EP2847714A2

EP2847714A2 - Computerised method and system for modelling aeraulic flows, in particular for the quantitative assessment of the risk of airborne contamination

Info

Publication number: EP2847714A2
Application number: EP13724358.0A
Authority: EP
Inventors: Philippe Le Goff; Catherine Hugel-Le Goff
Original assignee: Hygie-Tech SA; HYGIE Tech SA
Current assignee: Hygie-Tech SA; HYGIE Tech SA
Priority date: 2012-05-11
Filing date: 2013-04-10
Publication date: 2015-03-18
Also published as: US10073930B2; WO2013167942A3; US20150073758A1; WO2013167942A2; CH706502A2

Abstract

The invention relates to a computerised method and system for modelling aeraulic flows in an environment, in particular for the assessment of the risks of airborne contamination. The principle of the method according to the invention involves considering that, in a given unit volume or grid cell, which is defined so as to be sufficiently small, the final values of the primary variables (V, P, T) are the result of the primary values thereof calculated in an irrotational field consisting of a pressure-velocity coupling with a turbulence model adapted to incompressible Newtonian fluids belonging to the range of validity of the method.

Description

Method and computerized system for modeling airflow, in particular for the quantitative evaluation of the risk of airborne contamination

The invention relates to a method and a computerized system for modeling airflow in an environment, especially for the assessment of the risks of airborne contamination.

The fields of application of such a process concern, in particular, defense, the protection of sensitive installations, hospitals, clean rooms, premises with controlled environment, the pharmaceutical, cosmetic, electronic, agro-food and mechanical chemical industries.

The knowledge and the control of the airflow flows concern those skilled in the art since it seeks to control the environmental parameters for the purposes of comfort and safety. To achieve this goal, two approaches are possible: practical experimentation or mathematical calculation using recognized physical laws. Traditionally both approaches are considered competing and most often the fieldman prefers the experiment when the scientist favors the calculation. However, neither of these approaches can assume the quality of absolute accuracy when it comes to representing a flow of air.

The experimental simulation can only give a vision of the reality of the flows remaining dependent on the conditions of the experiment (nature of the tracer, modes of release of the tracer in the environment, aeraulic and operational conditions in the environment considered at the time of the 'experience). Also, if the experimental simulation can bring interesting lessons in the qualitative representation of airflow flows, it remains linked to the conditions of the experiment and should not be extrapolated to other conditions. Moreover, the margin of error of the experimental simulation in case Isolated quantitative study remains very important and can often be underestimated by experimenters.

Also, in a field of application including the notion of risk management which, by definition, requires the extrapolation of results and their projections in 5 different systems of operational configurations, only computational modeling is theoretically able to respond objectively. to the requirements of the problem posed. This computational approach is known to those skilled in the art as CFD (CFD: Computational Fluid Dynamics) and relies on the search for numerical solutions to approximations of state equations of fluid dynamics.

In practice, the CFD proceeds according to an always identical principle regardless of the method implemented. Indeed, the known methods are all based on the following steps:

1. Selection of fluid dynamics state equations adapted to the problem to be solved

The description of the flow of a moving fluid is governed by the equations of state, widely known under the name Navier-Stokes Equations, to which are added restrictions such as:

- Newtonian and incompressible fluid

0 - Constant physical properties

- Buoyancy effect governed by the Boussinesq approximation

- Neglected viscous dissipation

- Neglected radiation

- Monophasic fluid

5 - Monocomponent fluid.

These assumptions and restrictions, however, cover a very wide range of applications.

Under these conditions, the state equations are written:

- Law of conservation of the mass V · V = 0

- Law of conservation of the momentum

₊ V (V - V) = -VP + W ² V + ^ (T - T ₀ ) + F _H int

dt

- Energy Conservation Act:

The solutions to these equations provide the desired information on the fluid flow variables (also called primitive variables) that is: the velocity V = {u, v, w}, the pressure P and the temperature T .

2. Linearization of state equations with model input

approximation

The mathematical expression of the law of conservation of the momentum shows a non-linear term that makes the equation impossible to solve exactly. This impossibility is known as the "Closure Problem". As a result, it is necessary to approximate the solution. Many models have been proposed with varying degrees of success. Of these, the k-ε model is certainly the most commonly used.

The k-ε model is based on two equations taking into account turbulent kinetic energy (k) and turbulence dissipation (ε). In this model, the turbulent viscosity is determined empirically by the equation: ε

equation in which C _M is a constant (usually 0.09).

In fact, the implementation of the model k-ε requires the introduction of several other empirical functions such as F _m , Fi, F2 and E whose values vary according to the authors. Therefore, it is clear that conventional methods of modeling airflow can not provide exact solutions to the problem and the relevance of the approximated values is not always verifiable or even adjustable by the user.

3. Discretization of the computational domain to generate systems

algebraic systems capable of providing, by an iterative method, approximate solutions to the equations of these systems

The foundation of computational modeling is then based on a principle of discretization of space and indexation of modified state equations by the introduction of empirical values. The computational domain is meshed into a finite number of volumes called mesh M whose sum gives the entirety of this domain. In each of these meshes, the scalar variables P and T are evaluated.

