Friday 27th May, 2022

BULKtalk: Conveyor design software

Steve Davis, Senior Bulk Handling Expert, explores the software used for designing conveyors and how to make the most of it.

Steve Davis, Senior Bulk Handling Expert, explores the software used for designing conveyors and how to make the most of it.

As engineers in the bulk handling industry, we are often tasked with conveyor design, and armed with a flow sheet, an arrangement drawing and software of choice we run the design.

Today’s proprietary software will take the input and allow selection of component types and other data and provided there are no flaws in the data will calculate an acceptable selection of many of the components, such as belt, pulleys, idlers, and drive size. The software will generally allow calculation based on more than one design basis, CEMA, ISO5048 and others for specific use. The software cannot determine whether the input data and component selection are correct for the system.

I have seen some designs recently where the data was input ‘verbatim’ by engineers who had not been mentored on the many considerations to be made prior to calculations. There are several documented guides to the design of conveyors including belting suppliers’ manuals that effectively lay out design basis, and the CEMA publication “Belt Conveyors for Bulk Materials”. I have not yet seen guidance on the process parameters that define the inputs to be used in the design.

A worst-case example of this was some design calculations based on the premise of a particular ore and an annual throughput. The young engineer divided the annual tonnage by the number of hours in a year and selected ore properties from a generic source. Belt speed was defined in the flow sheet. It was assumed that the conveyor was a straight incline at maximum angle from the source. The selected parameters worked well in the software but gave a completely incorrect outcome for the application.

Generic bulk materials properties are just that. Ores vary from site to site, with different locations in the same site and with the weather and over time. Generic properties might be acceptable for a prefeasibility study but beyond this we need to be more precise. For conveyor design we need bulk density, minimum for volumetric capacity and maximum for power, and particle sizes. Maximum conveyor incline for the ore is useful. We have several laboratories in Australia that can test the ore and provide the correct numbers, and it is worth getting the additional parameters that relate to chute and hopper design etc. at the same time.

Now we need to consider maximum rock size and skirts. To limit blockage, it is preferred that we have skirt width set at minimum three times the maximum rock dimension. We prefer that skirts are set a maximum of two thirds of the belt width so that when the belt wanders the edges do not run out from under the skirts. With wider skirts the incidence of load point spillage and damage to skirts will increase.

Next, edge clearance. This is provided so that larger rocks, which will roll to the outside of the surcharge do not fall off the belt. If the edge clearance is 50mm and the rock size 150mm there will be spillage. Edge clearance also allows for belt wander caused by off centre loading, where the belt will move to one side or another. This could be caused by a poor or worn chute or an intermittent hang up in the chute.

Have we got the correct surcharge angle? This can be measured in a laboratory but is generally dependent on how the load chute behaves. I have seen several installations designed for 15-to-20-degree surcharge that run at almost zero, and sometimes below.

I reviewed a conveyor system that had significant spillage and found it had been designed for ore from an adjacent site. The new site’s ore was lower density, 1.6 t/m3 actual compared to 1.9 t/m3 design. The volume of ore on the belt was 19 per cent higher and surcharge angle was three degrees lower. Significant reduction in edge clearance resulted in spillage along the conveyors.

Consider how many hours the conveyor will operate? Conveyors are reliable overall, so perhaps 95 per cent available. What about up and downstream? Crushers, as an example might achieve 75 per cent availability, so the annual ore must be handled in 75 per cent of the hours. We have a crusher and conveyor in series, so availability is 0.75 x 0.95 = 0.7125 or 71.25 per cent availability. If there is another conveyor in series, availability drops to 67.7 per cent and so on. The capacity of the conveyor should be annualised hourly capacity / availability or 1.47 x annualised capacity per hour.

It is common for crushed ore to feed a mill. Mills are more reliable than crushers, so we have a crushed ore surge stockpile to manage the mill feed. The stockpile will be emptied after a long crusher shut so the crusher will be oversized for ‘catch-up’ capacity. This catch-up rate depends on stockpile size and how quickly we need to refill it, how quickly the crusher can be fed, and could be as high as 100 per cent for many hours or even days. Our conveyor should be sized for the catch-up rate. There are other parts of most materials handling plants where we might see similar equipment interfaces that have to be considered.

We generally design for steady flow; however, some machines and processes do not operate this way.

Apron feeders are mostly used to feed ore at the upper end of the size spectrum due to their inherent toughness. Although the drawdown rate will be reasonably constant, discharge is variable as ore stream breaks and shears at the apron feeder discharge. Deeper ore beds surge more. Although good conveyor load skirts can reduce the effect, it is not unusual to see 20 per cent to 30 per cent instantaneous surging onto the conveyor. This surge is contained by sizing the receiving conveyor to suit.

Other types of feeders can also produce surging, depending on design and ore. If there is surge on the first conveyor in a system, there will be surge on subsequent conveyors, and into any downstream process.

Many types of filter discharge onto conveyors. Most have a ‘continuous batch’ discharge, where a slug of cake is discharged at regular intervals. At best, the instantaneous flow rate might be twice the average rate. Conveyors and load skirts must consider this.

Ship, rail, and road unloading is often large discrete batches. Batches are intercepted in hoppers and feeders even out the flow. Shiploading is by nature discontinuous as hatches are changed and feed stopped. The average rate is always less than the continuous conveyor rate.

Longer conveyors can generate surge through movement of ore over the idlers if the design does not consider this. ‘Bunching’ or ‘hourglassing’ can change a well loaded belt into a series of discrete regular bunches of ore with gaps between. Ore can spread the full width of the belt and spillage results. This will only occur on longer conveyors, where there is a particular resonance. I have also seen an instance where the surcharge angle increased by a few degrees during transport over two km of horizontal conveyor.

We now look at the conveyor profile. Conveyors are best loaded on a horizontal section such that the ore can settle before entering a curve. For vertical curves a five-degree maximum incline for loading could be allowed. Loading up to the maximum incline is possible and is often done in brown field sites but should be avoided due to potential for material running back. For horizontal curves a straight section of belt before entering the curve is preferred. I like a minimum of 50 m and will go longer for large belts.

At the discharge it is also preferable to have a horizontal belt profile as trajectories are easy to manage. A straight section of belt after a horizontal curve gives time for the belt to stabilise into the discharge.

Avoid loading into a transition zone as it is more difficult to control dust and spillage when skirt design is compromised. In brownfields sites we must often do this as there is no space for the extra length.

Make allowance for the correct length of skirting. This should be on a straight section of belt so that the skirts have a chance to seal the load point.

Identify all the special components of the conveyor at the earliest time. Belt weighers, overbelt tramp magnets and samplers are best installed on horizontal straight sections of belting. All can be installed on an incline, but performance is likely to suffer. Weighers need weigh grade idlers before and after the machine and these must be on the same straight belt section as the weigher. Even more length will be required if calibration chains are proposed. Head end crosscut samplers will need extra height in the transfer, so the discharge from the conveyor may be higher by several metres.

When we went back to my example and checked all these issues, the conveyor became almost 50% longer, almost 10 m higher and was designed for almost 300 per cent of the original capacity based on corrected density, maximum particle, surge, availability, including a belt weigher and horizontal load and discharge. The design software is not able to make these decisions and gave a faultless design output for both cases.

Armed with this preliminary design, it is worth discussing with the layout designer to confirm, or otherwise, that the conveyor will fit. Be prepared for some iteration to get the best compromise.