Flac Modelling

[Garland Flugum]

FLACutilises an explicit finite difference formulation that can model complex behaviours, such as problems that consist of several stages, large displacements and strains, non-linear material behaviour, or unstable systems (even cases of yield/failure over large areas, or total collapse).

Perpetual, monthly lease, and annual lease licences are available as either a local USB-key (which is portable), a multiple-seat network USB-key or digital web licence. Academic institutions qualify for special pricing discounts.

FLAC has been developed primarily for geotechnical engineering applications in the fields of civil, mining, oil and gas, and power generation. FLAC is also a valuable tool used for research in rock- and soil-mechanics, particularly of localisation and evolution of shear bands in frictional materials. FLAC has also been used in the manufacturing field where the analysis of highly deformable materials is needed.

The explicit, time-marching solution of the full equations of motion (including inertial terms) permits the analysis of progressive failure and collapse, which are particularly important phenomena for mine design and geotechnical construction.

Dynamic Analysis: Can be performed with FLAC using the optional dynamic calculation module. User-specified acceleration, velocity, or stress waves can be input directly to the model either as an exterior boundary condition or an interior excitation to the model.

Two-phase Flow: The two-phase flow option in FLAC allows numerical modelling of both fluid-flow and fully coupled simulations (with optional capillary pressure) of two immiscible fluids through porous media.

FLACS-CFD 24.1 is the latest version of our software featuring significant upgrades in geometry handling designed to streamline your modelling processes.

There are also significant enhancements in the accuracy of fire simulations and safety assessments.

This webinar explores how the advanced capabilities of FLACS CFD consequence analysis software in simulating and analysing hazards can be exploited for informed decision-making in managing hydrogen safety in nuclear facilities.

Engie Lab Crigen presents the HyCARE Project, focusing on the development of secure hydrogen storage systems utilising metal hydride (MH). The discussion includes how FLACS simulation plays a critical role in verifying the safety of the design.

FLACS is utilised to investigate the occurrence of unexpected peak overpressures immediately after ignition at substantial distances from the ignition point, before the peak pressure within the flammable cloud is attained. This phenomenon was identified during simulations of stoichiometric flammable cloud ignitions under default boundary conditions advised by Gexcon.

This study aims to deepen our understanding of the consequence of ammonia releases and to advance risk management practices by evaluating different modelling techniques to identify the most effective tools for analysis and strategic planning.

Efectis showcases their application of FLACS in analysing the consequence of accidental high-pressure hydrogen releases, a crucial aspect in ensuring the safety of hydrogen facilities, specifically those utilising modular 20 or 40-foot containers.

In this presentation, Gexcon explains the validation methodology within FLACS and introduces the new validation GUI. The presentation also covers perspectives into the impact of validations on probabilistic risk analysis, specifically addressing how validation findings can be utilised to incorporate uncertainty into probabilistic risk assessments.

Gexcon outlines the development progress of FLACS-Fire, presenting an overview of its sub-models and thermal exchange principles, the existing validation framework, current challenges, and a review of the approach alongside future development plans.

Gexcon presents the engineering and scientific aspects of the large-scale experiments on hydrogen-nitrogen-air mixture explosions, aimed at refining the FLACS software to ensure the delivery of dependable outcomes for applications in nuclear and hydrogen safety.

This presentation serves as an introductory welcome from EDF to the 2023 FLUG Meeting, where EDF presented the application of FLACS-Hydrogen in their hydrogen risk analysis activities for nuclear power plants.

Gexcon working in partnership with EDF R&D is thrilled to invite you to the long-awaited return of the in-person FLUG Meeting which will be held on 10-11 October 2023 at EDF Lab Paris-Saclay in France.

In this study, the Flame Acceleration Simulator (FLACS-CFD) software was utilized to evaluate the behaviour of hydrogen release, dispersion, and flame propagation in the presence of multiple HFCVs in an outdoor parking space through the activation of thermally activated pressure relief devices (TPRDs).

