Fragmentation / Particle Size Distribution Measurement : Screening or Imaging ?
When talking about piled particulate material on conveyor belt, the answer could be both, screening and imaging.
Sampling and screening is about "product grading" and providing product certification particularly in the aggregates industry.
Imaging is about providing online or real-time analysis to rapidly alert operators to out-of-spec behaviour and avoid production losses, avoid excess damage and wear, avoid chute blockages, and prevent safety incidents. And additionally allow process control opportunities for crushers, mills, agglomeration, and other particle processes.
Our Principal Scientist Matthew Thurley, delves into this topic, originally featured on LinkedIn
After I wrote the article, So you want a low-cost system for fragmentation size distribution measurement in mining I had some discussions around 3DPM, the product I helped develop, that overcomes the limitations for reliable particle size distribution measurement outlined in that article.
3DPM is a system for fragmentation size distribution measurement on conveyor belt using uses high resolution 3D imaging, and advanced algorithms to mitigate critical sources of measurement error (outlined in the article linked above) so process engineers can get a stable, reliable measurement they can trust for decision making and automatic process control.
I got some feedback asking whether systems that image the surface of the material could give an accurate “product grading” because they do not measure the material that sits underneath.
In this context “product grading” meant that they wanted a system that can replace sampling and screening for providing product certification in the aggregates industry.
They are entirely correct that imaging the surface is not the same as sampling and screening, and I can’t see surface imaging techniques entirely replacing screening as the industry standard for product evaluation in the aggregates industry.
Piled particles on the surface of a belt are subject to segregation error, also known as the Brazil Nut Effect [1] or Granular Convection [2], where larger particles move to the surface when the particle bed is subjected to vibration.
Therefore, the representative-ness (I’m sure that’s a real word) of the surface particles will depend on many factors including, your range of particle sizes, the bed depth, segregation, and the nature of the process that loaded the material onto the belt. However, some of our 3DPM users implemented their own correction factors as their measurement situation (range of sizes and bed depth in particular) suited this adjustment.
We all know that sampling and screening is too time consuming and rarely done. So while the screening/sieving measurement is considered accurate, that accuracy depends entirely on the accuracy of the sampling, and one sample per day might hardly be considered accurate.
By comparison 3DPM can, and does sample 15000 times/day and about a 5m length each time providing a really enormous sample size. This could definitely improve the “accuracy” of the manual sampling process because you could correlate the manual sampling time with the measurements from 3DPM, evaluating the stability of the process at the manual sampling time.
The real value proposition of 3DPM is using it for the step change it can enable in terms of rapid response and automatic control.
If you have it on the primary crusher output, then the benefits include;
- track the PSD and more accurately track the effective crusher gap because you are measuring what actually matters, the output product,
- improved safety as you don't need a person to do the gap measurement on the live crusher
- you can respond if the PSD is trending towards out-of-spec, by dynamically adjusting the crusher gap (if your crusher allows this), or allow a closed loop automatic control strategy to adjust the gap,
- even without doing the dynamic control of the gap, one could schedule the best time to shutdown the crusher and manually adjust the gap, and
- detection of when the product is out-of-spec within minutes, instead of potentially producing hours of out of spec material and only noticing the result when the secondary crushing circuits become overloaded, and have high recirculating loads, possible chute blockages, falling rock safety hazards, and damage to belts, rollers, chutes, liners etc.
So if we ask the question “How accurate is it compared to screening”, I think we should also ask, "How accurate is the sampling?"
But, I think a more useful question is “How can automated online measurement with 3DPM help me optimise and control my process” so it increases production, improves production quality, increases safety, and allows better scheduling of maintenance.
[1] Rosato, A.; Strandburg, K.J.; Prinz, F.; Swendsen, R.H. (1987). "Why the Brazil Nuts are on Top". Physical Review Letters. 58 (10): 1038–41. doi:10.1103/physrevlett.58.1038 PMID 10034316.)
[2] https://en.wikipedia.org/wiki/Granular_convection
---------------
If you are interested to learn more about what is possible now to enable a future of reliable, robust decision making, and automatic control based on fragmentation size distribution, contact IMV and Principal Scientist Matthew Thurley
If you liked this article, please share, and check out the other articles in this series on LinkedIn
Article 2: So you want a low-cost system for fragmentation size distribution measurement in mining
Post 3: Data quality limits trust and acceptance of digital systems in industrial plants
Article 4: Fragmentation size distribution measurement : Screening or imaging
Article 5: Fragmentation measurement is notoriously tricky to do in a repeatable and quantitative way
Article 7: How much can you trust fragmentation measurement technologies? (part 1)
---------------
About the Author
Matthew Thurley has a background in software engineering, image processing, a PhD in 3D imaging for particle size distribution measurement, and has spent 20 years developing algorithms, technologies and systems for robust fragmentation measurement in order to support a future of feedback and control for drill and blast, caving, crushing, and grinding operations.