example
- - All results, everything is output (stress, strain, displacements,
internal variables, energy, etc.), unless users
requests no output for given elements, nodes, etc...
- - Post processing and results visualization is done using
ParaView visualizer, that is extended with
Real-ESSI Post-Processing tools.
Verification and Validation
- - Verification: each element, method, algorithm and procedure
is verified in detail and approximation accuracy documented
- - Validation: to some extent, looking for high quality validation,
experimental data
High Performance Computations
- - Template Metaprograms available within Real-ESSI for
both Sequential Real-ESSI and Parallel Real-ESSI
- - Parallel Real-ESSI is based on Distributed Memory Parallel paradigm
and runs on multi core / multi CPU PCs, clusters of PCs and large
parallel super-computers, such as EDISON@NERSC-LBNL,
CORI@NERSC-LBNL, STAMPEDE2@TACC, Euler@ETH, and on Cloud
computers, such as
Amazon Web Services/AWS
- - Plastic Domain Decomposition (PDD)
featuring dynamic computational load balancing, that is efficient
for elastic-plastic finite element problems where elastic-plastic
(slow) and elastic (fast) domains change dynamically during run
time, and for variable performance CPUs/compute nodes
- - Hardware Aware Plastic Domain Decomposition (HAPDD), that
extends PDD, for modern hardware platforms, with variable
network performance (core-core, CPGPU-CPU, CPU-CPU and
compute node-compute node) and variable CPU performance
(CPUs, cores and GPGPUs)
For additional information, please contact
Boris Jeremić ( jeremic @ ucdavis . edu ), use Subject: Real-ESSI.
Web page version: 30Oct2024
|