WRAP MDD018/23 WEEE separation
techniques
Delft University of Technology
Kinetic Gravity Separator trial report
Abstract
This report details a trial conducted on the kinetic gravity separator at Delft University of
Technology for WRAP project MDD018/23. The aim of the project was to trial innovative
techniques to tackle some of the more difficult separations encountered by primary and
secondary WEEE processors. Recovering fine copper from mixed WEEE is a notoriously
difficult separation. Primary WEEE processors typically use eddy current separators to
extract non-ferrous metals from shredded WEEE. These separators are unable to extract
fine wire and smaller non-ferrous items such as screws and washers because these
components are too small in cross-section to generate sufficient eddy current to create a
separation force. To allow high grade plastic recycling all non-ferrous metals must be
removed from the WEEE plastic fraction. Several techniques have been tested during this
project to attempt to find a solution to the problem.
The kinetic gravity separator is a piece of equipment which Dr Peter Rem and his team at
Delft University have developed to separate materials based on the principle of shape and
density. The separator exploits the different settling velocities of materials with different
shape and density. A large scale pilot plant is available at Delft University where research
into the relatively new technique is still in progress. However, one unit is currently in use in
an industrial application. The kinetic gravity separator (KGS) at Sluiskil in the Netherlands
has a capacity of 10 tonne per hour and separates aluminium and stone from heavy nonferrous metals into three size ranges: 1-3mm, 3-8mm and 8-12mm. The machine uses
conveyors to remove the material from the collection compartments.
The aim of the trial was to test the machine’s ability to separate fine copper, metal and glass
from plastic. Two different materials were tested:

The first sample was a mixture of plastics and fine copper wires with small quantities
of glass, wood, circuit boards, glass and other materials in the size range 8-12mm.

The second material was a coarser mixture of plastic and copper along with other
metals, stone and glass in the size range 10-15mm.
Both materials were processed in the kinetic gravity separator and a successful separation
was achieved. The copper and metals concentrated in the heavy fraction whilst the plastic
concentrated in the middle and light fractions.
WRAP MDD018 WEEE Separation Techniques Delft Kinetic Gravity Separator Trial Report The results of the trial with the first sample gave a copper recovery of 83% with
concentrations of 18% copper, 22% coated wire (assumed to contain 50% copper and 50%
plastic) and 9% other metals in the heavy stream, a total metal content of 38%. The level
of combustible plastic, rubber and wood in the heavy fraction, at around 43%, is still above
the recommended value of 5%. Therefore the material would need to be processed further
to reduce the combustible content to below the specification set by the copper smelters.
The second trial recovered 95% of the copper into a heavy fraction with a total metal
concentration of 76%. The concentration of combustible material in the heavy fraction was
8% which is close to the acceptance limit set by the smelters.
The equipment is a prototype and some problems with the design of the product off-take
system became apparent during the trial. It should be possible to resolve these issues in a
production unit.
If the equipment can be adjusted to cut the combustible content of the heavy product to
below 5% and also achieve throughput of at least 1tonne/hr the payback time would be just
over a year.
Overall the technique demonstrated good recovery of non-ferrous metals and glass from
plastic/ metal mixtures produced by secondary WEEE recycling plants and has the potential
to become a practical WEEE separation technique.
