首页 > 软件应用 > 内容详情

EDEM1.3提供固液耦合的最佳方案


DEM Solutions EDEM 1.3

written by

Al Dean

 

EDEM allows users to simulate, analyse and visualise the behaviour of particles. When used in conjunction with CFD and FEA it can provide a much more accurate picture of a product’s performance.

 

Let’s talk simulation for a moment. If you look at the widely adopted tools that receive the most attention, you‘ll see that these are split between two technologies. Finite Element Analysis and Computational Fluid Dynamics are the best known for structural analysis and simulation of fluid flow/heat transfer respectively. Now, while these technologies are widely adopted and their benefits proven, the facts are that these two don’t allow you to simulate everything that products interact with – the behaviour of particles, an area that is being pioneered by a company called DEM Solutions.

DEM Solutions has developed a Discrete Element Modelling application called EDEM, which is used for the simulation,analysis and visualisation of particulate flows providing highresolution information on particle kinematics, momentum,heat and mass transfer. But what exactly is DEM?

Discrete Element Modelling is a particle-based simulation technology that is increasingly being deployed in design and optimisation of a broad range of industrial manufacturing operations.The Discrete Element Method (DEM) is a numerical method which models the movement and contact between each particle. A DEM model particle can represent either a single particle or group of particles in the physical system. The most common and physically accurate implementation of DEM is the ‘soft contact’ approach which calculates the forces acting on each particle using models which can account for the particle mechanical and surface properties.

DEM is very well established as a powerful tool for studying the mechanics of granular bulks. DEM software tools such as EDEM, can model both particles and equipment and provide a framework for investigating the relative effect of particle material properties, environmental conditions and equipment design. Factors such as particle shape, surface properties (e.g. cohesion and electrostatic charge), temperature and moisture content can be accounted for and particle-fluid flow can be modelled by co-simulation with Computational Fluid Dynamics (CFD). This capability allows simulation of processes such as fluidized beds and powder inhalers.

DEM can be used at a number of scales depending on the process and the information required. Typical DEM simulations on a desktop computer can model between 104 - 106 particles. Parallel computing on a cluster increases this to 105 – 107 particles depending on the process time simulated. DEM simulation must therefore be used relative to the particle size, scale of the process equipment and the phenomena being studied. A common misconception is that it is necessary to model at full-scale in order to properly simulate a process. Given that the number of powder particles can be of the order of 1012 in some processes then it would appear that DEM has limited application. While it would certainly be of value to be able to model

every particle in a full-scale application there is a wealth of information and understanding to be gained by focusing on regions of interest. It is also possible to exploit planes of symmetry and to scale up model particles to represent the volume occupied by a larger volume of actual particles. While these techniques must be used with care they enable a very wide range of problems to be modelled.

DEM is now being applied to a wide range of industrial particulate solids handling and processing problems. The growing uptake by the process industries is rapidly expanding its application beyond the more traditional areas of mining and geo-mechanics to pharmaceuticals, chemicals, oil and gas for example. Continuing improvements in the price/performance of high performance computing makes DEM a realistic tool for use in a broad range of industrial product and process design and optimisation.

EDEM is an out of the box commercial software that runs identically on both Windows and Linux (both 32-bit and 64-bit) and is split into three applications the user interface for which is common across all three. These cover the pre processing, solver and post-processing that most analysis users will be familiar with. So lets run through the workflow and see what it can do.

 

EDEM Creator

As with all simulation tasks, the first task is to define the conditions and behaviour of simulation. Within EDEM, this is done in four steps. The first is to define the global conditions you’re working within – this includes standard properties such as gravity.

5An example of a mill with 100,000 particles. The mill was run at various rotational speeds to investigate particle-particle and particle-wall impacts as a function of mill speed and particle size distribution. Periodic boundaries are applied along the axis of the mill so only a slice of the whole mill is shown

 

 

You then need to define all of the materials that exist in your simulation job. Because we’re dealing with the interaction between particles and other objects, you have a slightly different set of inputs than for traditional analysis. Alongside the normal density, you also require Shear Modulus, Coefficient of Restitution and Coefficient of Static Friction as well as the coefficient of Rolling friction. Once your global set-up is done, the next step is to define the particles within your system. To optimise the process, EDEM uses a volumetrically accurate, user definable description of a particle (and remember that when we’re discussing particles, you can model almost anything from nano-scale ingredients in the pharmaceutical industry, through to potatoes - yes, it has been done). The basic particle form is a simple sphere and in many instances, this might suffice. Should you want something more specific, you can use the modelling tools within the system to create more descriptive particles using a series of spheres. Handily, you can load a CAD file, perhaps as a reverse engineering STL file of your typical particle and use that as a reference. You then calculate the properties such as mass, volume, moment of inertia XY and Z – much of these can be extracted from the model or loaded from an integral template.

The third is to define the physical objects, rather than particles. The first is the domain, a term CFD users will be familiar with. The Domain is the work area in which you’re working and is defined in terms of X, Y and Z dimensions. This allows you to restrict your focus on the critical areas of what could be a very complex CAD geometry of a whole product or process. If you’re looking at mixing, that’s the tank and the blades. If you’re working to discover the loading conditions on a backhoe loader, then it’s the form of the bucket. This can be translated from your CAD models via IGES, STEP, Pro/Engineer, Fluent mesh file, Neutral, STL, ACIS, Parasolid, Ansys, Catia V5 (CATpart and CATproduct). As with most analysis processes, the system uses a tessellated format to hold the geometry, so you have controls over that process.