This presentation shows the influence of the size of the gas cloud on the severity of the explosion and the effect of different ventilation arrangements. It also presents possible mitigation strategies to enable the safe deployment of hydrogen in applications that require close proximity to the public.

Due to its unique flammability characteristics, the use of LH2 in aviation presents several safety challenges, from the materials used through aircraft refuelling procedures to the storage of large amounts of LH2 at or near the airport.

At the 2022 International Loss Prevention Symposium, Gexcon will present the result of CFD-based explosion analysis of hydrogen-methane-air mixtures which was performed using FLACS-CFD consequence modelling software.

Using our flagship CFD modelling tool FLACS-CFD, Gexcon conducted several case studies on real facilities looking at the design of bulk storage areas in terms of safety during an ignited event, using relevant standards and documentation for guidance.

Most lossless audio compression codecs work very similarly: they split the audio stream in multiple blocks or frames and then compress each one of them. Compression is achieved identifying and modelling the structure in the audio signal: any repeating pattern is redundant by definition and can instead be represented more efficiently using a mathematical model and its parameters. These models need not to be complicated: constant values can already model silence blocks while linear predictive coding (a linear autoregressive model) is used for the other blocks. In order to recreate the signal without loss the approximation residuals of the models are also saved, but using a coding scheme optimised for their distribution. Linear predictive coding leads to residual Laplace-distributed and hence usually the Rice coding scheme is used.

The previous model could have been expressed as an instance of a general linear regression model whose covariates are the \(p\) lagged values. The solution would then minimise for the expected mean squared error directly without making assumptions about the underlying WSS property. This leads to a very similar but more expensive solution, as we lose the benefit of the WSS assumption.

It is easy to notice the similarities with equation \refeqn:yh2 : this solution is computing the sample estimates of \(\mathbfR\) and \( \mathbfr\), but using a different subset of samples for each estimate. Because of this the estimates have to be recomputed for each entry, and the toeplitz structure is lost, rsulting in higher estimation costs to build the matrices and higher matrix inversion costs to solve for \(\boldsymbol\alpha\). The computational complexity of this solution is \(O(n^2p + p^3)\), where the first term comes from the inner product of the matrix of covariates and the second term from its inversion. Usually \(n \gg p\) and thus this solution has complexity \(O(n^2p)\) which is much worst than the previous result of \(O(np)\) when using \refeqn:yh2!

Suppose one was to observe a realisation of the above mentioned stochastic process, for example an audio signal, and used LPC or some other model to approximate it. In order to reconstruct the original signal without any approximation error it is necessary to know what the exact realisation of the these error were and add them back into the reconstruction. The problem of efficiently storing these residuals is an instance of entropy coding.

when designing or choosing a code the objective is to minimise \(l(C)\) for the distribution of input words \(p(x)\), so that the encoded input will require on average the least number of symbols. This is the problem referred to as entropy coding

A good general code that works for any input distribution is the Huffman code [1], and is for example used in the popular compression tool gzip. When the distribution of the input is known a priori though it is possible to use a coding scheme tailored for that distribution.

For example for the case of reconstruction errors the occurrence of small values in the input stream is significantly more likely than large values, which makes having smaller code words that cannot be confused to be the prefix of longer code words very convenient to encode them. A code in which no code-word is allowed to be the prefix of another code-word is called a prefix code and for the case of alphabets following a geometric distribution it can be shown that the optimal prefix code is the Golomb code [2].

Golomb coding [3] was invented by Solomon W. Golomb in the 1960s. It takes the input \(s\) and divides it by the tunable parameter \(m\). First the quotient is encoded in unary coding and then the remainder is econded in truncated binary coding. Rice coding is a special case of Golomb coding where \(m = 2^k\), which is faster to encode and decode thanks to the usage of powers of 2. It is very intuitive to understand how the coding scheme works just by looking at the follwowing table example

FLAC is mainly based on shorten [4], with additional features for more convenient use in real case consumer scenarios. A high level overview of the codec can be found here, while the definition of the stream is defined more precisly here

3a8082e126