EEPA | John P. Morrissey

Calibrating DEM simulations of cohesive solids with the Uniaxial Powder Tester

Thu, 19 Oct 2017 00:00:00 +0000

This post was originally published on www.edemsimulation.com.

Introduction

Over the years, as the popularity of the Discrete Element Method (DEM) has continued to grow, it has been increasingly used as a design tool. The accuracy of these models largely depends on the chosen DEM model and its parameters, and as pointed out in a post by David Curry (What numbers should I put in? The perennial question for DEM users) the adage of “garbage in – garbage out” holds true when carrying out DEM simulations. For cohesionless materials we have a range of parameters to account for - size, shape, stiffness and friction being some of the more commonly considered. But how many materials are truly cohesionless? Not many, so we end up having to add some additional parameters to our DEM simulations to account for the additional cohesive forces in the material. The origin of these cohesive forces and their respective mechanics vary from long and short range non-contact interaction forces, mechanical interlock, electrostatic forces, solid bridges and liquid bridges to name but a few.

So how does one go about capturing all of these phenomena in their DEM simulation? The first step is selecting a suitable contact model – a liquid bridge model is not much use for capturing the behaviour of fine particles in an electrostatic field. Provided you have negotiated that stage and selected a suitable contact model; you are then faced with the much more difficult proposition of model calibration.

Calibration

To some, “calibration” is a dirty word that simply means ‘cheating until you get the right answer’, but we can define it more concisely as “the process of adjusting physical modelling parameters in the computational model to improve agreement with experimental data” ¹. Calibration isn’t a bad thing, it’s a necessary part of scientific life – we need to regularly calibrate our measuring instruments to ensure accuracy. Due to the assumptions we make in DEM; in relation to particle shape and number of particles we use to represent a system amongst other things; we need to calibrate our input parameters such that they provide a realistic result.

And much like the choice of contact model, the choice of calibration experiment is also important – you would not choose a quasi-static experiment to calibrate for a very dynamic regime as you may fail to capture much of the key behaviour such as impact dynamics. Obtaining suitable calibration data can be a challenge as one would like them to be both easily obtained and be highly repeatable.

The typical process of calibrating DEM contact model parameters involves tuning the observed response of the simulation to closely reproduce the macroscopic bulk behaviour measured in the calibration experiment. The mechanical behaviour of cohesive powders can be carefully measured using element tests such as biaxial test, true triaxial and hollow cylinder tests. However, in practice these tests can be expensive and slow to conduct and are rarely performed for industrial applications requiring material characterisation. More typically a test such as direct shear, ring shear, unconfined compression, triaxial compression or cone penetration is carried out.

The Case for Uniaxial Testers

For example, in industry the flow function of a cohesive solid is typically measured using either a ring shear tester or Jenike shear tester but these have some drawbacks. The Jenike test, whilst appearing to be quite straight forward is time consuming and can often be highly operator sensitive, both of which are problems when you want to have a high level of confidence in the result to use for calibration. Ring shear testers remove the operator sensitivity and repeatability problem by being highly automated, computer controlled devices but these can be very expensive to purchase and maintain. They also have particle size restrictions due to their small volume and may be limited to relatively low stress ranges which may not cover the range of interest.

Uniaxial testers offer an alternative method for characterising the flowability of powders and fine granular materials in industrial situations. Uniaxial testers apply a consolidation stress vertically and also apply the failure load vertically. This is appealing for two reasons: firstly it’s a mechanically simple device that’s easy to operate and secondly it’s physical similarity to the stress path associated with arching and the unconfined yield strength ². One of the challenges with uniaxial testers is achieving a uniform level of consolidation throughout the depth of the sample, as wall friction can mean that the vertical stress in the sample at the bottom is significantly lower than the vertical stress applied at the top surface. To combat this there have been many designs considered over the years to deal this, each with their own strengths ³^,⁴^,⁵.

The University of Edinburgh had previous experience of developing successful uniaxial testers for large diameter particles ⁶. From this experience the Edinburgh Powder Tester (EPT) was developed for measuring the flowability of highly compressible, high value powders. While many previous efforts have focussed heavily on accurate laboratory measurement of unconfined yield strength, the Edinburgh testers aimed at speed, robustness and high repeatability for industrial use, with a close match to a Jenike cell being a secondary objective ². The key differences between the Edinburgh Powder Tester and some previous devices lie in the attention to mechanical details for both the consolidation and failure load application and the strategic intent ², which means the EPT offers repeatability results in the range of 5-10% relative standard deviation (RSD) for a vast range of materials ⁷^, ⁸.

