I have been involved in the designing of transfers for over 30 years and while I understand you are using the Helix software to do a DEM simulation of the design you have developed I doubt that the issues you are raising will make much difference to the evaluation outcome. There are far more fundamental issues as to whether the Helix software will accurately simulate the flow through a chute. I would suggest that you read up on the many references available on DEM and transfer chute design.

Cheers

Colin Benjamin

Gulf Conveyor Systems Pty Ltd

www.conveyorsystemstechnology.com ■

Dear Colin,

To be fair, you could elaborate on the faults of Helix, since you are concerned about the misapplication of Helix to solve his flow problem. What are the fundamentals he need to address to obtain accuracy of the flow regime in his chute?

I have never seen the Helix DEM output. Maybe the thread starter can give an illustration of the problem in specifics.

I can say most codes would have a problem solving a 9300 t/h flow stream resolved down to 20 mm with a typical percent passing in 5 tier grouping (i.e. 100 mm, 80 mm, 60 mm 40 mm and 20 mm). How long does the simulation take in hours, for how many simulation seconds, what particle size distribution did the thread starter apply? Are all particle spheres or composites of spheres or other shapes? ■

Hi Larry,

The Helix Software is a commercially available product among many that are freely available and it is not appropriate on this Forum to single it out solely for critique. There are papers available that look at DEM Software that anyone purchasing such a programme should read before making the investment. The issues raised by the Post in our experience where we have seen this software used were not very critical, there were other issues of more import.

I am however prepared to make some general comments that are applicable to many DEM software programmes we have looked at and they are;

> The calculation of the material trajectory is in some cases not accurate.

> Some do not allow for the material size separation that occurs on a conveyor belt when projecting flow

> There are issue with both adhesive and cohesive material flow that is extremely complex to calculate and therefore simulate in DEM.

> There are issues with how DEM manages moisture content. This issue is made more complex by the surface tension affects that water and water laden material has with the conveyor belt and therefore the material trajectory off the head pulley.

> There are issues with ores that contain micro fines

Cheers

Colin Benjamin

Gulf Conveyor Systems P/L

www.conveyorsystemstechnology.com ■

Colin,

To quote you:

"There are far more fundamental issues as to whether the Helix software will accurately simulate the flow through a chute."

How do we interpret the above, if not a critique? I believe the comment is meant to create doubt, if you mean to or not.

You seem to offer comments without substance:

1. DEM often does not portray trajectories in an accurate manner - what proof? CDI is willing to take your test, which you are free to publish and criticize. We believe the code does an excellent job at trajectories and chute flows. In particular, we treat rock as non-round particles close to their known morphology and do not apply multiple spheres to do so.

2. I quote: "Allow for the material size separation that occurs on a conveyor belt when projecting flow". Size separation from what? I believe we understand the comment, but again it would help to give some insight, since I believe you are not discussing the percolation of smaller particles to the bottom of the charge, which is well known? I believe you are not talking about windage of fine particles, which DEM can simulate for understanding this effect in a solid-gas model with special techniques. Can you elaborate?

3. Many of us that practice particle physics do know a bit about adhesive and cohesive material and the science behind the attractive forces at the DEM and molecular level, the various rheology mechanisms, and their association within DEM. Not all DEM codes have introduced a cohesion model. Maybe some get it wrong. The forum is meant to help. Let us make that effort. Charaterization of an accurate cohesion model is more dependent on the critical moisture content that can be stress tested in a lab, which parameters can then be modelled for the specific flow behavior in better DEM codes.

4. Again, you lecture on whether developers of DEM software are equipped to model "liquid bridge" effects and the models that are given in point 3 above. Please demostrate with examples about "missing accuracy", without naming names, and show how your alternative is accurate. This will show your command of the science and go far in creating believers of your criticisms.

In conclusion, please teach us what you know we are missing with some illustrations, field measurements, or other instructions that will convince the forum readers. ■

Hi Larry,

Sorry for the delay in replying but I have been out of the office.

The summarised issues I raised are based on a great deal of work and research, In particular we have looked at a very large nmber of actual transfers over the last 6-7 years. In all cases we have been asked to look at them because there were performance issues ranging from rapid wear of liners, blockages, lack of throughput, belt tracking issues, belt damage etc. We have in all cases to use digital cameras and other means to assess the material trajectories. We have in most instances had access to the transfer design drawings which in most instances included the trajectories used to develop the designs. In some instances, because we were asked to forensically examine the issues, we gained access to the design methodologies used by those that had designed the transfer. We have been able to;

> Compare actual trajectories to what the designers have used and what we would have calculated.

> In some instances we have able to dynamically scale model and through this not only replicate the problems being encountered but also check whether our solutions would solve or mitigate these problems.

Many of these designs we looked at used as a means of checking and refining the design a DEM programme. It is based on this we make those general observations. I am aware that DEM software is getting more and more sophisticated and through this improving its overall accuracy. But with this sophistication comes a significant increase in the cost of the software and an increase in the time it takes to run a DEM evaluation.

We have exchanged emails privately as to the basis on which we would share this research with you. Meanwhile let me finish this note by saying back 10 years ago many of the transfers we saw going into the West Australian iron ore industry were modelled using DEM. Today most are evaluated using dynamic scale modelling or rely on a standard design approach.

Cheers

Col Benjamin

Gulf Conveyor systems P/L

www.conveyorsystemstechnology.com ■

Colin,

1. Our DEM-comminution code was sold to a major equipment supplier who took 2 years of due-diligence investigation. Extensive flow and breakage with gas and fluids were modelled. Rio Tinto ATD division, in Perth, requested we perform analysis to compare with their measurements. We met all their tests with approaved accuracy, before the equipment was sold or given to Mr. Donnecker.

We have many such test works for major mining company where flow and breakage modelling is performed. Thus, I see we have also met some major milestones of validation, including wear behavior, trajectory mapping, differences between particle shapes and size distribution.

2. What is an acceptable time? Tell us how long it takes to do a typical transfer chute flow model and what the input features are: a) tonnage, b) belt speed, c) size distribution and shape you model, d) material properties, and how long your model takes to extract critical date?

3. Do you model belt wear?

4. Does your book illustrate these claims and performances, and compare with know facts?

5. "DEM is becoming more sophistigated, with improved accuracy,but comes at significant cost of software and increased time to execute." What is an acceptable cost in dollars and time? ■

Larry,

Before you make reference to me by name, I would appreciate it if;

a) you get the spelling correct.

b) you get your facts correct. I have had no part in any transaction involving any work that you may or may not have done for ATD.

I personally did detailed comparison of physical modelling with field results when it was first developed. This involved removing the liner plates from chutes and determining the flow directions and comparing the effects of throughput, such as chute choking point, level of fill and so forth. This work all supported the efficacy of physical modelling. It has continued to demonstrate its effectiveness in the intervening 15 years.

A typical physical model has a turn around time of 3 weeks. Yes, all of the inputs you mention are of course used.

One thing I can tell you is that, running a team of typically 6 materials engineers with the best equipped wear laboratory in the country, for a period of 6 years, we were not able to correlate the wear of either belts or liners in chutes with wear tests. Provide a ranking, yes. Provide quantitative data, no.

So I am suitably impressed that you are able to wave your magic mathematical wand and come up with the answers to issues that we could not even measure effectively in the field over 40 man years of concentrated effort. Anyone familiar with wear in chutes will know that wear patterns are chaotic in nature and are strongly influenced by upstream and downstream effects, moisture, variations in material hardness and many other factors. I wonder if you can magically predict the wear mechanism. Micro-ploughing etc. etc?

No, I do not claim to be able to quantify wear. Indicate wear areas, sure. Demonstrate the contact area for wear, yes. Comment on suitable materials, yes. But quantify wear for any random combination of ore and liner, ridiculous!