2 WRAP MDD018 WEEE Separation Techniques Delft Kinetic Gravity Separator Trial Report Table of Contents
Abstract.......................................................................................................................... 1
1.0
Information on the trial ......................................................................................... 5
1.1
Photograph of trial equipment ............................................................................ 5
1.2
Description of trial equipment ............................................................................. 7
1.3
Trial objectives .................................................................................................. 9
2.0
Trial number 1 - Recovering fine copper from plastic mix....................................... 11
2.1
Trial objective .................................................................................................. 11
2.2
Feed material .................................................................................................. 11
2.3
Trial 1 ............................................................................................................. 11
2.4
Photograph of result samples ........................................................................... 13
2.5
Analysis of results samples ............................................................................... 15
2.6
Discussion of results ........................................................................................ 17
2.6.1 Throughput ................................................................................................... 17
2.7
3.0
Conclusions from trial....................................................................................... 17
Trial Number 2 - Recovering heavy copper and metal from a plastic mix ................ 19
3.1
Trial objective .................................................................................................. 19
3.2
Feed material .................................................................................................. 19
3.3
Trial 2 ............................................................................................................. 20
3.4
Photographs of product samples ....................................................................... 21
3.5
Analysis of product samples.............................................................................. 23
3.6
Discussion of results ........................................................................................ 25
3.6.1
3.7
Throughput .............................................................................................. 26
Conclusions from trial....................................................................................... 26
4.0
Economic assessment of the technique via a payback calculation ........................... 26
4.0
Overall final conclusion of trial ............................................................................. 29
3 WRAP MDD018 WEEE Separation Techniques Delft Kinetic Gravity Separator Trial Report List of Figures
Figure
Figure
Figure
Figure
Figure
Figure
Figure
Figure
Figure
Figure
Figure
Figure
Figure
Figure
Figure
Figure
Figure
1: Side view of kinetic gravity separator ................................................................. 6
2: Aerial view of kinetic gravity separator ............................................................... 6
3: Product off-take end of kinetic gravity separator ................................................. 6
4: Cut-away sketch of kinetic gravity separator....................................................... 7
5: Plan view showing the product off-take compartments relative the feed point ...... 8
6: Graph of the settling velocities for various particles............................................. 9
7: Photograph of Copper Rich Plastic Feed Material............................................... 11
8: Trial 1 heavy product fraction .......................................................................... 13
9: Trial 1 middle product fraction ......................................................................... 13
10: Trial 1 light product fraction........................................................................... 14
11: Schematic of the trial 1 results....................................................................... 16
12: Photograph of heavy copper plastic mix feed material ..................................... 19
13: Trial 2 heavy product fraction ........................................................................ 21
14: Trial 2 middle product fraction ....................................................................... 21
15: Trial 2 light product fraction........................................................................... 22
16: Trial 2 floaters product fraction ...................................................................... 22
17: Schematic of the trial 2 results....................................................................... 24
List of Tables
Table
Table
Table
Table
Table
Table
Table
Table
1:
2:
3:
4:
5:
6:
7:
8:
Settling velocity and compartment information for the kinetic gravity separator ..... 8
Results collected during trial 1 .......................................................................... 12
Results of hand sorting of trial 1 samples .......................................................... 15
Q and R separation efficiencies for trial 1 .......................................................... 17
Results collected during trial 2 .......................................................................... 20
Results of hand sorting of trial 2 samples .......................................................... 23
Q and R separation efficiencies for trial 2 .......................................................... 25
Economic assessment by a payback calculation.................................................. 27
4 WRAP MDD018 WEEE Separation Techniques Delft Kinetic Gravity Separator Trial Report 1.0 Information on the trial
Trial host: Recycling Laboratory at Delft University of Technology, Delft, The Netherlands.
Trial equipment: Kinetic Gravity Separator (KGS)
Trial date: 14th/15th January 2009
The Kinetic Gravity Separator (KGS) is a system developed by the recycling laboratory at
Delft University during their research into recycling and separation techniques.
1.1
Photograph of trial equipment
Figures 1-3 show views of the kinetic gravity separator. A detailed description of the system
follows.
Overflow chute with screen (Not in use) Rotating inner compartment
Feed hopper onto vibrating feeder
Product pipes carrying material to dewatering screens
Additional water supply to entrain Product pipes Water Pump
5 WRAP MDD018 WEEE Separation Techniques Delft Kinetic Gravity Separator Trial Report Figure 1: Side view of kinetic gravity separator
Rotating vanes Top up water pipe
Vibrating feeder Figure 2: Aerial view of kinetic gravity separator
Product dewatering screens Water reservoir Heavy products collection bin
Product chutes Light product collection bin Figure 3: Product off-take end of kinetic gravity separator
Middle products collection bin
6 WRAP MDD018 WEEE Separation Techniques Delft Kinetic Gravity Separator Trial Report 1.2
Description of trial equipment
This equipment exploits differences in the terminal falling velocities of particles in water to
achieve a separation. It is capable of separating the feed material into more than two
fractions and has been used by Delft University to separate light and heavy non-ferrous
alloys in the size range 2-10mm along with various types of plastics. Material which floats in
the separator is collected separately.