In addition, three complementary gratings offset in the directions of the unit trihedron {u, v, w} are added for the determination of the velocity vector V = {u, v, w} according to each of the axes. The integration of the partial differential equations in their respective mesh is then obtained by the determination of the term "convection - diffusion".

The algebraic equation can then be solved at the point A of the mesh M to know the value of the calculated variable.

It is useful to note that, to be solved, this equation requires the knowledge of the values of the variables calculated in the neighboring meshes. This interdependence of the set of meshes to build the final result in the computation domain requires the implementation of an iterative process to refine the values in each volume according to the values calculated in the neighboring cells until obtaining stabilized overall solutions. 4. Transcription of the iterative method into algorithms that can be processed by a computerized system

The search for convergence of variables in the discretized domain requires the use of computerized systems with considerable computing power. Indeed, the accuracy of the result obtained is conditioned by the size of the mesh: the smaller it is, the more accurate the result. However, the overlap of the domain by meshes of small sizes has the consequence of increasing the number and thus increase the calculation time necessary to obtain the result (note that the increase in the number of meshes is also o o conditioned by the equipment on which the calculation is conducted and

in particular by the size of the memory available for the calculation).

Most of the algorithms for this iterative calculation were developed in the 1960s to 1980s and have seen little improvement until today. The most used algorithms in commercial codes are known as SIMPLE, SIMPLER, MAC, SOLA (non-exhaustive list).

5. Introduction of initial conditions and boundary conditions satisfying both the requirements of the algorithms and the problem posed Whatever the algorithm used, the implementation requires the introduction 0 of data making it possible to "turn" the calculation . These data are

consist of initial conditions and boundary conditions of the computational domain. In the conventional approach of the CFD codes, the algorithmic is deployed from the meshes in which the values of the primitive variables (V, P, T) are known and diffuses towards the set of neighboring cells while having to satisfy the conditions boundaries defined on the boundaries of this area.

In conventional CFD, boundary conditions set the value of the velocity modulus at the boundary of the domain (usually zero velocity) and develop towards the interior of the domain according to the Dirichlet or Neumann models. 6. Systems Resolution

The algorithm of a conventional CFD code is then activated until the conditions of convergence (or stability of the solution) are satisfied. However, conventional CFD codes do not have the means to judge the relevance of the calculated solution. This means that the relevance of the result of the calculation is mainly conditioned by the CFD expert when selecting the most suitable code for the type of modeling to be performed and, for a given code, by the choice of the models and the values of the codes. variables and the introduction of initial conditions.

Although this human-factor conditionality of the calculated solution does not pose a problem in structures experimented with computational modeling (in the aeronautical field, for example), it nevertheless constitutes an obstacle for routine use by the man in the field. (as for example, for the modeling of aeraulic flows in a laboratory) with as a corollary the problems of misuse or management of the interfaces mentioned above.

The conventional principle of modeling by the calculation of aeraulic flows therefore qualifies the CFD codes for applications with high added value and restricted parametric variability (as in the aeronautical field) than for real configurations (laboratories, public place.) Prevailing in a risk assessment framework for airborne contamination.

However, the principle detailed above is adopted by almost all current codes, whether they are made available to the public or kept captively for internal uses by the organizations that hold the rights.

The object of the present invention is to provide a method for modeling the aeraulic flows in an environment occupied by a main fluid that is reliable while being less complex than the known methods both from the point of view of the technical resources needed and the view of knowledge required for a user to implement. In particular, the method according to the present invention aims to allow the assessment of the risks of airborne contamination in said field. Another object of the present invention is to provide a computerized system and a device for implementing said method.

The present invention relates to a method for modeling airflow in an environment according to claim 1.

The method according to the invention differs from any other method of numerical calculation of the approximate solutions to the Navier-Stokes equations (case of the conventional CFD "Computational Fluid Dynamics") by the fact that it makes the best of current technical knowledge to integrate them into a mathematical model and thus considerably reduce the number of iterations necessary to obtain an acceptable solution in the context of an assessment of the risk of airborne contamination.

This approach combined with a relevant mesh of space makes it possible to develop a method in two consecutive times in which a primary solution is first calculated in each cell and then a coupling algorithm is deployed on all the cells to refine the solution.

The principle of the method according to the invention consists in considering that in a given unit volume or mesh, defined sufficiently small, the final values of the primitive variables (V, P, T) are the resultant of their primary values calculated in an irrotational field composed by a "pressure - velocity" coupling with a turbulence model adapted to the Newtonian and incompressible fluids belonging to the area of validity of the process.

Furthermore, by integrating and parameterizing the information coming from sensors and measuring instruments such as laser telemetry - 3D scanner used to model the domain and the interfaces to be studied, the method according to the invention leads to the obtaining highly relevant solutions in the measurement where they are calculated in an environment faithful to the configuration of the interfaces they model.

