By: Mark Wagner
To calibrate, or not to calibrate; that is the question. And it's an important question to ask prior to installing your online bulk material analysis system, regardless of the industry you're in. This article will help you figure out why calibration may or may not be your best bet, and whether you will see value in calibrating your system. Read past the jump for more details.
One of the deficiencies with image-based rock fragmentation analysis technology is the inability to analyze underlying material running on a conveyor belt. So, in normal circumstances, without calibration, fine material is typically under-represented when using image analysis. The same thing can be said for blast fragmentation analysis, but I will touch on this a little later in the article.
If photo analysis technologies represent material on the top layer, and under-represent the underlying material, an operation can compensate for this deficiency by calibrating. (For the how-to guide on calibration, check out this link).
The way I see it, without sounding too much like a broken record, calibrated photo analysis technologies take the best of both worlds: Quantity and quality. Remember? (If you haven't read the other blog post relating to this, click here).
Take the following scenario as an example: Company ABC wants to adjust the SAG feed based on particle size analysis. By doing so, they need to know when to draw fine material from the stockpile and when to draw coarser material. Calibration allows for the proper "mix" of bulk material in order to optimize the process.
The Swebrec and Rosin-Rammler functions are great for adjusting the distribution curve to accurately compensate for fine material... But what if you are looking strictly at the coarse sizes, say, for oversize detection?
If this is the case, you may want to reconsider the calibration process. Allow me to explain:
A calibrated system is making certain assumptions about material underneath the top layer, and may even bias your coarse material fractions when trying to adjust the distribution curve. So, if you are planning on having a conveyor belt shut down if it detects material over x size, you may want to reconsider taking the chance of biasing your results. Instead, an uncalibrated system is going to get you very useful data that will allow you to stop/start/act on out-of-spec readings, and optimize your process.
Other reasons why clients don't calibrate?
When material is too large, it is much more difficult to manually sample material, and this could be cost-prohibitive. This issue arises most often on the blasting end of the application, where blast optimization relies heavily on detecting relative changes in material size.
For post-crusher analysis, some clients detect liner wearing by relative increases in the size of material, therefore making calibration unnecessary.
In summary, calibration certainly has its uses inside the mining and aggregate industries, and more specifically, when finer material sizes are a necessary part of the quandary; however, uncalibrated systems are still very useful in optimizing your process and for tracking relative change.