Figure 4: Cut-away sketch of kinetic gravity separator
Figure 1 shows the main vessel of the kinetic gravity separator. Material is fed from the
feed hopper into the kinetic gravity separator itself by a vibratory feeder. The product
removal pipes are visible in this photograph. These tended to block with heavy material
during the trial unless they were shaken regularly.
Figure 2 shows the rotating vanes which form an integral part of the separator. They are
approximately 20-30cm long, about 200cm deep and 5cm apart.
Figure 3 shows the dewatering screens and off-take chutes for the light, middle and heavy
products. The material collects in bins at the bottom of the chutes and has to be removed
by hand. The water from the de-watering screens is re-circulated to the water reservoir.
Figure 4 is a cut away sketch showing the inside of the unit and indicates which
compartments different materials will land in.
The feed material is directed by the vibrating feeder into the slots created by the vanes,
which act as isolated separating columns. Particles in each column fall through the water as
the vanes rotate. Since the vanes rotate, particles with high settling velocities will fall into
the collection compartments closer to the feed point, and those with low settling velocities
will be carried further around. The table below compares the rotation rate of the separator
to the settling velocities for each product off-take compartment.
Table 1 shows the range of settling velocities at the fastest rotation speed used in the trial
of 6 seconds per rotation and the slowest rotation speed of 35 seconds per rotation. It also
shows which compartments were used to separate the light, middle and heavy product
7 WRAP MDD018 WEEE Separation Techniques Delft Kinetic Gravity Separator Trial Report fractions. Figure 5 shows a plan view of the product compartments relative to the feed
point.
Compartment
Settling
velocity at 6
seconds per
rotation (m/s)
Settling
velocity at 35
seconds per
rotation (m/s)
Product
Fraction
1
0.21 - 0.22
0.03 - 0.04
Light
2
>1
>0.2
Heavy
3
0.5 - 1
0.1 - 0.2
Middle
4
0.33 - 0.5
0.066 - 0.1
Middle
5
0.25 - 0.33
0.05 - 0.066
Middle
6
0.22 - 0.25
0.04 - 0.05
Light
Table 1: Settling velocity and compartment information for the kinetic gravity separator
Feed 1
2
6
3
5
4
Figure 5: Plan view showing the product off-take compartments relative the feed point
Typically the separator will be adjusted so that the heavy fraction contains mainly metals
and stone/glass, the middle fraction will be lighter metals, plastics and possibly some
stone/glass whilst the light fraction will contain plastic only.
The rotation speed of the separator is adjusted so that the slowest settling material lands in
compartments 6 and 1. If the unit rotates too quickly the slow settling material will travel all
the way round into the heavy fraction. If it rotates too slowly light fraction material will land
in the middle fraction offtake compartments.
8 WRAP MDD018 WEEE Separation Techniques Delft Kinetic Gravity Separator Trial Report Delft predicts that the separator should be able to process at least 1 tonne per hour of feed
material. From previous trials they have learned that in order for the machine to separate
effectively the maximum particle size should be no more than three times the minimum size
and there should be a difference of at least 10% in the terminal velocities of the particles.
Prior to the trial at Delft tests were conducted on some of the mixed metal/plastic feed
material at Axion’s laboratory in Salford. Figure 6 shows the settling velocities of various
materials selected from the mixture. The majority of thick copper wires have a settling
velocity above 0.25 m/s. There is an overlap of the settling velocities of the fine copper
wires and the PVC coated copper wires in the range 0.1-0.2 m/s, so separating these into
two fractions from each other may prove difficult. There is some overlap between the
settling velocities of the fine copper wires, PVC wires and heavier plastic particles in the
range 0.8-.12 m/s.