In addition, the method according to the invention can be integrated into a computerized system and designed to be used by a much larger population than the experts to whom the conventional CFD software is addressed. The present invention therefore also relates to a computer program according to claim 15 and a device according to claim 16 or 17.

Also, by the performance of its calculation method and by its advanced ergonomics, the method and the computerized system according to the invention allow, in their field of application, a routine use of the modeling by the calculation of aerodynamic flows when the The information obtained by the experimental simulation is difficult to generalize or when the conventional computer-implemented CFD processes are too resource-intensive (memory, processor performance, etc.) to be applied efficiently and provide fast results.

The method according to the invention is especially particularly well suited for optimizing the design or the performance of an extraction and / or air blowing system or HVAC system (HVAC: Heating Ventilation and Air

Conditionning) in a building with strict specifications for air handling or controlling the risk of airborne contamination. Indeed, these two objectives do not fit well, for reasons of cost or security, treatment by experimentation or the implementation of heavy computing resources and inaccessible technicians directly in charge with the issues that concern them whereas the person skilled in the art knows that

the accumulation of interfaces with different levels of expertise in different technologies may preclude the relevance of the overall approach or even have a negative impact on the preservation of the confidentiality of the information. The method according to the invention also makes it possible, after modeling the aerodynamic flows of the study environment, to calculate the conditional probability (called plausibility) of having a quantity of contaminant, for example a collection of particles or a concentration of gas. present in a given volume of the environment at a given moment.

Thus, in a defined environment or area and thanks to the method according to the invention, the person skilled in the art can know accurately and quickly not only the aeraulic flow modes in this environment but also the possible dispersion profiles. 'an airborne substance with the

associated quantifications in space and time quantities.

An embodiment of the method according to the invention will now be described in detail by way of example with reference to the appended figures.

Figure 1 is a diagram illustrating the steps of the method according to the invention.

Figure 2a illustrates the result of a scan or scan of the study environment using a 3d laser scanner in four different positions and orientations.

FIG. 2b illustrates a modeling of the study environment in the form of a cloud of points obtained from the result of FIG. 2a.

Figure 3 illustrates an example of a study environment meshed by a finite number of unit volumes called meshes and whose sum forms the entirety of the environment.

FIG. 4 illustrates the forces exerted on a particle P of a

contaminant dispersed in the main fluid occupying the study environment.

FIGS. 5a to 5d illustrate an example of display of the results obtained by the method according to the invention. Figure 5a illustrates a qualitative dispersion profile of an airborne substance in a study environment, the profile being represented by a set of points. FIG. 5b illustrates the same qualitative dispersion profile as that of FIG. 5a, but being this time represented by a volume in three dimensions or "cloud". FIGS. 5c and 5d still show the same qualitative dispersion profile as that illustrated in FIG. 5b but superimposed on a modeling of airflow in the study environment.

With reference to FIG. 1, the method according to the invention comprises the steps E1 to E7.

The first step E1 of the method according to the invention consists in parameterizing and modeling the study environment in which it is desired to model the airflow flows.

To do this, it is necessary to collect the essential data relating to the study environment. These data include the physical and dynamic characteristics of the environment such as

the implantation of solid surfaces, the materials used, thermal equilibrium or local thermal insularities. This data can be acquired by telemetry, by scanning the environment by means of a 3d laser scanner or any other appropriate sensor for example or from pre-established plans of the environment and contained for example in a CAD file (computer aided design). These data are then used to create a faithful three-dimensional model of the study environment.

Said environment can be for example a laboratory, a clean room, an office, a public place such as a hospital or an airport hall or any other environment in which it is desired to model the flow of air.

In the following and to facilitate the description of the process according to the invention, it is assumed that the study environment is an indoor environment and that the fluid occupying this environment is air.

FIGS. 2a and 2b illustrate by an example a part of this first step E1 for a desk-type study environment. According to

structural characteristics of the environment, a series of scans / scans are for example made using a 3d laser scanner from various positions and in various orientations as shown in Figure 2a. A model of the environment consisting of a three-dimensional point cloud is then reconstructed using any appropriate software as illustrated for example in Figure 2b for the environment of Figure 2a.

This first step E1 modeling of the study environment is well known to those skilled in the art and is also an essential step in conventional CFD processes.

The use of 3D scanners and software that then make it possible to reconstruct the surfaces of the study environment as a three-dimensional point cloud are widely used in CFDs and are well known to those skilled in the art. These elements will therefore not be discussed in more detail here.

Also in this first step E1 of the method according to the invention, the technical characteristics of the blowing system and extraction of fluid in the environment - also called HVAC system (heating, ventilation and air conditioning) - are then determined. The system consists of aeraulic interfaces, the term aeraulic interface designating any element having a direct influence on the airflow in the study environment.

Aeraulic interfaces in the study environment include:

- the air vents of a ventilation system;

- the mouths of extraction or recovery of a system of

ventilation ;

- the surfaces of the environment having a significant thermal differential with the ambient temperature in the environment.