Edinburgh Powder Tester (EPT)

Consolidated sample exposed prior to failure

Failed sample

It is the simplicity and flexibility of uniaxial testers, coupled with a high level of repeatability, that make uniaxial testers ideal tools for calibrating DEM simulations of cohesive solids. For example, the EPT, which is a semi-automated uniaxial tester, can provide rapid measurements of various bulk mechanical properties of powders, including the stress–strain and the stress–porosity response during confined compression, the stress–strain response during unconfined compression including the peak unconfined strength and the caking strength over time consolidation, all of which can be utilised for calibrating you DEM simulation of a cohesive solid.

Recently, Freeman Technology; which has many years of experience of both powder testing and development of testing apparatus, including extensive knowledge of shear cells; entered into a highly productive collaboration with the University of Edinburgh, DuPont and The Chemours Company to develop a uniaxial powder tester for the commercial market that incorporates the key design elements of the EPT. The result of this collaboration is ‘The New Uniaxial Powder Tester from Freeman Technology’ which is a standalone uniaxial shear tester for powders, capable of testing a stress range of up to 100 kPa. The UPT comes in two versions: a manual version, which is ideally suited for quality control measurement in industrial situations and the advanced version with increased automation and reduction in operator inputs and advanced data logging capabilities which makes it very suitable as a DEM calibration tool for cohesive powders.

Freeman UPT

Freeman UPT Consolidation Station

A DEM Example Case

At the University of Edinburgh, we have used uniaxial testers (EPT/UPT) extensively for calibrating DEM parameters for the Edinburgh Cohesion Model in EDEM ⁷^,⁸^,⁹^,¹⁰^,¹¹^,¹² (which is currently available on the EDEM User forum and will be part of future versions of EDEM) due to the high level of repeatability and the flexibility to measure key properties quickly and easily in the laboratory.

As an example, the measured stress-strain curves from the EPT for an iron ore solid at varying moisture contents (levels of cohesion) are shown in Figure 6 and Figure 7. A typical set of unconfined test results that make up a single uniaxial flow function (the peak strength vs. the consolidation stress) are shown for the same iron ore fines in Figure 8 for a single moisture content.

Confined Stress Strain measured in EPT for iron ore fines

Unconfined Stress-Strain measured in EPT for iron ore fines

Measured bulk density variation of iron ore fines

These measured values provide the range of various properties to be captured within the DEM simulation. A typical DEM simulation of a cohesive solid in the EPT is shown in Figure 9. The DEM simulation consists of the filling, confined compression, unloading of the sample and the crushing to failure of the unconfined sample.

Figure 9 - DEM simulation of cohesive solid in the EPT: a) Filling b) Confined consolidation c) Unloading and removal of confining sleeve d) Loading to Unconfined failure (uUYS)

The measured properties provide the necessary information to define the loading stiffness, plasticity ratio and whether the material is exhibiting linear or non-linear stiffness, bulk density and level of cohesion to calibrate the DEM model. In order to capture the behaviour at different moisture contents, we need to calibrate the level of cohesion for each moisture content at one point on the flow function, as in Figure 10.

Experimental results and calibrated DEM results

Once calibrated, it is possible capture the entire flow function spectrum for varying moisture contents with a good degree of accuracy, as shown in Figure 11, with the effect of moisture influencing only one model parameter i.e. the contact surface energy $\delta \gamma $.
This process provides a fully calibrated DEM model for cohesive powders for use in DEM simulations.