A separation using a settling velocity of around 0.08 m/s should produce a clean plastic
fraction.
Figure 6: Graph of the settling velocities for various particles
The information on the settling velocities was used to adjust the settings of the kinetic
gravity separator during the trial to obtain the most effective separation.
1.3
Trial objectives
The main objective of the trial was to separate metal and if possible glass from other
components of mixtures derived from mixed WEEE, and produce a saleable non-ferrous
fraction. Copper smelters in Europe require a maximum of 5% combustible material in the
non-ferrous fractions that they process. Non-combustible materials such as stone and glass
9 WRAP MDD018 WEEE Separation Techniques Delft Kinetic Gravity Separator Trial Report can be present at much higher percentages because they do not interfere with the gas flows
in the furnace. They fuse and become part of the slag.
1.4
Trial samples
The following materials were chosen for the kinetic gravity separator trial, both of which are
produced at Axion’s polymer processing plant in Salford.
a) Plastic/glass/copper mixture in the size range 8-12mm, derived from small WEEE
separation; and
b) A copper/stone and plastic mixture derived from larger WEEE items. This consists of
copper, plastic and stone in the size range 8-15mm. The copper in this fraction is
much heavier and thicker than the copper found in the plastic copper mixture and
hence should be easier to separate.
In both the samples the copper wires tend to be finer than the plastic/stone/glass and can
be as small as 0.5mm (diameter) x 2mm (length).
1.5
Trial methodology
For the two materials tested during the kinetic gravity separator trial the same methodology
was followed. Prior to commencing the trial the samples of feed material were weighed and
left to soak in water for one hour. This was to ensure that the plastic was fully wetted prior
to adding it to the separator so that it would sink.
The test material was then processed through the machine. Changes were made to the
rotation speed of the machine for each run in order to optimise the separations but the
physical configuration of the machine remained the same during the trials.
Once all the feed material had been processed the product fractions were collected and
weighed. The samples were bagged and labelled ready for return to Axion’s laboratory in
Salford for analysis.
For both of the trials the same analytical technique was used. Samples of each of the
product fractions were taken and hand sorted into the respective components: wood,
plastic, rubber, copper wires, PVC coated wires, circuit boards, stone/glass, other metals and
fines.
The product and reject separation efficiencies were calculated for each trial run.
For this trial the product separation efficiency, Q, is the probability that the target material
(metal) is correctly separated into the heavy product stream.
The reject separation efficiency, R, is the probability that all other materials are correctly
separated into the light or middling streams.
10 WRAP MDD018 WEEE Separation Techniques Delft Kinetic Gravity Separator Trial Report 2.0 Trial number 1 - Recovering fine copper from plastic mix
2.1
Trial objective
The objective of the trial was to separate copper from other components of the feed and
produce a saleable copper fraction containing less than 5% combustible material.
2.2
Feed material
The feed material was copper rich plastic mixture containing fine copper wires and a mix of
plastic types along with glass, PVC coated wires, rubber, wood, circuit boards and glass.
Currently this material is sent to land fill because the copper content is too low for it to be
commercially interesting. This material has a size range of 8- 12mm, and is illustrated in
Figure 7.
Figure 7: Photograph of Copper Rich Plastic Feed Material
2.3
Trial 1
For this material the separator was operated at a rotation speed of 35 seconds per
revolution, corresponding to settling velocities over 0.2m/s for the heavy fraction
After 6 minutes of operation the rotation time was increased to 37 seconds per revolution as
too many wires were landing in the middle fraction. Table 2 shows the initial results which
were collected at the end of the trial. The gain in weight is due to the absorption of water
by the material.
