Depends on the DEM model you are running. Most models use artificial modulii and restitution values to allow a quicker time to the solution. It does not match with reality. There trick is to make the flow appear to meet the eye and not the mind.

Ask what are the true values and then compare with the DEM values. Publishing such here is questionable, since it gives others insight to what should be and realigns their verbal skills, and maybe in the long haul a correction to their code.

Better you should set up a standard set of tests by which all DEM codes must calibrate to. In this way, the running time is then considered a liability = getting it right causes long execution time. No spheres that provide low internal shear resistance, or sphere clusters, no fake bond breaking, no low modulus contacts with popcorn reactions at corners and edges, et al. ■

Hi Larry,

I was not going to comment but could not resist. Are you not confirming what I have been saying all along re most DEM programmes? If you have to do this for something as simple as coal flow which in most cases you should not have to model anyway, then what about the more complex materials where mositure,cohesiveness etc come into play? I know you are doing a great deal of research in this area and I commend this however there are a lot of transfer chute designers suffering under the illusions that the DEM programme they purchased generates real flow models, not pretty pictures.

Cheers

Col Benjamin

Gulf Conveyor Systems P/L

www.conveyorsystemstechnology.com ■

Dear Colin:

I keep reading about your dissing of DEM modeling. I think you are going out on a thin limb that will not support your mass. You published an article in Australian Bulk Handling Review March/April 2011 pages 44-45. You are quoted with many off-handed and misleading comments. I offer a number of cases in point where you published much about your flawed insight into DEM.

You quote: "we were criticized for not being aware of recent advancements in DEM. This was not true." You show a photo and subtext: "Lump-fines separation that DEM software is not able to model." DEM can clearly model the separation in the photo. We did so on the Yandi 5 km iron ore overland where big and small particle flow in chutes illustrates the separation with large lumps rising to the top. Many DEM published projects have shown, in many containment systems, the buoyancy of large particles in a bed of fines subjected to vibration or agitation. Do your research. Any DEM worth its salt will make the illustration clear and show the reason why. I now doubt you know the reason why such separation does occur.

You say: "DEM cannot identify mal-belt tracking due to (I assume) off-center loading at the transfer". Again not true. We and others advertise our ability to show belt side thrust that may occur from granular side force reactions, causing belt mistracking, discharging from the chute. We have made this argument in print and can easily demonstrate the phenomenon from chute modeling including the side thrust force magnitude = degree of mistracking.

You say: "There is no DEM software that can model varying amounts of water, large size distribution, clays, fine material, ....." Again not true. My guess is your exposure to good DEM code is highly limited. Maybe on purpose, since you are not convinced of its efficacy. Have you tried to talk to Dr. Paul Cleary, Prof. A.B. Yu,... lately? I doubt it.

You go to the less capable and less knowledgeable DEM marketers and make your less informed claims. All software behaves with the minimums of garbage in = garbage out. Can you predict the response of the material flow, only knowing the particle size distribution including micron size, surface water content, particle shape?

We get the impression that this is a marketing exercise to promote your book. However, the bias cannot be hidden from the overly negative and misinformed insight.

You claim the software is too expensive. The world-at-large seems to believe you are in a smaller and smaller minority. Maybe you have not tried a well formulated DEM code.

Stick with what you know. ■

what's coal???

-3mm wet Sticky fines with clay???

-25mm +3mm dry no fines???

Know the product, DEM or no DEM !!!

Applies to ANY material handling problem ■

Colin:

I admit to being harsh in commenting on your understanding of DEM. I, and others do not attack your understanding of chute physics and do acknowledge your ability to develop good practices, book included.

Likewise, I hope you take the time to learn what capabilities do exist in DEM. Then, maybe you will have better insight, leading to less critical commentary on advances in DEM technology by asking appropriate questions. If you wish to know more ask more. I am perplexed by your acknowledging there may be advanced codes that do good work, yet you select poor codes that do poor work and then claim "SEE, it does not work".

Any proper DEM code should have a means to calibrate the material flow mechanics, before conducting design efforts. I believe you labeled the process "reverse engineering". It seems you do exactly the same thing by model building, testing, modifying. I equate test, observe, change, and test, observe, change interations to be equivalent.

I am willing to take your test, if you take my test. Calibration is the key and it is not expensive. ■

In order to set the restitution value (R), you need to do a laboratory test. Vertical drop test is common. Measure height of rebound. You need sphere shaped rock specimens. By example, drop a spherical rock vertically, from 1000 mm height. Then, measure the height of the rebound. Assume the rebound is 0.090 mm. Knowing particle height at impact and height on the rebound (R) square is proportional to kinetic energy transfer (KE=.5mV^2). You obtain the restitution R equal to the ratio of velocities ( Vout/Vin) or square root of height ratios [R=(0.09/1.0)^.5 = 0.3]. Most experiments require also addressing spin energy effects and therefore you should not use irregular shaped rock. Spheres are less likely to induce spin on impact.

Shear Modulus? Shear modulus is a product of the internal resistance to rock flow. It is not a input value to most DEM models. Stiffness modulus like elastic modulus is the likely value you will require. This is the value of the applied force against the DEM particle, then measure the degree of deformation overlap in mm. The value depends on the DEM code used. Some try to speed up the code by inducing low stiffness coefficients. We see code with 200 N/mm to 10^6 N/mm. Most rock is very stiff in the later magnitude. ROCKY, our DEM code, uses the much stiffer value.

What code are you using? ■

Hi Larry,

Thanks for your reply. Once again I have to emphasise that I am fully supportive of the work you and others are doing developing better DEA models. This work will lead to a much better understanding of material flow dynamics just as the work we have done using dynamic scale modelling has also done. As I have said in previous correspondence however I am not prepared to interact on developing better DEA models unless I have access to the mathmatical logic otherwise there is no mutual development of our knowledge base.

My arguments in respect to DEM is all about its misuse as a design tool and this misuse stems from its promotion by you and others without any qualification as to the limitations of the technology. The first point is the vast majority of DEM programmes out in the market place and unfortunately the one's most engineers purchase as they are affordable are dangerous as they have basic flaws in their underlying logic. In the hands of inexperienced designers the end result is another poorly performing transfer chute (or worse failure). The more advanced programmes such as yours, EDEM and those developed by various Universities that require accurate calibration are expensive and require a great deal of training and know how to manipulate. The cost of doing a design evaluation using such programmes therefore is not much different to doing a dynamic scale model evaluation.

When it comes to conveyor system design there are far more underlying variables to consider in the design of a transfer chute than in any other part of the process plant yet so little attention has been paid to understanding the underlying design principles. In virtually all the operations I have visited in the last 7 years the major constraint to production, the major maintenance problem have all been transfer chutes. It was this that led to the writing of "The Transfer Chute Design Manual".

In the last couple of months I have once again been involved in looking at a number of serious transfer chute design failures. In both instances they had been modelled using commercially available DEM programmes. It led to me writing an article that I guess summarises my concerns. I have re-published a part of the article below;

"The basic problem is many companies are using poorly trained engineers to design transfers, engineers who do not have the experience or understanding of materials handling to take on what in many cases is a complex process. Such engineers see and read articles on DEM and believe that by acquiring such software they can ‘test’ their designs and therefore take on such tasks. Our observation of this issue was a prime motivation for writing our design manual. Another reason is that many of the DEM Software was ‘tested’ on sized, washed coal. This material has very predictable flow properties so any such evaluation gave a false impression of the ‘reliability’ of the software. Looking more specifically at the design of the available DEM Software the issues are:

When we are managing very complex ores such as those containing varying amounts of water, large size distributions, clays or micro fine materials (sub 200 micron) the flow dynamics are extremely complex. If we are to use DEM to evaluate a design in such circumstances we need a very advanced programme that allows us to accurately model the materials being handled. To the author’s best knowledge there is no DEM software capable of this, yet DEM advocates continually claim their programs will assist in designing the equipment in question.

