What’s New in EDEM 2022 – Release Highlights

David Curry · April 2022

We’re delighted to announce that EDEM 2022 is available now and is packed full of exciting new features! Whether you are dealing with fine powders, stalks and crops, or trying to optimize the design of large-scale systems then there is something in here for you. Here’s some of the highlights!

New EDEM Multi-GPU Solver Engine

Solver performance is key in granular materials simulation, and we know our users want accurate results in the shortest possible time. In recent releases performance has been a key focus for us and we have delivered in excess of 2.5 times speed up across the board with our CUDA-based solvers in the past 18 months. In addition, we introduced our super-fast ‘Hybrid precision’ solver that slashes run times whether you are using our polyhedral, sphero-cylinder, or multi-sphere shape options.

In EDEM 2022 we continue to push our solver by introducing a new EDEM multi-GPU solver! Developed using Nvidia CUDA, the EDEM multi-GPU means engineers can simulate much larger applications with more material than ever before.

20 million particles in a mill running on EDEM multi-GPU solver

Our new multi-GPU engine was designed from the ground up and allows you to make use of the latest Nvidia graphics cards. It is designed to be smart and will automatically balance computational workload and memory usage on your hardware. This means you get the highest utilization and performance without needing to fine-tune and adjust settings.

In our benchmarks we are seeing excellent scalability across our multi-million particle test cases – both with the computationally efficient multi-sphere shape, or with the computationally more demanding polyhedral shapes. The multi-GPU is available for all our particle types and is also compatible with our EDEM API.

Example of multi-GPU performance for 5, 15, 15, and 20 million particle simulations (based on a rotating mill)

Whether you are dealing with large numbers of fine particles or looking to simulate very large-scale applications, then our multi-GPU solver is for you!

If you have questions about how to get access to the right hardware to run our multi-GPU solver then please get in touch with your Altair account manager or reach out to support. They’d be happy to chat more about the range of options Altair offers for accessing multi-GPU hardware.

Additional GPU Solver Performance Enhancements

In addition to the multi-GPU solver, we continue to make general performance improvements to our GPU solver. This includes performance increases when running with size distributions and new options for tracking residence time on GPU, force and contact tracking enhancements, and factories. Please see the release notes for more detailed information.

Our focus on continued solver improvement brings additional speed increases across range of applications compared to 2021.2 release

Database for Powder Material Models

Nearly 6 years ago, EDEM delivered a world first with our GEMM Database – a vast library containing 1,000s of material models that users could search based on readily available real-world data.

In another first, we’re taking the same award-winning concept into the process industries sector by introducing our Database for Powder Material Models. Like the GEMM this is a library that offers 1,000s of material models designed for use in industries such as chemical, pharmaceutical, and food manufacturing sectors.

The EDEM Powder Material Model Database – gives users easy access to thousands of powder material models!

As anyone making engineering decisions based on DEM analysis knows, calibrating material model inputs is vital to give results that can be trusted. However, when it comes to powders getting the right set of inputs can be challenging due to the complex behaviors and wide variability they exhibit.

For many years, we have been working with our powder-focused customers to advise and guide on how to get the right calibrated inputs for their virtual material models. At times we have seen people view this step as a challenge (particularly for those new to DEM), or worse, are encouraged that calibration is unnecessary. In either case, the benefit that DEM simulation brings isn’t fully realized and so we have set out to distill and package our many years of know-how into an easy-to-use tool for those in powder-handling and processing industries.

The result is the EDEM Database for Powder Materials, and it addresses the concerns within these industries around lengthy calibration procedures, or how to even get started with DEM. Instead, everyone can access virtual material models they can trust within a matter of seconds.

Look out for future blog posts that outline more of this exciting, new capability!

Flexible-Fiber Modeling Improvements

Simulating flexible stalks, grasses, and fibrous materials has become even easier in EDEM 2022 thanks to updates to enable Meta-Particles to be composed of both Multi-Sphere and Sphero-Cylinder particles on GPU and mGPU. By using EDEMs Meta Particle technology, it is possible to create materials that will flex, bend, or even break.

Efficiently model stalks and fibers using EDEM’s meta-particle and Sphero-Cylinder capabilities

In addition, to make it easier to have equipment moving through a bed of fiber, it is now possible to use velocity preservation methods to ‘secure’ the base of the fibers in place relative to geometry objects and have the remainder of the particle free to move.

Increased Choice of Built-In Physics Models

EDEM 2022 further expands the scope of different applications that can be addressed by introducing a series of new built-in physics models.

Tavares UFRJ Breakage Model and Analysis Tools

Our world-leading Tavares UFRJ Breakage Model model has been developed based on previous well-established research at the Federal University of Rio de Janeiro (UFRJ) and is the result of close collaboration between Prof. L.M.Tavares, Prof. J.Ooi (University of Edinburgh) and the EDEM R&D Team. This model has now been integrated into EDEM to offer a streamlined user experience to set up and run breakage simulations.