Personally, I take objection to your constant belligerant attitude on this forum. I am not alone there, that is for sure. In my experience every forum has its bully, even if he be a knowledgeable one.

I have answered this because you have made reference to my name and got your facts wrong. I am not interested in continuing some war of words or entering into some tedious comparison of physical modelling versus DEM. They are both techniques with strengths and weaknesses and limitations. To claim that DEM is infallible is nonsense, and you know it. ■

In respect to your posting Larry,

Our book, (Peter Donecker is a co-author) does;

> cover the history and process of physical scale modeling.

> Once again the process of physical scale modeling as with DEM is not about actually designing but checking and through this refining the design process.

> We do not attack DEM in the book, we point there are limitations. We have in fact gone further in this posting in answer to your queries, to elaborate our concerns than we have in the book

Cheers

Colin Benjamin

Gulf Conveyor Systems P/L

www.conveyorsystemstechnology.com ■

At the risk of being 'flamed' by all those taking part in this somewhat acrimonious discussion surely publish a few examples of the theoretical v the practical so those of us who have to operate without the finances to use DEM or modelling can see the variance between different methods.

If the differences amount to a couple of inches thats totally different to a couple of feet.

(I now retire to a safe distance) ■

Sorry Mr. Donecker,

If you have been following the many postings, I have difficulty spelling late at night.

To your point (b), get your facts correct. I did not imply CDI did any work with you. I did not meet you until after the chute modeling equipment was removed from the ATD premises. I made no reference to your involvement on any CDI projects or visa versa. I have no interest in hostility and let the rest of your comments speak for themselves.

CDI can do a reasonable chute flow simulation (200,000 typical to more than 2 million non-round granular particles) in a few days and create a meaningful chute design in less than one week with wear properties. Calibration is required. This typically comes from known working systems.

I can say we are able to calibrate and predict belt and chute liner wear patterns. We have published these accounts for belts, chute liners, large and small gyratory and cone crushers, and AG and SAG mills. The work was evaluated by Metso, for almost two years, and met their laboratory and many field measurements to their commercial satisfaction. Metso concluded with the purchase of all software, hardware and personnel.

My guess is you are not impressed, as suggested by the repeated "magical" comments. I offer comment to those who wish to know more about such technology. It is more advanced than many acknowledge, some from ignorance and some from fright.

I had hoped, as was the case many years ago, that professionals could respect the efforts each apply to the common good. Sadly, civility has disappeared and is replaced by cynicism. I do not wish to denigrate you or your associates abilities, resources, insight, or experience, and I do respect the scale modeling methods. Enough said.

Colin:

We seem to look at the same problem and see different solutions. This is OK. Can you accept that the world of DEM may have moved beyond what you know? If it has, as a good technician, you might wish to become more attuned to the present and its consequences. Criticism is good, so long as you are right, and you are instructing those of lesser understanding.

Lastly, I would purchase the CHUTE book, if its teachings assure better understanding and engineering practice.

I have repeatedly challenged you on the book's content to assist CDI in this purchase. We do spend large sums of money on technology. Why not take advantage of this fact. Teach us about your wisdom that is beyond our understanding. We pay for fair value.

We will be placing the DEM code ROCKY with gas on the market in the near future, as academics are presently doing due-diligence. The code will allow many self-tailored non-round shapes and has certain rheologies dealing with cohesion and adhesion. ROCKY is a code specifically tailored for conveyors, their chutes, bins, bunkers, and related container designs. ■

Larry,

Thanks for your reply. I have never disputed that DEM is not advancing. Unfortunately there is a very large number of commercially available DEM software programmes that are currently being used that are not that advanced yet designers are relying on them. Secondly of those advanced ones that I have looked or been able to read up on the price tag makes the cost of our book look like chicken feed. Also the advancement of DEM (as with dynamic scale model testing) does not alter the following;

> If you do not know the basic design principles to develop a suitable transfer design in the first place, DEM won't help you.

> If you do not know the the basic design principles and how they relate to material interaction you cannot choose the right type of transfer and once again DEM cannot help you.