If you have moving components within your simulation job,you then define the movement. This can be done within EDEM,or you can link to a kinematic application such as Adams.

The final set-up stage is the creation of Factories. Factories are the entities which produce the particles and introduce them into the simulation. Factories are Dynamic or Static in nature, where the particles are introduced over time in a single time step. You also define the nature of particles. What’s key to note is that because the system can simulate multiple particles within a single simulation, you need to define variations of particle size, type using a number of methods. You also need to provide position,trajectory, velocity and orientation of the particles as they are introduced into the simulation.

While the above might seem somewhat intimidating, the truth is that the whole process is highly graphical and assuming that you have a solid level of understanding of both the product, its operating conditions and the process you’re looking to simulate, the whole set-up process is very efficient.

 

EDEM Simulator

The next step, once you have your simulation defined, is to calculate it. Here, the system is again pretty self explanatory. There are three tabs in the EDEM simulator application. The first is ‘ Time’ and deals with setting up the period over which you want to run the simulation. You input the time step, the total Run time (i.e.; what period you want the simulation to occur over, rather than the calculation time), then give it a Write-Out Period. This is key to ensuring that you have a workable amount of data. By writing out the results at a specified value (in seconds), you can collect the amount of information you want, without over extending the calculation time. The next is the Space tab. The domain in which youre working needs to be split into a grid of cells to assist with calculating the interaction between the particles. These cells should be around the size of your particles. The final stage is to set-up the simulation run, in which you define which simulation contact model youre using. The default Hertz-Mindlin model covers a wide range of applications but this can be substituted to cover specific requirements such as cohesion, conveyor belts or electrostatics.

As ever, you then hit run and leave the system running. The results can be viewed as they are calculated, allowing you to inspect the odd time step and see how things are looking as the system calculates how the particles and geometry are interacting. If something seems wrong, you can stop the analysis and rework some of the parameters without having to rebuild the model.

The final application in the EDEM suite is EDEM Analyst, which provides the post processing or results inspection and visualisation tools. You have full animation capabilities, so you can inspect the movement and interaction of your particles within the simulation along with colouring options, so you can colour particles by any of their inherent attributes (velocity, force, inertia etc). Of course, you’ll want to dive in and find out much more than just a general overview of the simulation, so there are clipping planes and such. The Streaming

tools, allow you to display the path that individual particles take over the specific time frame, which is extremely useful for tracing movement within a system. For example, alongside simple movement, it can also assist with tracing particles from an end condition/position, then trace backwards through time frames to find where they originated from, to assist in understanding and defining the key stages of any simulation. The greater sequenced insights provide a more detailed analysis for further investigation if necessary.

But, while visualising results is useful, for many, text export or charting tools are the most useful in terms of extracting relevant data. EDEM includes a full graphing suite that allows you to create the graphs and chart representations you need and is specialised for that purpose, rather than general purpose. But should you need to, you can also export directly out to Matlab and Excel. One thing to note is that the charting tools are not live; rather they run as a separate post-process activity. They don’t update as you’re working through the calculation - but you can stop the calculation at any point, run the Analyst tools to get a feel for the accuracy of the results you’ve got so far (as a sanity check if you will) and then continue - you can also dive back into the Creator tool, make a change to the inputs or variables and then continue the calculation.

5Distribution of particle velocity in a screw auger emptying from a hopper, using 200,000 particles.

 

In conclusion

Discrete Element Modelling is a new technology to the mainstream. To date, its use has been within several key application areas or industries, namely, pharmaceutical, chemical, oil and gas as well as mineral and materials processing. These industries are typically dealing with high value processes where design optimisation and development efficiency is key but resolution of production issues are often mission critical and extremely costly. But the fact remains that particle simulation is applicable to a much wider audience.

When you’re looking at any simulation process, there are assumptions which must be made, to both make the simulation process more efficient, but also to work around the limitation of existing tools. Consider a good example – that of earth moving equipment. If you use a static load on the bucket of an earth mover using FEA techniques, that load is static, but the reality is that the earth the bucket passes through is dynamic in nature, because the particles in the earth shift. Using FEA alone, you’re using a heavy assumption because that’s all the system can handle. But, using a system like DEM, you can simulate the movement of those particles and in turn, the loading on your bucket will be much more accurate.

EDEM is perhaps ahead of the game. While other codes are out there, this is the first time that I’ve seen a system that’s mainstream focussed. Yes, the parameters and variables you’re working with differ from more mainstream modeling tools, but the fact is that if you have a solid understanding of your product, process or application, then it’s eminently usable by anyone developing products where particles and their interaction are essential to that product’s performance. It interacts with other more traditional simulation tools allowing you to transfer data between applications. Key to EDEM is that if used alone it allows you to simulate particle dynamics, but in conjunction with other tools allows you to create a more accurate simulation. And from that, you can use that information to create optimised and higher quality products and that means one thing – competitive advantage.

 

 




您的姓名: 

验证码: 
看不清,点击图片换一张!
评论内容: