864 XXXI International Mineral Processing Congress 2024 Proceedings/Washington, DC/Sep 29–Oct 3
control to benefit downstream processing. Customization
of the analyzers to cater for shorter or longer measurement
times allows responses for different parcel sizes to occur.
Measuring over shorter times improves selectivity and each
measured parcel can be diverted into required categories:
ore and waste, low and high grade, or low and high del-
eterious content or combinations of these. Bulk sorting is
effectively an automated grade control process. Parcels that
are product quality are diverted to avoid unnecessary pro-
cessing at Assmang Khumani iron ore operations in South
Africa (Matthews and Du Toit 2011) where one third of
production bypasses beneficiation. Savings in processing
costs (US $5M/yr) are additional to GHG emissions sav-
ings (40,000 t CO2 e/yr) for that diverted material. Ore
losses are also minimized for the diverted product quality
material as jig efficiency is not 100 percent and some ore
would be lost if unnecessarily processed.
Bulk ore sorting is used in mines extracting porphyry
copper ores (Nadolski et al., 2018) where measurement
precisions of 0.02% Cu are achieved on 30 seconds incre-
ments. Copper ore grade variability occurs at different
scales, and this is utilized for bulk ore sorting to increase
ore grade and quality consistency in plant feed (Figure 3).
Analysis of pebble quality can prevent waste being recircu-
lated to the mill but no bulk analyzers are implemented in
this application yet to the author’s knowledge. Some opera-
tions are mine constrained so the diversion of waste from
the ore stream reduces tons available for the mill, unless
mining rates can be increased to compensate for this.
Shorter measurement times are particularly beneficial if
accompanied by high measurement precisions (Kurth 2022
and Scott et al., 2020). Unrepresentative measurements
increase the amount of misallocated material and small dif-
ferences in measurement precision can affect metal value
in ore sent for processing by tens of millions of dollars per
year (Kurth 2022). This highlights the need to customize
the measurement solution to each application.
Analyzer data is used to control blending of different
feed sources into the ore stream and can be used to dilute
high deleterious content as well as optimizing ore grade.
The blending application has proven beneficial at Sepon
copper-gold operation where average copper metal content
in leaching circuit feed is controlled based on analyzer data
(Arena and McTiernan 2011). Longer measurement times
of five minutes, composited over each 30 minutes, help
smooth the highly variable short term results to minimize
blend change frequency. Pyrite addition to the carbonate
ore at Sepon is controlled using analyzer data (Balzan et
al., 2016). Improvements in metal recoveries of up to 10%
are possible through more consistent process feed quality
(Goodall 2021) which can be achieved through using ana-
lyzer data in both blending and bulk sorting applications.
Feed forward control is possible when there is a good
understanding of the sensitivity of the process to changes
Source: Scantech International Pty Ltd
Figure 3. Copper variability each 30 seconds over 16 hours of conveyed flow of porphyry copper ore
control to benefit downstream processing. Customization
of the analyzers to cater for shorter or longer measurement
times allows responses for different parcel sizes to occur.
Measuring over shorter times improves selectivity and each
measured parcel can be diverted into required categories:
ore and waste, low and high grade, or low and high del-
eterious content or combinations of these. Bulk sorting is
effectively an automated grade control process. Parcels that
are product quality are diverted to avoid unnecessary pro-
cessing at Assmang Khumani iron ore operations in South
Africa (Matthews and Du Toit 2011) where one third of
production bypasses beneficiation. Savings in processing
costs (US $5M/yr) are additional to GHG emissions sav-
ings (40,000 t CO2 e/yr) for that diverted material. Ore
losses are also minimized for the diverted product quality
material as jig efficiency is not 100 percent and some ore
would be lost if unnecessarily processed.
Bulk ore sorting is used in mines extracting porphyry
copper ores (Nadolski et al., 2018) where measurement
precisions of 0.02% Cu are achieved on 30 seconds incre-
ments. Copper ore grade variability occurs at different
scales, and this is utilized for bulk ore sorting to increase
ore grade and quality consistency in plant feed (Figure 3).
Analysis of pebble quality can prevent waste being recircu-
lated to the mill but no bulk analyzers are implemented in
this application yet to the author’s knowledge. Some opera-
tions are mine constrained so the diversion of waste from
the ore stream reduces tons available for the mill, unless
mining rates can be increased to compensate for this.
Shorter measurement times are particularly beneficial if
accompanied by high measurement precisions (Kurth 2022
and Scott et al., 2020). Unrepresentative measurements
increase the amount of misallocated material and small dif-
ferences in measurement precision can affect metal value
in ore sent for processing by tens of millions of dollars per
year (Kurth 2022). This highlights the need to customize
the measurement solution to each application.
Analyzer data is used to control blending of different
feed sources into the ore stream and can be used to dilute
high deleterious content as well as optimizing ore grade.
The blending application has proven beneficial at Sepon
copper-gold operation where average copper metal content
in leaching circuit feed is controlled based on analyzer data
(Arena and McTiernan 2011). Longer measurement times
of five minutes, composited over each 30 minutes, help
smooth the highly variable short term results to minimize
blend change frequency. Pyrite addition to the carbonate
ore at Sepon is controlled using analyzer data (Balzan et
al., 2016). Improvements in metal recoveries of up to 10%
are possible through more consistent process feed quality
(Goodall 2021) which can be achieved through using ana-
lyzer data in both blending and bulk sorting applications.
Feed forward control is possible when there is a good
understanding of the sensitivity of the process to changes
Source: Scantech International Pty Ltd
Figure 3. Copper variability each 30 seconds over 16 hours of conveyed flow of porphyry copper ore