Gold accountability when treating high grades concentrates is not easy but still there is way:

• to run full batches. It means to completely empty/discharge the feed cone. It should help you with mass accountability (tons per batch).
• note of the total volume used in the leach batch (before washes).
• a solid sample (leach tails) before full discharge.
• take note of the transfer volumes and grades for pregnant and wash solutions (going to pregnant).

With all this information you can follow next calcs:

Feed Grade (g/t): (Solid Tails Grade * feed tons + Pregnant solution grade * total volume used) / feed tons

Note that “total volume used” does not include wash volumes…just the one used on the leach step.

Leach Recovery (%): 

• Grade – Solid Tails Grade) / Feed grade) * 100
• solution grade * Total volume used) / (Feed grade * tons)) * 100

Leached gold reporting to preg solution (g/batch):
Pregnant solution grade * transfer volume + Wash Solution grade * Volume

Leached gold reporting to tails (g):
(Feed grade – Solid Tails Grade)* tons – Leached gold reporting to preg solution

Think that I’m not forgetting nothing. You can also check the remaining solution volume on the ILR (and grade) but is not really important will be compensated on the next batch.

If, at the end of the batch you are leaving some solution on the “solution cone” then you should record grade of the initial volume of the batch to properly consider these gold too.

If you want more information send me a private email and will try to help.

Whatever, maybe you should contact the Gekko guys, I’m sure they have some spreadsheet to help you. 

Also, am not sure what you mean with "change in volume". To account feed gold shouldn't be an issue as you have the initial volume and preg solution grade. Then I'm guessing you have a closed circuit recirculating the solution through EW until you get the desire final grade and also that your final volume is less than the initial one due to evaporation. So based on it you must recalculate your tails grade:

Final grade * final volume / initial volume

Then

Plated gold = (feed grade - recalculated tails grade) * initial volume

Also do not forget to check gold in slags. If still there is a difference can be associated to cathode cleaning efficiency.

Just to be clear...
You cannot do feed grade - tails grade if there is a change in volume.
Grade is a relative measure of the concentration you always will try to take all to grams which is the absolute value.

Ah, missing gold from the books...

You probably need to look at assay accuracy since you are dealing with high Grades, also how accurate is your volume measurement for the solution, and any leaks in the system. There no solution staying behind in the cone. 30% is rather a high margin of error. Worst case ask Gekko to asist

How do you assay solution? AA or Fire by Pb boat method or?
IF AA check dilution errors can be significant.

Now the other end of the process is emptying the sludge from EW cell and refining.
How do you empty and collect the EW sludge?, and what's your slag assay?

Just to add on  comment on volume measurement. The loadcell calibration on the solution cone should be checked. ie Fill with water to overflow point and measurement in kg should match volume in L. Recalibrate if out.

Some really good advice regarding the accounting but it is very important to check the assays as advised.

Intensive cyanidation solutions are higher in impurities such as Fe and Cu which will interfere with AAS direct spray analysis and often causes high assays. Using solvent extraction before AA or fire assay is best.

I would recommend performing some solution fire assays since this will have the least chance of interferences.

If there is interference with the AAS you should use the fire assay results for accounting but you can still use the AA for plant control by applying a suitable factor.

Because of the high grade you are dealing with, a small assay error is magnified by the high grade. When final calculations based on the erroneous assays is done, the % is greatly magnified. Another way of checking is to run parallel assay methods for solution assay to see if error is consistent.

There is a lot to consider to accurate sampling and assaying of different streams at various stages of ILR batches. Therefore I want to focus on tails sampling first. How do you sample ILR tails? If the unit works properly % solids in the ILR tails would be about 5% (give and take as I am not trying to be too accurate here, but you should know what it is for your ops). Then the question becomes how do you accurately sample %5 solids in a (tails) slurry, especially given that solid particle size would probably vary from, say 0.5-1 mm, to below 37 microns? Particle density would also vary especially if you have free partial/unleached gold particles in the tails. You should expect to see some light density particles as well. How often do you need to sample? What (equipment) do you use to sample? At what point do you sample? These are all important sources of variability introduced to the gold balance, if you are trying to monitor in-out god balance. In one ILR 1000 batch commissioning I sampled the ILR tails slurry every minute for the entire discharge duration and found out significant variation in individual assays. This is expected. As for the estimation of feed solids tonnes, again this is another important source of variability. The feed cone is never fully of, only solids as it contains water as well. What about fine gold that is discharged from the feed cone to the ground via overflow pipe? How do you account this in your balance?
However if you want to focus on how much gold you are smelting then volume x assay of pregnant barren EW solutions would be the key to accurate monitoring, as you would have a better idea on the weight and assay of the gold bar. These can also be important source of variability especially if the expected pregnant solution grade is high (can be as > 1500 ppm or >2000 ppm. Correct assaying and use of correct dilution factors would be important here.