The technical characteristics of the blowing and extraction system include in particular the list of the air interface present and influential in the study environment, the implantation of said aeraulic interfaces in the environment and the aeraulic and climatic characteristics of said aeraulic interfaces. The aeraulic and climatic characteristics of the aeraulic interfaces forming part of an HVAC system are available to a person skilled in the art. Indeed, the contribution of each aeraulic interface at a given point of the environment, defined in relative coordinates with respect to this interface can be deduced from specialized technical works such as H. Recknagel, E.

Sprenger and E.-R. Schmarek, Climatic Engineering, Dunod, Paris, 2007, 1804p. (ISBN 978-2-10-048353-2) or technical data sheets published by the respective manufacturers of the air handling interfaces.

Finally, in this step E1 of modeling the environment, are still determined the characteristics of the main fluid occupying the study environment. In the case of air, these characteristics depend in particular on the location of the environment: altitude,

pollution ... The person skilled in the art will easily identify said characteristics.

It should be noted that even if step E1 of the method according to the invention has been described as a succession of substeps to be performed, the order of these substeps is of course irrelevant. In particular, it would be possible to begin by collecting the characteristics of the main fluid before modeling the study environment.

The study environment is the area of computation in which air flows will have to be modeled. That is, in which the fluid flow variables will have to be calculated. These variables are called primitive variables and include the velocity V = {u, v, w}, the pressure P and the temperature T.

As in a conventional CFD method, the determination of these primitive variables for the study environment is based on the discretization of space. To do this, the second step E2 of the method according to the invention consists in creating in the study environment at least one mesh network called Matrix. By definition, the term "mesh" designates a unitary volume type element in which the values of the primitive variables (V, T, P) (speed, temperature, pressure) are calculated. In extension, the term "mesh" designates the space formed by the union of all the meshes.

Preferably, in this step E2, an array of cells called a grid is also created.

The grid is a network of cells in two dimensions covering the surfaces of the environment and its infrastructures. This grid can be generated by any appropriate method, for example from the point cloud given by the 3D laser scanner after environmental scanning. The structural characteristics of the environment and ventilation interfaces are associated with the grid.

The environment is then meshed by at least one network of a finite number of meshes, the sum of which gives the whole of the study environment. This mesh will be called Matrix in the following.

Preferably, two meshes are generated in the study environment, said meshes being of different size and not superimposable. The Matrix then designates these two meshes. It will become clear later that although this is not necessary for the implementation of the method, there is an advantage to working with two distinct meshes, in particular in that, during calculations, the second mesh serves as a control for the implementation of the method. solution obtained in the first mesh and vice versa.

The matrix and grid as described above and their generation in an environment are known to those skilled in the art. In particular, the mesh or meshes can be hexahedral or tetrahedral according to the shape of the mesh component. In particular, a mesh or grid in an environment can be obtained by using the Delaunay triangulation. The fineness of the mesh and the grid is left to the choice of the skilled person and the user implementing the method, knowing that the smaller the size of the mesh, the more accurate the result will be.

Once the grid and the matrix have been created, it is necessary to determine the zones of exclusions inside or at the periphery thereof. These exclusion zones

are areas in which no calculation or modeling will be performed. This may include a closed volume placed inside the environment such as a cabinet, a pillar or other infrastructure.

The purpose of the method according to the invention is therefore to evaluate the primitive variables in each of the cells composing the matrix. According to the method which is the subject of the invention, this evaluation is done in two stages, described below in detail as third and fourth steps E3 and E4 of the process.

In a third step E3 of the method according to the invention, the primitive values (V, P, T) are first calculated in each of the meshs of the Matrix 15 in an irrotational field for a fluid supposed perfect and incompressible.

For this step E3 of the method, the fluid occupying the environment is therefore assumed to be perfect and incompressible and% M designates for each mesh M of the matrix a primitive primary variable calculable in the mesh M at point A. ¾ is determined by the infrastructure of the study environment. Of

Preferably, the point A is the center of the mesh M, but A can be any point of the mesh M.

We then calculate:

¾ ⁼ ^ ^~ ¾ ⁾

where ^T is the function calculating the value K of the primary primitive variable ^X M in the irrotational field for a supposedly perfect fluid.

The function ^τ is specific to the process according to the invention. For each mesh M of the matrix, it expresses for this mesh M the airflow result perceived as a weighted linear combination of all the interfaces. aeraulic environment that can influence this resultant. The fact of defining the function ^T only for an irrotational permanent field makes valid the hypothesis of linearity. It can indeed be demonstrated that a velocity potential is defined on an irrotational and simply connected fluid domain and that ^v = grad ( ^< P), \\ is interesting to note that if Φι and Φτ are two potentials on the same domain then Φι + 4> 2 is also a potential on this domain. Moreover, if the fluid is incompressible, ^ satisfies the Laplace equation:

Also, the function ^T is defined ar:

equation in which ( ^x is the value of the primary primitive variable ¾r at point A in the mesh M subjected to the influence of the air interface i; Pi represents the weighting factor of the air interface a in the mesh M .

