Please post details of the TUNRA method so we are able to comment on it. ■

Feeder overpressures can be awkward to assess due to the many variations of bin geometry and variations in the properties of bulk materials, particularly according to whether the bin is of mass flow design and the feeder extracting from the total interface area, but I would make three more points.

- Starting loads can be much inexcess of running loads, so an allowance is necessary to deal with start-ups after long periods of standing.

- 'Normal' running conditions with mineral applications does not cover exceptional product condition that may be expected from time to time.

- finally, It is better to be over-powered than under-powered, so at least the method appears to be on the right side ■

I had hoped that by now DaveB would have posted details of the TUNRA calculation method :-(

I expect that there are a lot of us who use methods that we, and those around us, have developed over a period of time. If you make feeders you have to have some idea of what power to to install.

As Lyn has said, better to be over powered than under powered. Complaining about a motor being a bit bigger than absolutely necessary will be as nothing compared with the Engineering/Anglo Saxon terminology used with an underpowered machine that won't start when you switch it on!! ■

I can well understand concern about an over-power factor of four when dealing with a substantial drive. Starting torques can reach this level in many cases and, whereas squirrel cage motors can sustain a short period of excess torque that is useful for overcoming high levels of initial shear, if linked to an inverter the overload potential is reduced so some additional allowance must be made.

However, the bottom line is that feeder start-up torque can be very sensitive to fine details of design or material condition. A theoretical analysis is dependent on various assumptions and cannot compete with actual measured values. A design formula has to be conservative to deal with the general case, so there is an inherent dilemma in balancing commercial efficiency with the prospect of an inadequate design that could have significant repercussions. Academic research does not allow the luxury of full scale proving trials so, if the issue of of sufficient importance, it beholds industry to involve research organisations to refine the methodolgy. It is most unlikely that the mathematics or the measured properties of the bulk solids are in question, so queries must relate to the sensistivity of the values of the active pressures developing the shear strength and wall frictional resistance.

Looking at what information is available for assessing feeder loads, one is reminded of a comment in a first world war cartoon associated with 'old Bill', sheltering in a shell hole - 'If you know of a better hole, you can go to it'. The breakdown of factors and application of basic mechanics and bulk technology to the problem by Alan Roberts of TUNRA is the most comprehensive and detailed approach available. Refinement calls for more accurate measurements in large scale applications.

An interesting point is that down-pressures at the outlet of a hopper very sensitively depend on how much the opening size exceed the critical arching dimension. Obviously, if an arch forms there is no overpressure. The greater the pressure, the more difficult it is to shear, so a low extraction force with reliable flow indicates that the outlet size to the feeder is well balanced. If a material is used that flows easier it will tend to increase the feeder load so what is best for the feeder is worst for the hopper outlet, and visa versa. This illustrates the need to test bulk material properties in the full range of potential condition. ■

Surely it beholds those seeking assistance to maximise the information they provide. As it is, we have no data on the applications that have proved overpowered and, until the reference was given, no idea what TUNRA formula was. It's not for those who might give comment to have to rummage round on the net.

Reference is also made to

. Have they been all the same or have they varied in dimension and capacity. If so, a comparative analysis between them may establish a correction that can be applied to the TUNRA formula.

I have a formula I might use, but until I have an application with all details (including the TUNRA results) not much I can say.

PS,

Why not??

Clearly, since I am employed in the industry I have greater freedom to make comments ■

The basis for computing loads on feeders used by manufacturing companies is unlikely to be in the public domain because of the commercial value associated with its application. Users may benefit from sharing such information, but manufacturers cannot be expected to educate competitors or, as may be the case in some instances, expose their ignorance of a scientific method and rely on empirical data.

I have to say that the derivation of extraction forces from a hopper is very sensitively related to many factors of the equipment design and bulk material properties. A well-designed hopper can minimise these to the extent that even a factor of four or five can remain a reasonable power margin to deal with variations in operating conditions, but the geometry for achieving such low values is definitely not universal practice. No doubt the scale and consistency of this deviation raises eyebrows, but nothing like the emotions of being underpowered.

A way forward would be to do a public service and publish the geometry and measured property values of the four installations so that others may examine the findings and possibly contibute better to the discussion. ■

Exactly so! ■

Dave,

There are several methods that I've seen commonly used in Perth for designing belt feeders. One is the method that uses test data commonly prepared by TUNRA. The other assumes a certain height of material on the belt as discussed in the Apex catalogue and also called Bruff's method.