In an attempt to address the above, many researchers now use data gathered from other transfer applications handling the same or very similar material and, through a process of reverse engineering, duplicate the outcome on a known transfer and then apply it to the new transfer. However the limiting factor is once again the inability for DEM to model complex materials that are moist, with large size distributions containing micro fine materials.

In some cases we read that Universities are attempting to undertake the aforementioned reverse engineering using dynamic scale modelling. The problem with this approach is that the transfer designer must have access to mathematical logic that was used to develop the DEM Software in order to ‘tune’ the software to the material if the outcome is to be accurate. Designers of such advanced software are very reluctant to share such material which inevitably means the software is extremely expensive or one must rely on a third party to run the DEM simulation.

Due to the mathematical rigour and computing resources needed to model complex material characteristics, many commercially available DEM programs treat the material flow dynamics in a very facile manner (simple spherical particles). These provide relatively quick solutions however the easier and quicker the DEM programme is to use the more likely the output is a ‘pretty picture’ with no relevance to reality. In analytical transfer chute design the first fundamental step is calculating the material trajectory. In the two examples highlighted above, it was found the DEM trajectories did not match those produced by analytical modelling, and consequently all downstream chute design aspects were flawed. Compilation times for DEM simulations comprising non-spherical particles (e.g. super-quadratics) can be greater than two months, which is of little benefit if a chute design solution is required for client immediately."

Once again let me emphasise I am not against the development of better DEM programmes. I am against the way DEM has been promoted without qualification and the result is literally thousands of poorly performing transfer chutes around the globe because the designers relied on DEM instead of their good basic engineering design skills.

Cheers

Colin Benjamin

Gulf Conveyor Systems Pty Ltd

colin.benjamin@gcsm.com.au

www.conveyorsystemstechnology.com ■

Originally Posted by designer

what's coal???

-3mm wet Sticky fines with clay???

-25mm +3mm dry no fines???

Know the product, DEM or no DEM !!

Applies to ANY material handling problem

Science might take the fun out of the job. ■

Originally Posted by louispanjang

Science might take the fun out of the job.

I'm all for latest techniques, but with computer programmes

GARBAGE IN = GARBAGE OUT

Last time I discussed coal at a power station the general engineering agreement was that these days coal was "generally black, has a calorific value, and is the cheapest thing buyers can get hold of!!". ■

Dear designer, Louispanjang,

An engineering job done, based on science is much more rewarding than a job done on an educated guess that cannot be explained.

And what is science then meant for?

Is an engineer without a computer program better than when the same engineer uses a computer program with garbage in = garbage out?

I sense the generation gap between old (grumpy?) engineers and young engineers (real engineers, not managers).

There has been a time that the old engineers were young engineers, probably facing the same generation gap.

In my opinion, the computer and the scientific software, is a great tool for understanding and designing technical installations.

And there are many young engineers out there, who are smart, well educated and highly motivated.

Give them a chance.

Cheers

Teus ■

If you read too many posts its -

give me the formula for this

give me the factor for that

what's the computer solution for this

I use computers all the time to solve problems but if you don't know your material properties how can you get the right answer

Bulk materials aren't like steel, EN8 has these properties, EN32 has those properties.

Remember the question originally posed "the properties for coal".

How vague is that■

Whem properties are unknown there is little point in claiming that computerisation will reveal all. I use parametric 3D modelling daily: interspersed with sessions in ANSYS or StaadPro and Mathcad. Accordingly I am reasonably aware of the limitations of modern engineering techniques.

I still have my old slide rule: just in case man ever decides to go to the moon again. ■

I can imagine that it must be disconcerting to have purchased a DEM package and have no idea how to set the basic input parameters such as coefficient of restitution. I assume that you are referring to the coefficient of restitution between the particles themselves, rather that the CoR between coal and steel, or ceramic, or rubber or accumulated coal buildup.

I presume that Mr. Nordell’s comment about the drop test was a little tongue in cheek. If you had to make spheres of every material that you wanted to consider and conduct drop tests then you would be up for quite an experimental program. Particularly for highly variable ores.