Calibrated Flow function vs Experimental flow function

References

American Society of Mechanical Engineers., Guide for verification and validation in computational solid mechanics, American Society of Mechanical Engineers, 2006. https://www.asme.org/products/codes-standards/v-v-10-2006-guide-verification-validation (accessed July 13, 2017). ↩︎
T.A. Bell, E.J. Catalano, Z. Zhong, J.Y. Ooi, J.M. Rotter, Evaluation of the Edinburgh powder tester, in: PARTEC 2007 - Congr. Part. Technol., Nürnberg, 2007: pp. 1–6. http://scholar.google.com/scholar?hl=en&btnG=Search&q=intitle:Evaluation+of+the+Edinburgh+Powder+Tester#0 (accessed October 5, 2013). ↩︎
L.P. Maltby, G.G. Enstad, Uniaxial Tester for Quality Control and Flow Property Characterization of Powders, Powder Handl. Process. 13 (1993) 135–139. ↩︎
M. Röck, M. Ostendorf, J. Schwedes, Development of an Uniaxial Caking Tester, Chem. Eng. Technol. 29 (2006) 679–685. doi:10.1002/ceat.200600068. ↩︎
R.E. Freeman, X. Fu, The Development of a Compact Uniaxial Tester, in: Part. Syst. Anal. 2011, Edinburgh, UK, 2011: pp. 1–6. ↩︎
Z. Zhong, J.Y. Ooi, J.M. Rotter, Predicting the handlability of a coal blend from measurements on the source coals, Fuel. 84 (2005) 2267–2274. doi:10.1016/j.fuel.2005.05.023. ↩︎
S.C. Thakur, H. Ahmadian, J. Sun, J.Y. Ooi, An experimental and numerical study of packing, compression, and caking behaviour of detergent powders, Particuology. 12 (2014) 2–12. doi:10.1016/j.partic.2013.06.009. ↩︎
J.P. Morrissey, Discrete Element Modelling of Iron Ore Fines to Include the Effects of Moisture and Fines, University of Edinburgh, 2013. https://www.era.lib.ed.ac.uk/handle/1842/8270. ↩︎
S.C. Thakur, J.P. Morrissey, J. Sun, J.F. Chen, J.Y. Ooi, Micromechanical analysis of cohesive granular materials using discrete element method with an adhesive elasto-plastic contact model, Granul. Matter. 16 (2014) 383–400. doi:10.1007/s10035-014-0506-4. ↩︎
S.C. Thakur, J.Y. Ooi, H. Ahmadian, Scaling of discrete element model parameters for cohesionless and cohesive solid, Powder Technol. (2015). doi:10.1016/j.powtec.2015.05.051. ↩︎
A. Janda, J.Y. Ooi, DEM modeling of cone penetration and unconfined compression in cohesive solids, Powder Technol. 293 (2016) 60–68. doi:10.1016/j.powtec.2015.05.034. ↩︎
J.P. Morrissey, J.Y. Ooi, J.F. Chen, K.T. Tano, G. Horrigmoe, Measurement and prediction of compression and shear behavior of wet iron ore fines, in: 7th World Congr. Part. Technol., Beijing, China, 2014: p. 8. ↩︎

The Edinburgh Elasto-Plastic Adhesion (EEPA) Model... in EDEM 2018!

Tue, 10 Oct 2017 00:00:00 +0000

This post was originally published on www.edemsimulation.com.

Since Altair’s acquisition of EDEM in 2020 the original post has been taken down. A cached copy of that post can be found here.

Introduction

EDEM has just announced the new 2018 version of the software that includes many new features with a focus on performance on productivity. I don’t normally show much interest in the release of new version of software but this is a bit different. I’m excited to announce that the Edinburgh Elasto-Plastic Adhesion (EEPA) Model¹^,² which started it’s life in my Ph.D.

This is a significant milestone and highlights how popular the model has become since it’s release as an API model to EDEM users just over 3 years ago. EDEM is considered the leading Discrete Element Method software application for bulk material simulation in the market and is used my many companies across fields as varied as mining, food manufacturing and pharmaceuticals. EDEM is also used by thousands of researchers in more then 200 universities around the world.

In parallel, a new contact model for modeling complex cohesive materials such as fine dry powders, organic materials, soil and ore fines is now available as a standard built-in contact model in EDEM. This model, called Edinburgh Elasto-Plastic Adhesion (EEPA), offers a solution for cohesive granular solids whose behavior changes depending on the stresses experienced by the material beforehand. It can help realistically simulate applications such as material adhesion to earthmoving equipment, soil-tyre interaction or for instance a cohesive powder compaction process such as tabletting.

This will allow users to take advantage of the EEPA model without the computational cost of running the API. This could be as much as 30% reduction in runtime for your simulations - effectively for free!! All of the features of the API model ³^,⁴ exist and are nicely integrated into the GUI. There should also be a GPU implementation of the modelling making bigger simulations more tractable.

EEPA now available in Contact Model Physics

Inputting EEPA Parameters

I hope this helps making simulation of cohesive granular materials more accessible to more user and that the models popularity continues to grow.