(* t w and Pi are calculated in known manner from the technical characteristics of the aeraulic interfaces determined in the first step E1 of the method according to the invention.

This third step E3 of the method according to the invention allows, for any mesh M of the calculation domain to determine a raw value ^X M of the desired primitive variables. These values must then be refined by a coupling calculation of the respective interactions between the set of aeraulic interfaces, this time considering a perfect and incompressible fluid, but the main fluid whose characteristics have been determined in step E1. This refining is carried out in a fourth step E4 of the process according to the invention.

It should be noted that if the matrix is composed of two or more meshes of distinct size and not superimposable then the calculation of the third step E3 of the process is performed (simultaneously or not) in each of these meshes. As already mentioned above, this offers a possibility of controlling the solution obtained at this stage. Indeed, the convergence of the solutions obtained in each of the meshes makes it possible to ensure the coherence of the solution whereas the divergence of the solutions obtained can indicate that an error occurred in the computation.

The fourth step E4 of the method according to the invention therefore consists in refining the raw value of the primitive variables ^X M 'determined in the previous step E3.

To do this, we define the JV function calculating the value ¾ of the primitive variable ^X M - after coupling "pressure - velocity" and introducing a turbulence model adapted to the main fluid occupying the study environment. The chosen model therefore depends on the main fluid considered. The model used in the case of air or a Newtonian fluid is based on the Navier-Stokes equations.

The N function is structurally comparable to that built into conventional CFD codes. The function w is a digitization of the Navier - Stokes equations supporting the computation of the Newtonian and incompressible fluids and adapted to the Matrix developed in the second step E2 of the process.

This coupling step considers in the first place the equation of

conservation of momentum

said

ρ- = -νρ + μΑιι + F _B

ut

whereas the conservation equation of mass v. = o _is taken as the condition of control on which the convergence criterion is based.

For the numerical calculation, each of these equations is discretized on the Matrix, preferably on each of the two networks of meshes constituting it, said networks acting in first order and in second order.

The determination of solutions follows the principle of weak Leray solutions defined in a Sobolev space. This calculation is an iterative process to refine the values in each cell according to the values calculated in the neighboring cells until stabilized overall solutions are obtained.

Thus, at the end of this step, we obtain for all mesh M of the Matrix an evaluation of the primitive value ^X M at the point A of the mesh M. These primitive values therefore determine the aeraulic flows in the study environment.

The method according to the invention thus makes it possible to initiate the conventional phase of calculating the conventional CFD solution by the N function, which phase constitutes the fourth step E4 of the process according to the invention, on an approximate solution obtained by the developed ^T function. on each mesh of the Matrix constituting the domain of calculation. The method thus considerably reduces the number of iterations necessary to satisfy the acceptance criteria of the solution and limits the risk of divergence of the calculation. Indeed, in CFD

Conventional values of primitive variables in a given mesh are conditioned only by the system of initial conditions and boundary conditions as well as by code-specific models for which the user does not always have sufficient knowledge to be able to parameterize them. convenient way. It then appears that the value of the primitive variable in conventional CFD in any mesh from within the mesh domain is mainly controlled by the conditions introduced by the operator in charge of the calculation. As these are most often determined on the boundary of the domain, it follows immediately that the areas in which the interior volume grows faster than the surface of its border become both difficult to control (risk of obtaining a wrong solution for lack of framing parameters) and slow to converge towards the accepted solution.

In contrast, the method according to the invention ensures speed and relevance of the solution. The first calculation step E3 of the method according to the invention involving the function ^T makes it possible to obtain an enriched matrix giving improved initial conditions for launching the algorithm of the function N.

The conditions of the grid transposed in the enriched matrix are translated by the method according to the invention into initial dynamic characteristics in each cell (step E3). The principles of conservation of the mass and conservation of the momentum (contained in the Navier-Stokes equations) are then introduced into the calculation according to an iterative coupling mode on the near neighbors (step E4); note that the method according to the invention differentiates between first-order neighbors and higher order neighbors to increase the convergence speed of the solution. The iterative calculation is continued until a stable solution is obtained at the required accuracy.

The implementation of steps E1 to E4 of the method according to the invention thus makes it possible to model the airflow in a given environment and to calculate the flow of the homogeneous fluid in the study environment, this flow being characterized by the values calculated primitives ¾.

Once aerodynamic flows are modeled, it is possible to assess the risks of contamination in the study environment if a contaminant such as an airborne substance is dispersed there.

In the following, it is assumed that the contaminant is an airborne substance.

To evaluate this risk, the method according to the invention also comprises a fifth step E5 which consists in acquiring the data relating to the contaminant. The main parameters to be taken into account are the nature of the contaminant, the emission point defined with respect to the meshes, the emission conditions (speed, direction, etc.) as well as the contamination threshold or concentration limit value. that is to say the rate of presence of contaminant in the main fluid from which it is estimated that there is contamination. The sixth step E6 of the method according to the invention consists in determining a qualitative diffusion profile of the contaminant in the environment knowing the modeling of the aeraulic flows in the environment obtained in the fourth step E4 of the process.

The calculation of the diffusion or transport of airborne substances is based on the following assumptions:

^■ an airborne substance can decompose into a finite number of

individual components called Particles;

^■ the flow model of the carrier fluid (main fluid) is known in the vicinity of the Particle. Indeed, this model is obtained as a result of step E4 of the method according to the invention;

^■ the movement of the particle is expressed as the resultant of the forces acting on it, the force exerted by the carrier fluid is just one of them and its intensity will depend on the physical characteristics of the particle. FIG. 4 illustrates said forces exerted on a particle P:

^■ g represents gravity,

^■ f is the friction,

^■ e represents an optional electrostatic force,

^■ a refers to buoyancy;

^■ p is the thrust exerted by the air flow V flow into the environment on the particle;

^■ i denotes an inertial force.

A new mesh is generated to follow the trajectories of the particles. This mesh of the contaminant takes into account only the cells of the matrix in which the contaminant has an impact.

This approach makes it possible to calculate the particle trajectories, having the characteristics of the contaminating airborne substance, in the study environment by deriving the double advantage of having:

^■ an initial state based on the result obtained in step E4; ■ a "self-adaptive" contaminant mesh size and shape giving accuracy in the vicinity of the Particle and less refined values outside the direct sphere of influence of the airborne substance.

In addition, the property inherent in the method according to the invention of creating a "self-adaptive" mesh around the contaminant, that is to say a mesh whose boundaries are not predetermined, ensures that the solution will not remain confined. in the imposed volume before starting the calculation but only the grid representing the physical environment will limit its expansion.

In practice, the method according to the invention allows the calculation of an airborne substance release from a user-determined source (including the uncertainty of the position of the origin point) with the definition of concentration limit value giving a risk of exposure in each mesh M of the Matrix, for the zones within the limit envelope (ie for all the zones except for the exclusion zones defined in the step E1).

To further increase the relevance of the calculated solution, when it comes to determining the plausible amount of a given substance at a point in space, the method according to the invention performs a coupling of the Euler-Lagrange descriptions of the movement of the particles of the airborne substance in the fluid vector.

In Euler's description, by the resolution in step E4 of the algorithm associated with the function W, the method according to the invention determines the zones and the vorticity profiles of the airflow.

The method according to the invention records these singularities obtained during step E4 and reintroduces them during step E6 in the Lagrange description for monitoring the limit trajectories of the envelope of probability of presence of the contaminant at the determined concentration threshold. For each singularity

(vorticity or uncertainty node), the method according to the invention determines the the most unfavorable trajectory (that is to say the one that disperses at most the contaminating substance while remaining above the defined concentration threshold) and thus introduces the notion of plausibility of contamination. In the context of the method according to the invention, the plausibility differs from the probability by the following notions:

> the probability expresses the certainty fraction (for example on a scale of 0 to 1 or on a scale of 0% to 100%) to have a quantity of product present at a given point;

> the plausibility corresponds to a combination of conditional probabilities whose final result is a recovery of the space of the study area when the different parametric systems influencing the calculation of the probability have been compiled.

The method according to the invention integrates these uncertainty schemes into the Lagrange description algorithm for determining the product risk envelope. In this, the method according to the invention opens the concept of roCFD (roCFD: risk oriented Computational Fluid Dynamics).

Thus, the implementation of steps E1 to E6 makes it possible to determine a qualitative diffusion profile of the airborne substance with plausibility of

contamination.

In addition, in a seventh step E7, the method according to the invention allows the user to build, where he wishes, in the study environment a virtual volume in which the method according to the invention will make it possible to evaluate the amount of stagnant or through material at a given time as well as the plausibility of contamination based on the results obtained in steps E4 for the flow aeraulic and E6 for the diffusion profile of the contaminant. For example, in the case of the study of an environment in which a worker always occupies a fixed place, it is possible to construct a virtual volume at the position of the worker's head. The evaluation of the quantity of airborne substance stagnant or passing through said volume will correspond to an evaluation of the quantity of substance inhaled by said worker. On the basis of this It will be possible to determine whether or not to test or not to treat the worker.

The method according to the invention is preferably designed to model the aeraulic flow and the transport of airborne substances in a defined area of validity as follows:

• the main fluid is a Newtonian and incompressible gas;

• Airborne substances are present in low concentrations of

preferably less than 5% by weight in the main fluid and are: o either miscible with the fluid (then considered Newtonian gas and incompressible);

o be heterogeneous and in this case the substance must be such that the size of the particles composing the airborne substance is:

^■ large enough, preferably greater than 0.1 pm so that the fluid can be considered continuous to its scale;

^■ sufficiently small, preferably less than 1Ό00 pm, so that the trajectory of the substance can be influenced by the characteristics of the fluid in motion in which this substance is.

The method according to the invention is specifically parameterized to provide reliable results in this area of validity. However, the conceptual basis of the method according to the invention is not limited to this area alone. Other fields of application are indeed conceivable such as water, nanoparticles, mixtures inside a reactor .... It is enough to consider the equations of state of the dynamics of the corresponding fluids in place and Navier-Stokes Square.

