BULKtalk, Conveyors, Transfers, Chutes

BULKtalk: Weighing in on belt conveyors

Bulk handling expert Steve Davis explains the ins and outs of belt weighing systems and how they should be selected, installed and maintained to get the best accuracy possible.

When transporting bulk commodities on belt conveyors it can be important to know the tonnes moved. Reasons for this could be to inform process monitoring, storage management, custody transfer and feed metering.

To gain this insight, belt weighing systems (belt scales) are used. However, these devices should be selected, installed and maintained to meet specific aspects that define whether the system will achieve accuracy goals.

General aspects

Belt scales do not directly weigh bulk on the belt. They are electromechanical devices that convert an instrument output to a weight indication. There are four generic types:

1. Electromechanical belt scales, using load cells and a deflecting load support frame, are the most common by far, and generally the most accurate. Deflection of the frame under load varies strain on load cells. Belt speed is measured, and an integrator converts the outputs to a weight.

2. Optical systems measure cross section of the load on the belt. The belt shape and speed, and the load bulk density must be consistent as these define the weight in the integrator. These are best with finer materials and consistent feed rates.

3. Nuclear systems measure absorption of radiation in the load and converts to a weight. Variations in the belt and the material coefficient of absorption can affect accuracy.

4. Impact weighers are also available for discharge chute installation. Deflection of the impact plate correlates to the flow of material in the discharge. These are suited for dry fine product.

There is a common misconception that belt scales are 100 per cent accurate, which can lead to concern when two weigh systems do not match exactly. Conversely, there might be something wrong when two systems match consistently closer than realistically possible, for example, within 100 tonnes on a 200,000-tonne batch with a combined error of 0.05 per cent.

Belt scales are sold based on error bands, such as being accurate to ±0.5 per cent. This accuracy cannot be attained just by installation on a conveyor. The accuracy of all belt scales depends on calibration, and repeatability through good installation and maintenance.

Related stories:

External influences must be considered and mitigated. When all aspects of the belt scale and installation are in good condition, calibrated correctly, and external influences are minimal, the integrated weight readout will vary between 99.5 per cent and 100.5 per cent of the true weight on a ±0.5 per cent accurate scale. For a detailed explanation, I recommend reading Control Systems Technology Knowledge Bases. Good accuracy is possible, but it doesn’t just happen.

Belt scales have a span, which is the range in which they deliver repeatable outputs. Go beyond the span and accuracy will drop. The best accuracy is generally in middle to top of the span range.

Zero checking is not calibration. It is a regular reset of the electronic zero to accommodate change due to temperatures and other daily variables. Zero movement should oscillate around a centre value. Gradual drift in one direction likely indicates a fault.

Calibrating a belt scale is done by simulating a load on the belt scale. There are two common methods; hanging known weight on the scale and using roller weigh chains over the scale. The belt is running empty in both methods.

Stored in place, static weights are safer and simpler to apply, easier to maintain, and if split correctly will cover the load span in two or more stages. Chains have several safety risks associated with installation and operation and need to be stored carefully. Neither method is effective if there are any aspects of the belt scale design or the condition of the equipment that prevents this. If precise calibration is expected then the calibration weights should simulate full load, and preferably one or two intermediate loads.

There are two accepted and similar methods used to certify belt scales for custody transfer. NMI / OIML R50-1 is the Australian Standard and used in many other countries while the USA and others use National Institute of Standards and Technology handbook 44. These standards define how to test the scale and provide minimum requirements for various aspects that are necessary to be tested. The lowest error band for which there is a regulated test available is ±0.2 per cent, and this is for a fully installed and operational conveyor with a belt scale, designed to be tested according to one of the two standards.

There is no other regulated method for testing belt scale accuracy on a conveyor, although there are many ways to confirm accuracy. Purchasing a weigh scale with a stated error band is only the start. It is design, installation, calibration, operation and maintenance of the belt scale that turns this into a precision instrument.

Specification

Do not specify a belt scale, specify a conveyor with a weighing system. Understand the reason for weighing and the accuracy necessary for the process using the weight. If there are multiple scales in a process, consider developing a matrix that identifies maximum cumulative errors in accuracy. Is a belt scale the most appropriate and accurate?

Why do you want to measure weight? What precision? Which conveyor(s) is it best for scale installation? Most belt scales will offer better than ±3 per cent to ±5 per cent accuracy in a basic installation with some regular attention. If this is sufficient, then a simple single/dual roll belt scale could be the solution. Is accurate weighing to ±0.2 per cent required for legal custody transfer and payment basis or as a check for draft survey? Is a repeatable accuracy of ±1 per cent acceptable? What developments are available in belt scales?

Is the load over the scale constant, variable or random? What are the maximum possible upset forces on the scale? What is more important, flow rate or total flow? Will the belt scale output be used to control any aspect of operation? If so, consider location and time delay.

