Quality control of Limestone by PGNAA

Hello Joseph,

We have been using Thermo Fisher PGNAAs for quality control of preblended limestone/clay stockpiles at Cement Australia for over 10 years now. The system has excellent availability and reasonably good overall agreement with XRF values, even at relatively low concentration levels.   Typical statistical comparisons for MgO, K2O and Na2O are shown below;-

MgO 

Mean of MgO analyses = 0.65%

90% of the time MgO is within +/- 0.15% of XRF values

70% of the time MgO is within +/- 0.10% of XRF values

K2O

Mean of K2O analyses = 0.26%

90% of the time K2O is within +/- 0.06% of XRF values

70% of the time K2O is within +/- 0.025% of XRF values

Na2O

Mean of Na2O analyses = 0.15%

90% of the time Na2O is within +/- 0.07% of XRF values

70% of the time Na2O is within +/- 0.035% of XRF values



This link has further information;-

http://www.thermoscientific.com/content/dam/tfs/ATG/CAD/CAD%20Documents/Third-Party%20Papers/Cement,%20Coal,%20Minerals%20Sampling%20and%20Online%20Analysis/Cement%20Online%20Elemental%20Analyzers/D00459~.pdf

Regards,

Ted.

Hello Ted, 

  Thank you for your kindly response for this question.  The performance of your PGNAA looks pretty good.  Can you tell me how to collect XRF and CAN data and what the interval of your statistic is?  And is there any relationship between limestone thickness on conveyor and CAN analysis results?

Best regards,

Joseph

Hello Joseph,

The data represents five years worth of data, or about 900 values. (each value being an individual limestone/clay stockpile). The XRF value is the average of the K2O, Na2O and MgO of the raw meal produced from those individual stockpiles.

When we first installed the PGNAA we did some tests on conveyor material thickness versus analysis results and found that there was no appreciable difference between results from full belt loading (1000t/h) down to about 300t/h. Below 300t/h the differences in PGNAA and XRF analyses became unacceptable. We therefore set the conveyor loading in the PGNAA software to only report and aggregate analyses if the belt loading is >300t/h.

The method we used to determine the critical lower limit for the belt loading was to firstly take a representative bulk sample of crushed limestone/clay by crash-stopping the belt when fully loaded and then  shoveling off the contents of a random  2m section. This sample weighed about 300kg. The whole sample was then crushed to <10mm in the laboratory and 5 representative portions of it were analysed by XRF and the mean and std.dev. for each element calculated. 

The bulk sample was then transported to the PGNAA belt and manually layered on the belt in a 2m long strip, positioned just before the entrance of the analyser . We calculated the kg/m belt loading for several tonnage rates between 100 and 1000t/h by dividing the required t/h by the belt speed (m/s). 

e.g.  At 100t/h the belt loading in kg/m for a belt travelling at 2m/s is [(100 * 1000)/(60 * 60)] / 2 = 13.9 kg/m. So, for a 2m strip of belt the weight required is 2 * 13.9 = 27.8kg

   t/h     Weight required for 2m strip

  100                27.8 kg  

  200                55.6 kg     

  300                83.3 kg

  500               138.9 kg

  800               222.2 kg

1000               277.8 kg 

Using these weights we layered the calculated amounts of analysed limestone/clay on the belt just downstream of the PGNAA and then manually moved the conveyor forwards until the 2m strip of material was directly under the analyser detectors. We analysed the sample for 1 hour and collated the individual one minute analyses. These were then statistically compared the XRF analysis values for the material.

It shouldn't be too hard to perform this same test with any PGNAA installation to determine the minimum acceptable belt loading for the instrument.

Hope this helps you, Joseph....

Regards,

Ted.