Citations

John P. Morrissey, 2013. Discrete Element Modelling of Iron Ore Fines to Include the Effects of Moisture and Fines. Ph.D. Thesis , p.331. Available at: https://era.ed.ac.uk/handle/1842/8270

Cite PDF Project

S. C. Thakur, John P. Morrissey, Jin Sun, Jian-Fei Chen, Jin Y. Ooi, 2014. Micromechanical analysis of cohesive granular materials using discrete element method with an adhesive elasto-plastic contact model. Granular Matter , vol.16, pp.383-400. Available at: http://link.springer.com/10.1007/s10035-014-0506-4

Cite PDF Project DOI

References

Morrissey, J.P., Discrete Element Modelling of Iron Ore Fines to Include the Effects of Moisture and Fines, University of Edinburgh, 2013. https://www.era.lib.ed.ac.uk/handle/1842/8270. ↩︎
Thakur, S.C., Morrissey, J.P., Sun, J., Chen, J.F., & Ooi, J.Y. (2014). Micromechanical analysis of cohesive granular materials using discrete element method with an adhesive elasto-plastic contact model. Granular Matter, 16(3), 383–400. doi:10.1007/s10035-014-0506-4 ↩︎
Morrissey, J.P., Thakur, S., & Ooi, J.Y. (2014a). EDEM Contact Model: Adhesive Elasto-Plastic Model. University of Edinburgh. doi:10.13140/RG.2.2.10139.18724 ↩︎
Morrissey, J.P., Thakur, S., & Ooi, J.Y. (2014b). Using the Elasto-plastic Adhesion Model: Example Problem. University of Edinburgh. doi:10.13140/RG.2.2.10978.04809 ↩︎

Models for the Manufacturing of Particulate Process (Models MPP)

Wed, 27 Apr 2016 00:00:00 +0000

Introduction

Particle processes in industry are both common and challenging. The complexity of particulate systems means that it is prohibitively difficult to gain a comprehensive understanding of the particulate mechanics which control a process using solely laboratory experiments. Even though some numerical models are available in the literature which describe particle processes, these tend to be inadequate to meet the requirements of industry. Industry requires robust models which contain the key relevant physics, which are user-friendly and which are focused on delivering the output product specifications given a limited number of measured input parameters.

Despite the significant academic activity in the development of multi-scale, multi-phase modelling approaches, few of the models originating in academia are fully implemented in industrial practice. As a result, the full benefits of model-based development have not been derived. The creation and implementation of a better approach to implementing useful models from academia into industrial practice could therefore lead to significant economic benefits relating to enhanced speed of development, reduced development costs and improved product quality.

This project aims to create a generally applicable framework for transferring academic innovations in the modelling of particulate materials into industrial practice in the UK. The process of twin-screw granulation has been selected as an exemplar industrial process which is simulated across multiple scales using the coupled methods of population balance modelling and the discrete element method.

Project Members

The project involved two academic partners, the University of Edinburgh and the University of Sheffield, and six industrial partners: Pfizer, AstraZeneca, Procter and Gamble, Johnson Matthey, Process Systems Enterprise and DEM Solutions (Now Altair EDEM). The project was led by the Centre for Process Innovation with funding coming from Innovate UK.

Project Summary

The two-year Models for the Manufacturing of Particulate Process (Models MPP) project focused on establishing a generic framework and core capability to improve industrial productivity.

As part of the drive to provide more leverage and integration of the wealth of formulations expertise in the UK, the £700,000 project, funded by Innovate UK as part of CPI’s National Formulation Centre Strategic Projects Programme, has facilitated the translation of the UK’s world-leading expertise in computational modelling and simulation for the manufacturing of particulate processes.

Connecting industry and university research groups with state-of-the-art technology has resulted in the development of a generic framework for translating particle models of industrial relevance into industrial practice. The developed framework incorporates a decision support methodology/tool for models development that will allow easier access and adoption by leading UK-based industrial organisations. These models are expected to be deployed within industry to accelerate the development of innovative products and improve productivity and cost-effectiveness of product manufacturing processes.

This framework will provide guidance for the coupling of well-established computer modelling techniques at different size scales, to enable development scientists and engineers to develop models that simulate different parts of manufacturing processes from the dispensing of input materials through mixing, agglomeration and other subsequent processing steps.

Twin-screw wet granulation, a method used in a number of industry sectors, was used as an exemplar. Wet granulation is a process in which small primary particles are joined together using agitation and a liquid binder. The purpose is to improve the properties of very fine cohesive powders used in formulated products such as pharmaceuticals, ceramics, detergents and fertilizers. Granulation is commonly used in the pharmaceutical industry during tablet manufacturing. Fine powders are granulated to improve flow prior to tableting and reduce the potential for dusting and other production issues. The formation of granules also helps to reduce segregation and improve the content uniformity of the final product which is critical for consistent product performance.

The academic partners at the Universities of Edinburgh and Sheffield conducted a review of the current state of the art and used this insight to develop coupled models using EDEM and PSE software platforms and the technology vendor’s expertise in developing industrially robust models. These models were validated using real life trials at the industrial partner’s production facilities.