Cone crushing using the Tavares UFRJ Breakage Model courtesy of Colombia Steel which allowed them to increase mass flow by 15% through design improvements

In addition, EDEM 2022 also provides new built-in EDEM Analyst tools for analyzing breakage results such as querying particle size distributions and the novel fines distribution analysis from UFRJ that eliminates the need to simulate at a super-fine scale.

Spray Coating Model

Ideal for use in applications such as tablet coating, EDEM now offers a built-in spray coating model. This model captures how a spray of coating material is transferred to particles during a simulation and is compatible with all particle types.

Spray Coating of tablets in a drum coater. Spray location and flow rates can be varied and EDEM will track the coating of each particle

Wet Mixing Model

Wet mixing is a critical activity in many industries, and EDEM offers several solutions for wet mixing and granulation through our API (check out the EDEM Forum on the Altair Community Knowledge Base to find out more). In this release of EDEM we have included one of the more popular models within the user interface. This model is ideal for the wet mixing of materials such as concrete and captures how moisture is transferred between particles and the resulting changes in cohesion.

Built in wet-mixing capability in EDEM 2022 - ideal for investigating complex particle-moisture interaction.

Volume Packing improvements

EDEM’s novel volume packing capability is a quick and easy way of producing packed beds of material with predefined compression in arbitrary shaped volumes. Since we introduced this capability, we are seeing it used extensively in many industries to reduce the set-up time needed to create a starting condition of material.

EDEM’s novel volume packing tool creates complex shaped beds of packed material quickly and easily

In EDEM 2022 we have further enhanced this capability with several new capabilities including support for user-defined particle size distributions, ability to use periodic boundaries, extended compatible physics models to include bond and heat transfer physics, as well as performance enhancements to further increase the speed of material generation.

Volume packing tool now supports periodic boundaries and user-defined size distributions

Whether you’re creating large beds of materials for agricultural or off-road vehicles to travel on, or you need to generate silos of material with specific solid fraction then the volume factory will offer a fast, effective way of doing this.

Unlimited Analysis with EDEMpy Custom Properties

In EDEM 2022 we are extending the functionality of our popular EDEMpy capability to allow unlimited custom properties to be introduced and modified as part of the post-processing stage of your simulation results. The result is that users can interrogate and report on a wider variety of data, even those very specific to their application, through the user interface.

The EDEMpy Custom Properties allows you to add, remove, and edit custom properties directly in the simulation file which means that this custom data can be visualized, animated, graphed, and exported like any other standard EDEM property.

Calculating particle residence time in a custom region of a simulation – custom properties defined entirely during post-processing

If you were previously running an API model to assign custom properties for post-processing, these APIs can now be removed from your processing chain and means the solver is focusing solely on crunching the numbers needed to simulate your application.

New Cone Penetrometer Calibration Kit

The EDEM team lead the way for virtual bulk material model calibration – we have always stressed the importance of this step for meaningful engineering design decisions. We also know that it can offer challenges for some users, particularly those new to bulk materials simulation. This is why our team were the first to create ‘Calibration Kits’ that guide users on how to run key virtual tests.

In EDEM 2022 we’ve extended the range of tests to include a new Cone Penetrometer test. This test is a popular one in applications involving off-highway vehicles – for example, tractors in a field or off-road vehicles traveling over rough terrain – to understand the behaviour of soils, sands, and other ground materials.

New Cone Penetrometer calibration kit is ideal for easy calibration of soils and other ground materials

The new EDEM Cone Penetrometer Calibration Kit contains an input deck for the test as well as the EDEMpy post processing tools. As with the other tests a report is provided and the test can be used with tools such as EDEMCal or HyperStudy to simplify the calibration process by setting up large numbers of jobs.

In addition, all our existing Calibration Kits have been updated to make use of the latest versions of EDEMpy and APIs.

And the Rest!

And that is just the start – in addition to the items shown here, there have been several additional improvements across the product. To find out more about what’s new in EDEM 2022 then check out the release notes that are available on Altair One.

phamad · April 2022

Personally, I think the CUDA-based Multi-GPU solver in EDEM 2022 is the most amazing feature to break the limits of DEM simulations.

Hope to see more detailed benchmarks from the EDEM team!

Thank you.

ElsayedRagab · May 2022

where is including of wet mixing model?

Stephen Cole · May 2022

ElsayedRagab said:
where is including of wet mixing model?

Hi,

You can find details of the wet mixing model if you go to Tools > Options > Additional Components and select "experimental contact models". You can then choose wet mixing from the physics particle-particle or particle-geometry section and set a moisture content in the Factory.