The method according to the invention can advantageously be implemented by computer. Thus, the subject of the invention is also a computer program or a set of computer programs or computerized system for, once the program implemented in a processor, to model the airflow flows in an environment. Said program includes in particular the instructions necessary for carrying out the method according to the invention described above.

The program or set of programs may include

The algorithms for modeling the study environment from the essential environmental data transmitted to the program by the user,

• the algorithms for generating the grid and the matrix in the study environment according to the conditions defined by the user and,

The calculation algorithms making it possible to evaluate the primitive variables in the matrix according to the method and in particular the steps E3 and E4 described above.

The program or program package may further include

• the algorithms intended to evaluate the risks of contamination by a previously defined airborne substance and to define a profile of qualitative diffusion of said substance, and

Instructions or algorithms enabling the user to determine a virtual volume in the environment and to evaluate the quantities of stagnant or through airborne substances in said virtual volume at a given instant.

Those skilled in the art obviously understand that as regards the algorithms or program for modeling the environment and for generating the mesh, any appropriate algorithm can be used in particular the algorithms or programs known by it.

The invention also relates to a device for modeling airflow in an environment comprising a processor programmed to implement the method according to the invention. The computer program and the device may also include display means for displaying the result obtained by the method according to the invention. Any appropriate display mode of the airflow can be implemented. Thus, the modeling of the aeraulic flow in

the environment can be for example represented by graphic elements indicating the calculation point, the direction and the speed of the flow.

Figures 5a to 5d illustrate an example of application of the method according to the invention implemented by computer.

The computerized system is adapted to take as parameter and process the input data necessary for the calculation: data relating to the environment, the aeraulic interfaces, the main fluid and possibly the contaminant, limits of the domain of calculation. Once all the data necessary to implement the process transmitted to the system, it can perform the calculations (step E2 to E6 of the method).

In the example illustrated in FIGS. 5a to 5d, a dispersion profile of an airborne substance was determined after calculation of airflow in the study environment. The results of the calculations can be displayed by the computerized system in different forms. FIG. 5a illustrates the diffusion profile of the airborne substance at a given moment in the form of cloud points from the source of the substance S. FIG. 5b illustrates a diffusion profile of the same substance but this time in the form of a three-dimensional volume representing the dispersion over time and whose interior is evaluated as having a concentration of substance greater than the fixed limit only. In Figure 5b, said volume is superimposed on the cloud of points of Figure 5a. Finally, in FIGS. 5c and 5d, the dispersion profile of the airborne substance is superimposed on the airflow flow model of the environment. In these figures, the flow model is represented by graphic elements carrying the following information: ^■ the point of the environment in which the airflow has been calculated,

^■ the direction of said flow,

^■ the speed of said flow determined by a color scale. The difference between Figures 5c and 5d lies in the number of points in which the aeraulic flows have been calculated: there is more point in Figure 5c.

The process according to the invention has been described above by way of example only. In particular, the method has been described in the case where the main fluid occupying the study environment is air. It goes without saying that the method is in particular applicable to any incompressible Newtonian fluid.

Similarly, the method according to the invention has been described in the case where the study environment is an indoor environment. However, the method according to the invention is also applicable for an external environment provided that the conditions having an influence on the aeraulic flows are known and controlled.

Similarly, one skilled in the art obviously understands that the process according to the invention is applicable in the case where the contaminant is diffused from several sources as long as the contaminant streams are sufficiently diluted. Indeed, we can then assume that the total flow of contaminant at point A is equal to the sum of the flows emanating from the different sources at this point.

The advantage of the process according to the invention is twofold:

• by the optimal use of the memory blocks and by the reduction of the number of iterations necessary for the resolution of the equations dealing with

singularities, the method according to the invention implemented by computer decreases the calculation time while responding more to the risk management objective;

By the introduction of the concept of plausibility, the method according to the invention precisely meets the expectations of those in charge of security and the prevention of the risk of airborne contamination by determining the certainly uncontaminated, and therefore safe zones, by subtracting plausibility areas of presence.

This method marks a clear break with the set of computational approaches implemented in the field of aerodynamic flow modeling.

Indeed, this method offers the advantages:

• the speed of calculation, by reducing the number of iterations necessary to satisfy the convergence criteria;

· The control of the purely mathematical solution by the integration of

technical knowledge available;

• the explicit character for the user of the settings required by the code to define the computing environment;

• the automatic adaptation of the spatial volume in which the plausibility of presence and the associated concentration for a given contaminant are quantified in each unit cell of the domain.

In addition, the method according to the invention can therefore advantageously be implemented by computer.

None of the references of the prior art describes or mentions a principle comparable to that of this inventive method to allow, in any point of the field considered, a quantitative determination of the plausibility of finding an airborne material at this point. This quantitative determination must also be feasible by a man on the ground who is not necessarily an expert in

CFD modeling and the results be obtained in a time compatible with the constraints of routine use.