In any situation, the system including the belt scale, the conveyor, the location, and the installation and ancillaries, and provision for calibration and maintenance defines belt scale accuracy. These must all be appropriate if the required weighing accuracy is to be attained.

Many belt scale suppliers have installation handbooks that provide guidance to achieve accuracy. These guides are minimum requirements. The best suppliers offer specific guidance for each application.

What matters?

Support and service are key. A belt scale supplier should understand your requirements and provide guidance. A robust scale that will withstand the rigours of the operation. Real precision is achieved by custom design and longer multi roll-weigh frames.

Conveyors that provide the best possible location for the scale are ideally at ground level, horizontal and with a robust isolated structural support. Transmission of vibration, external loads and the like must be as low as possible. Wind fences may be required. Install the scale according to supplier instructions and all moving parts must be free of restraint.

Use best quality weigh-class idlers with micro adjust frames on the scale and in the lead-in and out zones. Weigh frame and idler alignment must be excellent. Install the belt speed generator correctly. Conveyor belt tracking should be good, and the load should be central to the belt. The belt should sit correctly into all weigh idlers from empty to fully loaded.

The electrical installation should be excellent, especially the shielding of signal wires and earthing. Locally mounted integrators deserve to be in an ingress protection rating IP66 cabinet.

What changes?

Several common changes affect the accuracy of the belt scale.

• As the conveyor belt wears out, its weight reduces. Idler rolls also wear and lose weight. Calibration mitigates this. Replace seized idler rolls at the earliest opportunity, as they change the dynamics on the weigher and replace all idlers on a frame at the same time with weigh class idlers.

• Any spillage build-up or carry back on the weigh idlers or frame will change the weight reading. Cleaning will remove build-up, but it is better to prevent this.

• Spillage build-up on gravity take-up weights will increase belt tension and change the weight reading. Deliberate change to the take up weight has the same impact.

• Any damage to the structure will affect alignment and accuracy. Addition of pipes and cable trays to the structure may have a similar impact.

• Changes to the conveyor, such as increasing speed or loads will change the scale dynamics, which will change the accuracy.

• With age, conveyor components are replaced often with different components to the original. Belt weight and idler type may not affect the operation of the conveyor, but they will affect belt scale accuracy.

• Damaged, corroded or incorrectly installed calibration weights or chains give false calibration.

• Wear on the speed sensor wheel causes belt speed to be incorrect. Bouncing of the wheel caused by a wheel or belt irregularity gives incorrect speed.

• Damaged belt or splices affects belt scale idlers and affects accuracy.

• High pressure hosing may damage the belt scale.

• Flooded belt load bends or breaks the scale.

• Operator expresses doubts over accuracy of the scale and installs a second check scale resulting in two different weight readings. Which is correct? No two scales will provide the same weight reading due to the error band, but they should trend in parallel.

• Perception that belt scales do not need maintenance. Any moving part will eventually wear, seize or fail. Load cells gradually fail over time, from fatigue and overload, moisture ingress and cable failure.

Case histories

1. A belt scale had a random step change in reading under steady operations. It was caused by a combination of water-undercut foundations, loose structural bolts, and a nearby large compressor that randomly started and vibrated the belt scale. The issue was corrected by repairing the damage and using isolation pads on the compressor.

2. An impact scale located in a discharge chute after an air slide could not control the material flow. When the flow set point was increased, the scale increased flow through the upstream feed control. The trajectory changed and some of the material missed the impact plate, meaning the scale did not measure all flow, causing a further increase through the control system until the system choked. This was corrected by changing the size and location of the impact plate and recalibrating the system.

3. A nuclear belt scale was manually calibrated using sacks of ore. This was the incoming conveyor from a mine. In operation, the scale consistently measured 10 per cent lower than the mine supplied through a truck weighbridge. After several checks and recalibrations, the trucking contractor was observed weighing every 10th truck twice, the belt scale was ok.

4. A customer complained that the certified belt scale for custody transfer at a port consistently did not agree with the draft surveys of the ships loaded over the scale. The unrealistic expectation was of an exact match. On evaluation, we found that the difference between scale and survey was a randomised ±50 tonnes on a cargo of 75,000 tonnes. With draft surveys being at best ±2 per cent or ±1500 tonnes there was cause to doubt one or other. The surveyor used the belt scale reading and randomised a fictitious survey.

5. A night shift plant operator used a brick on the belt scale weigh frame to reduce plant feed but maintain feed records. At lower feed rates, the plant was easier to manage.

6. During a site visit to an older facility we were advised that the weigh feed system no longer controlled feed correctly. Many possible reasons were given for the problem, but no one had thought to check the belt scales. They were dirty, corroded and had no maintenance for some time after a maintainer had retired. The system worked well after some maintenance and calibration.

Send this to a friend