Thanks for the help. I will give you feed back on the results of the tests. First we will check the assays by running parallel.

l assay methods direct AA and fire assay if the solutions. For the solution tonnages we have verified the numbers. The sludge in the electro-winning cells has gold slimes and we have made it procedural to clean and smelt the sludge every time we smelt.

I am assuming that the change in volume that you are talking about is the difference between the totalizator reading at the start and the end of the process. I agree with Micheal that you should check the accuracy of the volume measurement from the loadcell. Most of the metallurgical accounting problems generally emanate from the inaccurate mass measurement. On the other note it will be interesting to know how the elution estimates compare to the actual production. I am assuming you have a similar set up for the elution and the ILR.

May I ask which figure is higher than the other (actual gold or wet call)? Fact that one value is consistently higher than the other may point to a systematic/procedure issues. I had an experience where actual was always higher and it turned out that we had missed accounting for gold during the washing/rinsing. Process audit recommended

Yes we do have a similar set for elution. For elution side we do not have major variances there are within+/- 5%.

This suggests the error is either in the assaying technique or in harvesting the cell (if you are getting less than predicted)

Another possible contributing factor (although I haven't observed it in practice) could be if the EW feed tank is not mixed and you are taking your start solution sample only from the bottom. It is possible that the initial transfer of high grade pregnant solution is not being mixed with the subsequent low grade wash solution. Check this by taking your normal sample; then a sample from the top and bottom of the tank and finally by mixing the tank and again taking samples from the top and bottom of the tank.

How do you measure your volume, using the load cells or via flow meter, then are you continuously sampling both the pregnant eluate AND the first wash transfers to your electrowinning and your barren eluate after electrowinning prior to transfer/disposal? the reason for my winded question is that 

Your load cells might require calibration therefore you might consider installing a flow meter?

Are you accounting for the density as you might be dealing with a dirty solution thus your solution might be less than expected.

I might be stated the obvious, but remember that usually both your pregnant and first wash goes to electrowinning.

I don't see anyone talking about your tails grade of the concentrate you are leaching. Aren't you sampling it also? Because to me its a key factor in the determination of your GEKKO recovery. Your need the weight feed to the drum and the tails grade to have an overall picture of your gekko performance. So you need to sample the gekko tails once your batch is over.

And Haven't you got a way to get the exact tons of the concentrate you are feeding like a load cell on your feed come to the gekko? because it will prevent you from using the volume and the SG to determine the tons fed to the drum. So it may certainly reduce the error factor in your calculations.

The ILR at the mine is an early manual (non PLC) model with limited instrumentation, so many of the design features which are now standard are not available. These include load cell monitoring and PLC controlled and recorded met accounting numbers.

As I understand it, the problem is reconciling the gold production predicted from EW feed solution volumes and assays against the actual mass of gold harvested from the cell. This is actually after the ILR itself.

The other measurements, which modern ILRs monitor automatically, are important for overall accounting (eg. feed concentrate mass, preg solution mass, wash solution mass, etc) but not directly relevant to the problem. Similarly there are automatic samplers on all PLC controlled ILRs for solution and tails samples.

The main things to look out for would be:

• ILR dry mass added for the batch - review the load cells accuracy on your concentrate hopper.
• Pregnant & wash solution volumes transferred - review the solution cone load cells accuracy.
• Review sampling protocols for the solutions and residual solids.
• Assay accuracy from your assay laboratory to identify the errors involved by the assay methods.

Reconcile the total solution transferred with holding tank to monitor Load cell errors. Cross examine with electrowin cake sample and weight to back calculate the metal recovered.

What are the best automatic sample cutters for the solution streams. I think the problem with solution accountability has been identified. it was due to interference with Fe ions with the use of AAS direct.  With future assay the problem has been rectified.Thank you for your assistance.