The fields of application of such a process concern in particular the defense, the protection of sensitive installations, hospitals, clean rooms, premises with controlled environment, pharmaceutical, cosmetic, electronic, agro-food and mechanical chemical industries.

Claims

Method for modeling aeraulic flows in an environment occupied by a main fluid, the method comprising the steps consisting of:

has) . Extract from the study environment the essential characteristics of said environment, its infrastructures, the aeraulic interfaces and the main fluid and model said study environment (E1);

b). Construct at least one mesh of the study environment composed of a finite number of unit volumes

(E2) whose sum represents the entire environment;

vs) . Calculate in each of the meshes of the mesh a raw value of the primitive variables sought by determining for a fluid assumed to be perfect and incompressible, the influence of the aeraulic interfaces on the primary primitive variables determined by the infrastructures of the study environment (E3);

d). Calculate in each of the cells of the mesh the primitive values modeling the aeraulic flows in the environment by discretizing on the mesh the equations of state of the fluid dynamics corresponding to the main fluid, the raw values of the primitive variables calculated in the previous step being considered as the initial conditions for these equations (E4).

Method according to claim 1 in which step a) comprises at least one scanning of the environment using a 3D laser scanner to obtain a modeling of the environment in the form of a three-dimensional point cloud.

3. Method according to one of the preceding claims, in which step b) comprises the construction of two distinct, non-superimposable meshes whose meshes are of different sizes and in which steps c) and d) are implemented on each of the meshes simultaneously or not.

4. Method according to one of the preceding claims, in which said raw values of the desired primitive values calculated in step c) and denoted ^X M' for each mesh M of the mesh, are obtained by solving the equation ¾ = ) in which the function J ^" is given by

Or

^X M designates the primary primitive variable in the mesh M of the mesh for a fluid assumed to be perfect and incompressible;

(*Î.)M is the value of the primary primitive variable ½ in the mesh M subject to the influence of the aeraulic interface i, i varying from 1 to the number of aeraulic interfaces in the environment;

Pi represents the weighting factor of the air interface i in the mesh M, i varying from 1 to the number of air interfaces in the environment;

(*i >M and Pi being determined from the characteristics of the air interfaces determined in step a).

5. Method according to one of the preceding claims, further comprising the steps consisting of:

e. Extract the essential characteristics of an airborne substance;

f. Calculate the dispersion profile of the airborne substance from a specific source in the environment; g. Determine an exposure risk (plausibility of contamination) in each mesh cell according to a predefined concentration limit value of the substance.

6. Method according to the preceding claim, in which the essential characteristics of the airborne substance include its physical characteristics and the initial conditions of dispersion of the airborne substance in the environment as well as the concentration limit value before contamination.

7. Method according to one of claims 5 or 6, in which step f) comprises the construction of an adaptive mesh around the airborne substance, the adaptive mesh being constructed to allow precise values to be calculated near the substance and less refined values outside the direct sphere of influence of the airborne substance.

8. Method according to one of claims 5 to 7 further comprising the step consisting of

h. Construct a virtual volume in the study environment and determine the qualitative dispersion profile of the airborne substance.

9. Method according to one of claims 5 to 8 in which the airborne substances are present in a concentration of less than 5% by mass in the main fluid.

10. Method according to one of claims 5 to 9, wherein the airborne substance is miscible in the main fluid, said fluid being a Newtonian and incompressible gas.

11. Method according to one of claims 5 to 8 in which the airborne substances are heterogeneous and the particle size is between 0.1 and 1000 pm.

12. Method according to one of claims 1 to 4 in which the main fluid occupying the study environment is a Newtonian and incompressible gas.

13. Method according to one of claims 1 to 10 in which the main fluid occupying the environment is air.

14. Method according to one of the preceding claims in which the study environment is an indoor environment.

15. Computer program for modeling aeraulic flows in an environment occupied by a main fluid comprising instructions for carrying out the method according to any one of claims 1 to 14.

16. Device for modeling airflow in an environment occupied by a main fluid comprising processing means programmed to carry out the method according to any one of claims 1 to 14.

17. Device for modeling aeraulic flows in an environment occupied by a main fluid comprising:

has). A processor designed to

i. Extract from the study environment the essential characteristics of said environment, its infrastructures, the aeraulic interfaces and the main fluid and model said study environment (E1);

ii. Construct at least one mesh of the study environment composed of a finite number of unit volumes whose sum represents the entire environment (E2);

iii. Calculate in each of the meshes of the mesh a raw value of the primitive variables sought by determining for a fluid assumed to be perfect and incompressible, the influence of the aeraulic interfaces on the primary primitive variables determined by the infrastructures of the study environment (E3); iv. Calculate in each of the meshes of the mesh the

primitive variables modeling the aeraulic flows in the environment by discretizing the state equations of fluid dynamics on the mesh

corresponding to the main fluid, the raw values of the primitive variables calculated in the previous step being considered as the initial conditions for these equations (E4).

18. Device according to claim 17, further comprising

display means for displaying the calculated primitive variables in the form of graphic elements which can be interpreted.