I have used and seen both methods used. Start up loads are significantly higher than the run loads. Assessing these loads from first principles is not easy and will depend among other things the way in which the plant is operated. It is well documented that the startup loads can be three or four times the run loads.

I do recall a situation some years ago where we were required to check the calculations of feeders under a stockpile that had completely locked up. Someone had evidently miscalculated the loads. That leaves everyone in a very difficult situation, especially the designers.

If you contact the university of Woolongong you should be able to get papers that have been published on this topic by people who use TUNRA data.

http://www.saimh.co.za/beltcon/beltcon2/paper27.html ■

Why not do as Lyn has said and

As a matter of interest how were the various formulae referred to developed? I presume by some guy sitting down with paper and pencil and "imagining" what was happening and the loads that would result? ■

TUNRA has revised the method for the peak load, it is now the average of the old peak and run figures.

If there is any funnel flow happening, or build up on the walls or corners, the loads can be reduced.

The details are critical. Did TUNRA get the detailed drawings?

Perhaps the design case hasn't been seen yet. Is the moisture level, density as per the test data? TUNRA use the worst case condition.

Does the ore have the same properties as the ore passing the bin? It is typically impossible to make the connection between the ore used for the TUNRA test and the ore in the hopper. What was the northing and easting and depth of the ore sample location?

Some think they are using the TUNRA method, and their method has errors. Recently, when a large error was noted, the supplier was not prepared to discuss. ■

Hello Gary,

Has this change been documented anywhere? ■

In the absence of factual data on the installations in question my final comment on this subject is that there are two important components to the bin extraction load that have a significant impact on the powder needed.

The first is that the determination of overpressure at the feeder interface is extremely sensitive to the degree by which the opening size exceeds the critical arching dimension. Once flow is mobilised, the magnitude of the overpressure reflects the support afforded to a collapsing arch, so if this is small it indicates a low margin of safety to accommodate adverse variations in the flow properties of the bulk solid. A low starting and running power for this reason would suggest that the detailed design is offering more resistance to flow that should be expected or that the condition of the material being handled is near the worst design condition for the application. Product that is in a more free-flowing condition may be expected to increase the load required to extract the product.

My second point is that the formula published in paper 27 of beltcon2 shows what appears to be a final vertical neck to the belt in Fig.29 and in Chap. 4.5, that the assessment of extraction force is based on tangential shear at the hopper interface. Readers will appreciate that shear strength is a global value, with constituents of molecular particle attraction, particle surface friction and mechanical interference of particle shape that requires expansion of the shearing layer to take place so that the particles can pass one another. Confining forces, such as overpressures and restraint on volume change, constrain the expansion of the failure layer and therefore the condition of confinement during bulk failure has a large influence on the shear resistance that is generated by the latter component. A simple experiment to demonstrate this is to fill a cup containing a spoon with granular sugar and then try to rotate the spoon. The expansion required to permit the material to shear in a confined space can only be secured at the expense of compaction of adjacent particles, and these are virtually incompressible and capable of sustaining a high load to resist compaction.

Rather than shear tangentially, inclining the failure plane from the back of the opening to the depth of the outlet gate will provide a limited, but highly effective, divergence of the failure plane from the dynamic arch running alone the transition zone. Tensile and particle friction forces will still prevail, but the content due to mechanical interference of particle overlap will significantly decrease due to this relaxation of confinement. Such features of fine detail in a feeder interface design can have a remarkable influence on the efficiency of an application and may account for the low power consumption of the installation in question. ■

Derek

TUNRA run courses on materials handling. The updated methods used by TUNRA are available in the course notes.

Prof Roberts publishes papers frequently. The updated methods will be published.

Minerva always uses TUNRA to calculate the feeder forces. ( same for stockpile volume etc) We now provide the calculations from TUNRA to the suppliers during the bid process. We have found the some suppliers typically want to provide a bigger machine. They may have overlooked the reduced loads due to the funnel flow behaviour of the hopper etc. Providing the forces from TUNRA forces a common bid across the suppliers. The suppliers can offer a option if they insist on a bigger machine.

We currently have projects in our system with 11 apron feeders. From D4 to D10, and pans from 1000mm to 2400mm.

Some clients also lock onto say a D10 machine. Even if the numbers from TUNRA say D8, the client insists on D10. The D10 figure is written into the contract with the client. ■

Thanks Gary. It sounds like you are satisfied with the results provided by calculations using Tunra testwork. ■