The first comment I would make is that coefficient of restitution is not some fundamental physical constant in the same way as, for example, density. Even for a simple experiment with a tennis ball, and a concrete floor, the CoR varies with the drop height. The fact that coefficient of restitution varies with velocity has been known for at least 180 years. With coal, the situation is even more complicated because coal is generally anisotropic, due to the way it is formed. That is why we often get acicular particle shapes in coal. Furthermore, CoR is strongly influenced by porosity, and in particular, water filled porosity. There are separation processes based on just this fact. I developed one myself years ago that separated bauxite on this basis. The patent application is available online and I have mentioned it here before.

So, what you need is an ‘effective’ CoR that will suit your purposes.

Given the situation, a search of relevant recent literature would seem a reasonable approach. I have had a look at some information that is available online, and there are two papers that I would like to draw to your attention.

The first is a paper co-authored by Paul Cleary, mentioned by Mr. Nordell. I met Paul 20 years ago and he has been doing this stuff for a long time. It is titled “Summary of ACARP Project on Cross Belt Cutters” and is published in the Journal of the South African Institute of Mining and Metallurgy in June 2010. In that paper, the DEM simulation of a cross belt coal cutter sampler is described. The coefficient of restitution used for the main calculation was 0.25. I could not see any elaboration of the origin of that number. Perhaps it was selected to suit the simulation and prevent wild scattering of the particles or to counteract the fact that the size distribution was truncated at 12 mm. In any case, there is one possibility. Elsewhere in the paper, results for a CoR of 0.5 are reported, so that might be a further possibility.

The next paper that I would like to draw your attention to reports work done at the University of Wollongong and is titled “On improving the Calibration and Validation of Computer Simulations for Bulk Materials Handling Systems” and it was published in Australian Bulk Handling Review, Sept/Oct 2009. Perusal of that paper will reveal that actual tests were conducted to measure the CoR of a sample of washed, dry coal. They used high speed cinematography. The result was 0.55. This suited one of the tests conducted in the paper, but in another case, a CoR of 0.78 was found to provide a better fit.

So, there we have it, values of CoR in the range 0.25 to 0.78 used in DEM simulations of coal in recent literature.

To illustrate the breadth of that range, it covers values that range from a squash ball dropped on the floor to a golf ball hit with a steel club. The choice is yours, but I think if I was in your shoes I would opt for something around 0.5 for a start to minimise errors. And looking at those papers, perhaps it is not as important a parameter as one might think. ■

Originally Posted by louispanjang

I still have my old slide rule:

Mine is a Faber Castell 2/82, plus a clip on magnifier for 'greater accuracy' ■

I had an ARISTO slide rule.

Type: Slip - Stick

Teus ■

Admitting K&E, Aristo, Castell, I guess will carry the label: “ The Old Grumpy Man’s Guide To The Universe." I think most have never held the marvel.

Regarding Coefficient of Restitution (CoR): Many tests have been performed on hard rock where the rock sphere machining is possible.

Some coal is in a world apart. It begins to crumble on impact with sub-bit. Anthracite can be sphered.

Does coal have a higher rebound? Recall it also depends on the particle size for most rock groups except the coal Indonesian stuff. As the rock size is comminuted, its apparent strength increases as the internal flaws are reduced = higher rebound height.

First met Paul in 1995 where we began to collectively collaborate on DEM advancements. We each had our own methods and wished to see if combining the technologies and companies would yield a product superior to either. We now have ROCKY, first named by Dr. Noel Barton, Paul’s boss. ■

CoR = 0.78 must be more efficient than a ping pong ball. It can rebound to 78-80%, claimed by experiment, of its initial drop height? Pure latex rubber, made into a ball has trouble achieving this value. I also tried silly puddy. I wonder what is missing?