While granulation is an important process in many industries, including pharmaceuticals, foods and agrochemicals, the framework provided by Models MPP means it can be broadly re-applied to other manufacturing operations and industries.

The legacy capability established from this project will be applied to other CPI strategic projects, e.g. in the development of a digital twin with a sister National Formulation Centre strategic project focused on a fully PAT enabled twin screw extruder (‘PROSPECT CP’).

Project Gallery

EDEM Contact Model: Adhesive Elasto-Plastic Model

Wed, 18 Jun 2014 00:00:00 +0000

Click the Cite button above to get publication metadata for your reference management software in .bib format.

Using the Elasto-plastic Adhesion Model: Example Problem

Wed, 18 Jun 2014 00:00:00 +0000

Click the Cite button above to get publication metadata for your reference management software in .bib format.

Micromechanical analysis of cohesive granular materials using discrete element method with an adhesive elasto-plastic contact model

Thu, 01 May 2014 00:00:00 +0000

Click the Cite button above to get publication metadata for your reference management software in .bib format.

Transporting, handling and storing behaviour of iron ore fines

Sat, 27 Jul 2013 00:00:00 +0000

Introduction

This project attempts to deal with the challenges associated with handling and storage of cohesive solids in the mining industry. An adhesive-frictional model has been recently developed for DEM simulation of cohesive particles at the University of Edinburgh. This project will exploit the new method for modelling cohesive particulates for specific problems, such as effect of fines in silo discharge and the effect of time consolidation.

Fines are produced during transportation, handling and storing of iron pellets. The presence of fines can significantly affect the behaviour of iron ore pellets during these processes. In addition, increased volumes of sinter fines are being produced and handled and their behaviour can also be significantly affected.
Silos and other equipment handling the fines can experience such phenomena as caking, arching and alteration of flow pattern, leading to operational difficulties.

The behaviour of fines is extremely complex, affected by numerous factors such as particle size, size distribution, stress level (e.g. the maximum vertical stress in a silo), stress state (i.e. the stresses in different directions), stress history, temperature, moisture content, time and chemical bond/fusion. Different combinations of these factors can lead to different behaviours, such as caking with different strength and size, which in turn result in different effects on the solids flow in silos and other handling scenarios in the mining industry.

DEM simulations can also be affected by the prescribed boundary conditions and as such the effect of oversimplification of the degrees of freedom for a Jenike cell are explored in detail, through the use of fully coupled multi-body dynamics for all geometry movements.

Image Gallery

Video Gallery

Discrete Element Modelling of Iron Ore Fines to Include the Effects of Moisture and Fines

Mon, 01 Jul 2013 00:00:00 +0000

Click the Cite button above to get publication metadata for your reference management software in .bib format.

Granular Mechanics and Industrial Infrastructure Research Group

Sun, 27 Sep 2009 00:00:00 +0000

Introduction

The Granular Mechanics and Industrial Infrastructure research group is part of the School of Engineering at the University of Edinburgh. This group contains more than 20 members and is led by Prof. Jin Ooi. Members use a combination of experimental testing, analytical methods and simulations (with appropriate verification and validation) to investigate the behaviour of a diverse range of granular materials.

The group is characterised by a multi-scale philosophy: the application of multiple complementary approaches to a given system to enable exploration of a broad range of scales from the single particle to the bulk material. The group also actively research multi-phase systems in which one or more fluids must be considered to fully understand the system’s behaviour.

Research emanating from this long-established group has had a considerable impact. Within an academic context, group members disseminate their research results in peer-reviewed journal publications and at major international conferences, and the group collaborate with many leading academic research groups around the world. Outside of academia, many well-known companies have benefited from the group’s expertise through consultancy and/or collaborative research projects, e.g., AstraZeneca, Johnson Matthey, P&G and Pfizer.

In recent years, the group have been highly successful in winning research funding from EPSRC, the European Union, the Royal Society and IFPRI, in addition to industrial funding. The group at Edinburgh currently coordinate the EU FP7 Marie Curie Initial Training Network TUSAIL (2021–Pres.), which follows on from a previous highly successful FP7 projects T-MAPPP (2014–2018) and PARDEM (2009–2013).

Research Activities

The research activities of the group may be divided into the following broad themes

Featured Research

Furthermore, the group have a strong track record of research commercialisation. Two software companies, DEM Solutions (now Altair EDEM) and Particle Analytics, have been formed as spin-out companies from this group, and in 2016 Freeman Technology launched the Uniaxial Powder Tester based on the Edinburgh Powder Tester with the technology licensed from the University of Edinburgh.

Edinburgh Powder Tester