Need suggestion for choosing sensor.

[Boopathysuntharan]

Guys need a suggestion for choosing a sensor that could sense the tear or any damage in a conveyor belt. The application is that we have to find the area where the belt is damaged. We need a sensor that could be used for this application. We are connecting the sensor with a microcontroller Atmega8. I have searched online but couldn't get a suitable one.

1. Conveyor type : Open
2. Material it is carrying : Coal
3. Conveyor speed : 5 meters/second
4. Conveyor Length : 500 metres
5. No. of conveyors : 7

 

[danadak]

Seems like thats a tall order because of so many ways belt could be considered a fail.

At some point the belt becomes unloaded such that it could be examined with optical
sensor, such as image recognition. Or simple variation in a line array of sensors spanning
belt width.

Possibly a field of IR sensors looking for heat variation in a damaged region.....

A belt painted with an array of white dots on backside looking for timing variation
between successive dots (eg belt stretch or voids)......

Strain gauges seeing variation in belt tension sensing imminent failure .....?

Is there a specific predominant failure mode that could allow a narrowing of detection
design solutions ?

One can think of a learning system of several sensors arrays of different types, sensing
varying belt properties......


Regards, Dana,

 

[Nigel Goodwin]

I agree with danadak, multiple failure modes would require multiple sensors.

However, conveyor belts carrying coal are VERY, VERY old technology, I'm sure there must already be loads of different sensor systems commercially available.

 

[Tony Stewart]

Define your specs then expand with more details

1. define your acceptance criteria for an output ( light flash ? digital time relative to Index (start of belt) video image ??
2. define existing methods( visual gaps, falling dust, vibration or sound effects etc)
3. If optical , visible or IR (?) laser scanner, optical transmission thru or reflected underside
4. How to detect? in real-time with what marker or alert?

Video loop with image loop storage and timestamps ? with visible, audible, vibration triggers?

Foundations™ Learning Center | Material Spillage

Material Spillage |

foundations.martin-eng.com

High tech solutions

Belt Tear Detection for Coal Mining Conveyors

The belt conveyor is the most commonly used conveying equipment in the coal mining industry. As the core part of the conveyor, the belt is vulnerable to various failures, such as scratches, cracks, wear and tear. Inspection and defect detection is essential for conveyor belts, both in academic...

www.mdpi.com

Either way there must be a budget for loss, prevention costs and detection costs with margin to overload risks and targets with real measurements on costs to create a budget plan.

 

[rjenkinsgb]

The simplest I can think of is a line of light sources & sensors, working on the return section of the belt.

Have the sensors arranged so there are a decent number spread over the belt width, plus some either side at the same regular spacing.

Any damage through the body of the belt should let light through where there should not be any; eg. differences in the number of obscured sensors.

Edge wear would gradually decrease the obscured sensor count.

And, I think that any damage to the tensile structure should cause the belt to kink to one side to some extent before it fails; the stronger section will naturally be pulled towards the centreline to some amount, giving a sideways kick around the weak area.

Continuous monitoring of the overall sensor array should allow detection of any side to side movement significantly above whatever usual random drift rate there is with a good belt, indicating a failing section.

 

[Tony Stewart]

In order to block stray light and detect a faint light from a deep top-side scratch, they tend to use enclosed Xrays and lasers but pulsed narrow 1A IR emitters can be used in sequence at high speed to act as a laser sweep. with synchronous IR detection on the bottom side to create a video image. But you need an index marker for position 0 for vertical a nd horizontal sync then store only the anomalies on the belt thru light transmission in memory. But this may not be good enuf.

In 1977 we designed a system to detect nano-flaws in Monel steel tubes in secondary heat exchangers for Candu reactors. It could detect a change in wall thickness or < 1 mm pinhole out of thousands of km of tubing. Each heat exchanger has 2k tubes x 20m with many heat exchanges and many reactors. Heavy water prices back in the '70's were $10k/kg . If a tube was found bad an explosive plug was used to block the end hole during shutdown The eddy current signature was captured every 0.2mm and could scan 5000 mm/s . (if I recall)

 

[Pommie]

I think any optical solution will be hampered by dust. How to keep sensors clean in a coal environment is a big challenge. Having ridden a belt out of a mine (in the 80s, before all the H&S stuff) I know that dust gets everywhere.

Mike.

 

[ZipZapOuch]

I think any optical solution will be hampered by dust. How to keep sensors clean in a coal environment is a big challenge. Having ridden a belt out of a mine (in the 80s, before all the H&S stuff) I know that dust gets everywhere.

Mike.

A roll of clear PET film can be fitted across the lens, it just keeps moving. New water jet lens covers are also available - soon to be on cars with Lidar - pressurized water followed by a puff of air clean the lens regularly. Coal dust should be easier to rinse off than hi-way speed bug gut splatter.

 

[BR-549]

One might also think of the belt as a dielectric in a capacitor. One could monitor the capacitance of sections of the belt. This might not need near the processing needed for imaging or patterns. A light bar could display the sectional capacitance with a row of bars representing the belt. Defects could be mapped and inspected in person or with tv. One could even video the belt leaving the detector. A belt trail camera. For defect hits.

You can vibrate(sound or mechanical) and air blow to clean the returning belt. Then feed belt thru a top and bottom sectional spring loaded capacitor plates. The plates need to be soft and cheap. Taking the wear, but keeping the area. Sequentially apply signal thru sections and record and display sectional capacitance. One might have to adjust plate size, or limit detection range, for the property of the material dust being conveyed. Both the dust and plate material will become embedded in the belt.

I would try the simplest first. Perfect job for today's microprocessor. See if maintenance will build you a belt cleaner and capacitor jig on returning belt. A torn edge or a rip might be easy to detect. The lacing should be very easy to detect.

Maybe just cleaning the return belt could increase it's life.

 

[ZipZapOuch]

Hire a team of interns to review slo-motion video of the belt.
What is the value of planned/scheduled downtime for repair vs unplanned spontaneous outage? A few real people looking at the belt or video of the belt may be a good value to the owner.

 

[BR-549]

On second thought, the conventional cap setup requires much fabrication, high maintenance and wear and tear on the belt. Use a set of pinch rollers for the cap plates. A section of the roller can be used as a cap and several sets can cover the entire belt. A slip ring for the return pinch rollers(cap) is all that's needed.