Most competent rocks fits in the range of CoR = 0.30 from testing. Of course there is the Pilbara 68% grade iron ore with a pot metal bounce. ■

Originally Posted by nordell

CoR = 0.78 must be more efficient than a ping pong ball. It can rebound to 78-80%, claimed by experiment, of its initial drop height? Pure latex rubber, made into a ball has trouble achieving this value. I also tried silly puddy. I wonder what is missing?

Most competent rocks fits in the range of CoR = 0.30 from testing. Of course there is the Pilbara 68% grade iron ore with a pot metal bounce.

No, if you read the paper, which is available online, you will see that it is simply a value determined by best fit to an experiment involving many particles. Fiddling with parameters to match an experimental observation. From the look of the graphs, there is not a lot in it.

Yes, iron ore can be bouncy! Not just from other particles or steel either. I have filmed large pieces breaking free from the pack on a banana screen feed and leaping three decks down the surface. ■

Originally Posted by nordell

Admitting K&E, Aristo, Castell, I guess will carry the label: “ The Old Grumpy Man’s Guide To The Universe." I think most have never held the marvel.

Don't run away with the Moderator Moniker. After all your bouncing around just tell the man what CoR and Shear Modulus he should be using. If you can't define that then what is the point? ■

No bounce is this step. Most readers will have little knowledge of how to applied the details. It just gives our competition sensitive information. Buy ROCKY and you will know more.

Unfortunately, to be accurate there is no one value. Some typical restitution values (R) from testing of materials machined to 25 mm spheres that are dependent on impact velocity onto a thick steel plate:

1. Glass ............... 0.9

2. Steel cast iron .. 0.70 - 0.80

3. Brass .............. 0.45 - 0.60

4. Limestone ........ 0.55 to 0.60

5. Coal ............... 0.25 t0 0.50 (Anthracite through some Bituminous; (Lignite ??))

6. Clay ............... 0.12 to 0.17

Since DEM codes come in all flavors, what is right? Know your technology. There are other values for spheres of like materials, et. al. ■

I do not know what a Shear Modulus is in rock parlence as it applies to DEM.

As I said earlier normally, the DEM value depends on the rock shape, size distribution, and related factors. Please read a snip from the web:

http://ascelibrary.org/gto/resource/...sAuthorized=no ■

I have attached a ROCKY brochure with notes on the next generation of DEM that can be applied to chute designs. Faster, more accurate, large size ratio, larger non-round sizes and shapes with gas. Some basic calibration capabilies are included in the User's Manual.

This will release in 2 weeks.

Attachments

rocky brochure_051011 (PDF)

■

All these posts emphasise once again that understanding the underlying basic design principles is essential if you are going to design a transfer chute. We can argue all we like about what are the correct calibrations to use but before we even get to this, what is the basis of calculating the trajectory of these particles off the head pulley? Get that wrong and your damned before you start. The broad range of particles, shapes and sizes in ROM or primary crushed ore is myriad, how do you simulate this? I know in your new programme "Rocky" you can input particles sizes and shapes but the reality is we need basic design fundamentals to come up with transfer shapes that will allow for these broad range of materials. Whether a DEM programme can successfully simulate such material flow is really academic unless the designer knows what to do when confronted with such issues.

Once you have a design based on good underlying design principles you can elect to evaluate it in whatever manner you like. If you use DEM then if it is a poor programme it won't enlighten you at all but it will give you a pretty picture. If it is a good programme it could assist by identifying high pressure zones but here is the rub, how does a designer decide what is a good or a bad programme without knowledge of the underlying mathmatical logic, eg how the trajectories were calculated. Form our viewpoint we can only do this by trialling the software against known transfer problems that we have dealt with in the past using a broad range of design situations. Alternatively you let others purchase new software and wait for feedback as to its usefulness.

Cheers

Colin Benjamin

Gulf Conveyor Systems Pty Ltd

www.conveyorsystemstechnology.com ■

Dear Colin, Graham,

PLease open attached to see fines bouyance causing large lumps to rise to surface along belt.

There are three images. First two illustrate flow in rockbox with non-round DEM. First image shows the fines in bottom corner of rockbox and outer idler wing roll with only fines and with no large lumps on bottom of belt surface.