Regards

Stephen

Joshua García · May 2022

I have some problems with EDEMCalc in this version. For example, the jobs created parameters are incorrect, the interface create itself parameters, etc.

Joshua García · May 2022

And with the new version of EDEMPy in the sript of AoR i have the following error code:

runfile('C:/Users/Josh/OneDrive - Universidad de Las Palmas de Gran Canaria/8. Tesis/8.5 Calibracion de materiales/8.5.1 LDPE/Hollow cylinder/JobFileLocation/Job0/Static_angle_of_repose_analyst_v3.py', wdir='C:/Users/Josh/OneDrive - Universidad de Las Palmas de Gran Canaria/8. Tesis/8.5 Calibracion de materiales/8.5.1 LDPE/Hollow cylinder/JobFileLocation/Job0')
-------------------------------------------------------
Loading: Static_angle_of_repose_example0.dem
-------------------------------------------------------
-------------------------------------------------------
Processing: Static_angle_of_repose_example0.dem
-------------------------------------------------------
Traceback (most recent call last):

File ~\OneDrive - Universidad de Las Palmas de Gran Canaria\8. Tesis\8.5 Calibracion de materiales\8.5.1 LDPE\Hollow cylinder\JobFileLocation\Job0\Static_angle_of_repose_analyst_v3.py:73 in <module>
for n in range(len(deck.CreatorData[0].h5PTypes)):

AttributeError: 'Deck' object has no attribute 'CreatorData'

David Curry · May 2022

Joshua García said:
I have some problems with EDEMCalc in this version. For example, the jobs created parameters are incorrect, the interface create itself parameters, etc.

Hi Joshua - thanks for the note about this. I recommend you contact our support team - you can do that here - who will be able to help you.

David Curry · May 2022

Joshua García said:
And with the new version of EDEMPy in the sript of AoR i have the following error code:

runfile('C:/Users/Josh/OneDrive - Universidad de Las Palmas de Gran Canaria/8. Tesis/8.5 Calibracion de materiales/8.5.1 LDPE/Hollow cylinder/JobFileLocation/Job0/Static_angle_of_repose_analyst_v3.py', wdir='C:/Users/Josh/OneDrive - Universidad de Las Palmas de Gran Canaria/8. Tesis/8.5 Calibracion de materiales/8.5.1 LDPE/Hollow cylinder/JobFileLocation/Job0')
-------------------------------------------------------
Loading: Static_angle_of_repose_example0.dem
-------------------------------------------------------
-------------------------------------------------------
Processing: Static_angle_of_repose_example0.dem
-------------------------------------------------------
Traceback (most recent call last):

File ~\OneDrive - Universidad de Las Palmas de Gran Canaria\8. Tesis\8.5 Calibracion de materiales\8.5.1 LDPE\Hollow cylinder\JobFileLocation\Job0\Static_angle_of_repose_analyst_v3.py:73 in <module>
for n in range(len(deck.CreatorData[0].h5PTypes)):

AttributeError: 'Deck' object has no attribute 'CreatorData'

Hi Joshua - I recommend you contact our support team - you can do that here - who will be able to help you out. Thanks for getting in touch on this!

Stephen Cole · May 2022

Joshua García said:
And with the new version of EDEMPy in the sript of AoR i have the following error code:

runfile('C:/Users/Josh/OneDrive - Universidad de Las Palmas de Gran Canaria/8. Tesis/8.5 Calibracion de materiales/8.5.1 LDPE/Hollow cylinder/JobFileLocation/Job0/Static_angle_of_repose_analyst_v3.py', wdir='C:/Users/Josh/OneDrive - Universidad de Las Palmas de Gran Canaria/8. Tesis/8.5 Calibracion de materiales/8.5.1 LDPE/Hollow cylinder/JobFileLocation/Job0')
-------------------------------------------------------
Loading: Static_angle_of_repose_example0.dem
-------------------------------------------------------
-------------------------------------------------------
Processing: Static_angle_of_repose_example0.dem
-------------------------------------------------------
Traceback (most recent call last):

File ~\OneDrive - Universidad de Las Palmas de Gran Canaria\8. Tesis\8.5 Calibracion de materiales\8.5.1 LDPE\Hollow cylinder\JobFileLocation\Job0\Static_angle_of_repose_analyst_v3.py:73 in <module>
for n in range(len(deck.CreatorData[0].h5PTypes)):

AttributeError: 'Deck' object has no attribute 'CreatorData'

Hi Joshua, creatordata is now indexable, all references to creatordata can be replaced with creatordata[0].

Previously creatordata read the first time step infomation by default, now creatordata[n] can be set, this is if you change the creator settings after running the script can pickup the changes, however for like-for-like behaviour with older versions creatordata[0] can be set.