> When doing complex transfers or transfers with variable and difficult feed then you have to do many many iterations to come up with the optimum solution. Unless you are aware of the source code (mathematical logic) of the DEM programme you are effectively flying blind and relying on the software developer. With dynamic scale modeling the results and evaluation are very transparent. This is why I find it far easier to work with Peter than some compex DEM programme at this time. It does not mean that I am not interested in DEM although I doubt I will devote time and energy myself to developing a software model when I currently have a fully viable alternative that is proving very cost effective. If I get the opportunity to work with a DEM developer on an open basis I may change my mind.

In the real world of day to day engineering we have developed a culture that the computor is God. I am seeing countless number of young engineers who have no idea of the first principles of materials handling design yet have access to software programmes and are given a licence to design.

I see reliability engineers and maintenance engineers similarly afflicted who cannot work through the basic design principles and from this understand the root cause of a maintenace or operating problem. This is not in the exotic major overlands but in the every day fixed plant. While I accept there is a need to explore the edges of computor capability and advance engineering unless we also accept that those that we want to influence and move forward with us must first do their apprenticeship and understand the basics so they build their knowledge base on firm foundations then as an industry we are in trouble. I have visited many many operations over the last few years. In no site have I seen management in control of maintenance, in fact exactly the opposite with the consequence for production and operating costs. In today's high commodity prices production is king so you only get attention when production is affected, This will eventually change as it always does. In these fixed plants that I talk about (coal mining excepted) transfers are the major source of these maintenance hassles. In some cases the lot of the reliability engineer is just to get the wear liners to last until the shut cycle which can be as short as 8 weeks. When you have difficulties getting liners to last just 8 weeks then the state of transfer chute design is in a very poor state.

Our book does go into some detailed empirical design that is just as relevant to the developer of a DEM software module as it is to a transfer chute designer but it's focus is on the basics and sharing a knowledge base from first principles. It took a lot of time and cost a lot to publish. I dare say much much more than what has been spent developing many of the DEM programmes that I have looked at. There are items in the book that may interest you but we were not targetting you as a potential purchaser. Where I challenge you is on this ongoing defence of DEM as some sort of magic bullet that once purchased will miraculously solve your transfer chute design issues. The reality for the vast majority of designers, maintainers and reliability engineers is that it will only confuse you and take you further away from the solution you need unless you use it with the full base knowledge of the design principles required to design a transfer.

Cheers

Colin Benjamin

Gulf Conveyors Systems P/L

colin.benjamin@gcsm.com.au

www.conveyorsystemstechnology.com ■

Perhaps it is time to put personal differences aside and address the issue in the OP. As I understand it, the question is how to set the various parameters needed to perform a DEM simulation.

I am certainly not an expert in DEM, but I do understand that any package requires the setting of parameters for particle interaction and other factors. I imagine that an experienced practitioner can make estimates of these. For the inexperienced practitioner, this is clearly a problem as the OP indicates.

As Mr. Nordell mentions, it is all a matter of calibration. In this regard, I draw attention to the following paper from the University of Leoben where such issues are discussed. Note in particular the use of a physical scale model to calibrate the DEM model. This raises the question of why it was necessary to run a DEM model at all, instead of just using the physical scale model.

http://www.mas.bg.ac.rs/istrazivanje...4FKessler.pdf

I find this an interesting study, because we have run a physical scale model of a chute of the same style of design as that illustrated in the paper, handling coal in a very similar application. Through that modelling exercise, we were able to draw the designer’s attention to issues of wear and throughput as well as some of the basics that Colin Benjamin mentions in our book.

In addition, the process of physically constructing the chute drew attention to shortcomings in the drawings, something which we frequently encounter. These arise no matter what form the drawings are received in, be it as PDF or TIFF files, or output from any of the CAD software programs.

During the course of the modelling exercise, through a process of discussion, the designer made appropriate modifications, which we tested and the chute is now in service, operating without problems at last enquiry. ■