Second image shows rockbox geometry profile with

belt trough at right side of image.

Third image shows similar fines segregation on idler wing section on belt surface. ■

In case you missed it and wish to see it, click on last posting in the next to last line inserted "Rockbox segregation.doc" . It was also called "Attachment 27871". ■

Hi Larry,

This is buta small part of the story. Fine you simulate lump/fines separation but then you need to apply differential trajectories based on air interaction with the fines, moisture interaction etc etc. Once again until I can run my own simulation of your programme against known outcomes that we have done I remain extremely cautious as to the claims you continue to make. Your brochure for instance claims your programme does lots of design work, let me ask what design parameters can I insert into your programme to create a design? The answer is I cannot. I have to insert a proposed design. How do I go about designing the transfer or is all by guess and good luck? The issue is simply you have to know how to design a transfer before you subject it to any design analysis. You promote DEM as the way to do this, I question the accuracy of the DEM modelling process. If I use a scale model at least I can interact in a way that allows me to test the design, with DEM it is all about the mathematics used and whether it is robust enough and as anyone buying a DEM programme has no way of accessing the mathematics we have no idea whether the programme is good, bad or a disaster unless we can evaluate it against known situations that represent true tests of the technology (not sized washed coal).

As a final comment I refer you to an earlier post of yours where you commented on calibrating a DEM programme for a coal transfer and suggested a range of resistivity factors, none of which had any empirical merit, all were designed to "make " the DEM programme work

Cheers

Colin Benjamin

Gulf Conveyor Systems Pty Ltd

www.conveyorsystemstechnology.com ■

Colin,

Few points:

1. Separation, then need apply differential trajectories? I see you do not comment on your earlier observation that DEM cannot show segregation. Then we show it. Any change of heart?

2. Air interaction? I showed you an image of gas dynamics interacting with flow stream. The gas and fines would need to be calibrated. Since we cannot model micro size material we do like you do and scale the details. Large particles with lighter mass to achieve simular turbulent response on particle groups and not on individual particles.

3. Maybe my memory is failing. Where did we claim the code designs transfer chutes? The designer must know something about the business.

4. What does "resistivity" have to do with transfer chutes? "Restitution" maybe? If you mean "restitution" and say this measurement has not merit? I suggest you read up on the definition. It is a real life, commonly measured parameter, as I noted in a prior posting on this thread. Do you need references? If you mean something else, please quote the point. My memory is good for just a few thoughts at a time.

5. Even panel beaters can be chute designers if they want to know what are possibilities with granular flow.

6. We offer a few simple calibration devices that anyone can build and take measurements on for many granular materials and moisture contents. I will not work for highly cohesive/adhesive values where the material will not carry its own weight on granular sizes the exceed 20 mm. There are ways to tune these materials as well, but, it takes more knowledge than dropping off the bus. In time these calibration methods may also be academic.

There is no substitute for knowledge. Many wish to be able to know and do more. This gives them the tools to explore today. Read the 120 page manual.

I will let you use ROCKY for 30 days. You let me read your book for 30 days. I promise not to copy the book's contents. I would attempt to model some of your chutes, if they are sufficiently dimensioned. Certainly would study your knowledge of DEM. I return the book or buy it, if it is worthy. I will even give credit to its insites and advancements, if you will do the same, if ROCKY is more than a toy. I will not diss the book or your effort. You do not have to be as kind, just ask questions before going public with comments.

ROCKY uses multicore capability. You should have 4,6,8, or more cores on a modern Intel machine with at minimum 8 GB memory. The program makes big files to enable mining the data such as wear surfaces and keeping a history of all interactions to enable replay.

What say you? ■

Hi Larry,

I had already replied to your email invitation and would like to confirm that I am willing to run your new software through some known transfer situations. In the article published in ABSH that I think you refer, I also stated that I very much support the work you and others are doing on DEM. If your programme proves to be effective even in a limited range of applications, it will be very good in deed as most commercial DEM software is not only very limited, in the hands of inexperienced designers, very dangerous.

As far as the book, if you want a copy we will sort this out via email but the book is not about DEM, it is about design. It has now been sent to 19 countries and the feed back has been very good. Looking forward to agrreing arrangements for the evaluation

Cheers

Col Benjamin

Gulf Conveyor Systems Pty Ltd

colin.benjamin@gcsm.com.au

www.conveyorsystemstechnology.com ■

Hi Larry,

I was interested to see your Word doc with the simulation pictures. Thank you for sharing that.

The chute in the first two pictures interested me. I have not seen one like it. Is it one of your own design? Is it included as an example in your software evaluation package? If so, can you vary the flowrate? I would love to have the chance to play with that. I could easily make up a physical model of that, given the normal parameters.

As I say, I have never seen a chute like that, but I was surprised to see that there was no sign of particle size segregation across the width of the outgoing belt. But you never know what you are going to find. Surprises abound.

You know, every chute that I model should, in theory, be able to be duplicated by your DEM program. But I suspect that there will be divergence. In fact, that is a known. We are both dealing in models of reality, not enhanced reality. We need to keep that in mind. Paul said it elegantly in one of his papers.

I am open to collaboration with yourself and Colin. ■

Colin, Peter,

I will send a ROCKY copy for review this week. Let us see how fair minded you are. Do you have AutoCAD Inventor? If not, we will have a more general format within the next week.

Regarding the images in the thread above: These images are of a chute designed by another. We were asked to critique and correct, which we have done. The big losses with the intial design are belt wear life reduction and big power draw penalty. ROCKY can alter the flow rate and mix of particle sizes, shapes, material properties, cohesion and adhesion, over a wide range of conditions within each particle set. You will be amazed at the cost.

Thanks for the softer tone. ■

Peter,

Good eyes. Yes, it does not show lump segregation. I missed this point as a reality check. We do often see the larger particle separation across the width with rockbox right angle transfers. I must look at this result a little further. ■

Hi Larry,

We have both Autocad and Inventor. We will just need to ensure there is enough RAM on the computors on which these are loaded.

Cheers

Col Benjamin

Gulf Conveyor Systems Pty Ltd

colin.benjamin@gcsm.com.au

www.conveyorsystemstechnology.com ■

Please spec your machine details:

1. 32 or 64 bit architecture

2. number of cpu cores on one machine - test up to 64 cores with hyperthreading

3. memory -2 GB for 50,000 particles, 4 GB for 100,000 up to 24 GB for 1 million, 48 GB for 2 million using non-round faceted particles

Since this is a DEMO, it is good for 30 days. We will send when you give the specs.

Note: There is special password protection. Do not let it roam, we will know. Will you be able to allow Mr. Donnecker usage? The manual is given in confidence. There are additions yet to be released. ■

Hi Larry,

I will load it onto one computor of sufficient power and then get together with Peter. We are in different States but I visit WA very frerquently. We will do the evaluation together. First step will be understanding how it works and what we have to do to make sure it runs well. We will then run it through a series of different scenarios. I am hoping all goes reasonably well as I am more than prepared to include a good DEM programme into our design arsenal even if there remains some limitations. My aim is to work through scenarios that define where we can be confident and where we have concerns. It should also be very good feed back for you.

Cheers

Col Benjamin

Gulf Conveyor systems Pty Ltd

www.conveyorsystemstechnology.com ■

Colin,

What company name should be security coded: Gulf Conveyor Systems Pty LtD. or Conveyor Systems Technology? ■

Originally Posted by nordell

Peter,

Good eyes. Yes, it does not show lump segregation. I missed this point as a reality check. We do often see the larger particle separation across the width with rockbox right angle transfers. I must look at this result a little further.

Yes, a common issue with right angled transfers and in many other transfers with significant size distributions in the feed.

I have emailed you separately about software issues. ■

Colin; Peter,

You have now had ROCKY for more than 2 weeks. Please give an offering of your thoughts. What type of chute models have you investigate. Give some details if possible on computer and models. ■