11 WRAP MDD018 WEEE Separation Techniques Delft Kinetic Gravity Separator Trial Report Trial
Material
Rotation
Outlet position
Feed
Start time
End time
Throughput
Product fraction
weights Total
Gain/Loss
sec/rev
1
2
3
4
5
6
kg
1
Cooper rich plastic
35
full
full
full
full
full
1/2 open
5.81
12:11
12:18
kg/hr
Heavies
Middle
Lights
Floaters
kg
kg
49.8
0.77
5.42
0.42
0
6.61
0.81
Table 2: Results collected during trial 1 1
The photographs in the following section show the product fractions. Figure 8 shows the
heavy product fraction where the copper and PVC wires concentrated. Figure 9 shows the
middle product fraction which mainly consists of plastic. The light product fraction is shown
in Figure 10 and there is some wood visible along with plastic.
1
Rotation time was increased during the trial from 35 seconds per revolution to 37 second per revolution after 6 minutes. 12 WRAP MDD018 WEEE Separation Techniques Delft Kinetic Gravity Separator Trial Report 2.4
Photograph of result samples
Figure 8: Trial 1 heavy product fraction
Figure 9: Trial 1 middle product fraction
13 WRAP MDD018 WEEE Separation Techniques Delft Kinetic Gravity Separator Trial Report Figure 10: Trial 1 light product fraction
14 WRAP MDD018 WEEE Separation Techniques Delft Kinetic Gravity Separator Trial Report 2.5
Analysis of results samples
Samples of the heavy, middle and light fractions were taken for hand sorting to determine the composition of each fraction, with the results in
Table 3.
Trial
1
Material
Copper
Rich
Plastic
Fraction
Total
Weight of Fraction
Wood
Plastic
Rubber
Copper wire
g
g
kg
g
g
%
g
Heavy
0.77
770
0.7
0%
33.2
Middle
5.42 5420 14.4 6% 183.5 74% 5.7 2%
Light
0.42
Total
6.61
420
Weight of hand sorted sample
Composition 2.7
%
%
18% 5.8 3% 34.6
1% 113.3 57% 9.8 5%
%
PVC wires
g
%
Stone/
Glass
Circuit Boards
g
g
%
%
Fines
g
%
Other metals
g
%
g
18%
42.2 22% 29.7 16% 1.4 1% 24.5 13% 17.6 9%
189.7
1.0
0.4%
9.0
4%
3.0
1%
3.5 1% 28.3 11%
0.1
0%
248.5
3.0
2%
4.5
2%
4.1
2%
2.8 1% 47.9 24%
9.6
5%
197.7
Table 3: Results of hand sorting of trial 1 samples
15 WRAP MDD018 WEEE Separation Techniques Delft Kinetic Gravity Separator Trial Report Trial 1
Lights
kg
%
Total
0.42
7%
Wood
0.01
1%
Plastic
0.24
57%
Rubber
0.02
5%
Copper wires
0.01
2%
PVC wires
0.01
2%
Stone/Glass
0.01
2%
Feed
kg
%
PCB's
0.01
1%
Middle
kg
%
Total
5.81
100
Other metals
0.02
5%
Total
5.42
93%
Wood
0.32
6%
Fines
0.10
24%
Wood
0.31
6%
Plastic
4.38
75%
Rubber
0.17
3%
Copper wires
0.17
3%
PVC wires
0.38
6%
Stone/Glass
0.19
3%
Heavy
kg
PCB's
0.09
2%
Total
0.77
13%
PCB's
0.08
1%
Other metals
0.09
2%
Wood
0.00
0.4%
Other metals
0.00
0%
Fines
0.82
14%
18%
Fines
0.62
11%
KGS
Trial 1
Copper Rich Plastic
%
Gain
0.8kg
Plastic
0.13
Throughput
50 kg/hr
Rubber
0.02
3%
Copper wires
0.14
18%
PVC wires
0.17
22%
Stone/Glass
0.12
16%
PCB's
0.01
1%
Other metals
0.07
9%
Fines
0.10
13%
Plastic
4.00
74%
Rubber
0.12
2%
Copper wires
0.02
0.4%
PVC wires
0.20
4%
Stone/Glass
0.07
1%
Figure 11: Schematic of the trial 1 results
16 WRAP MDD018 WEEE Separation Techniques Delft Kinetic Gravity Separator Trial Report Q
R
66%
93%
Table 4: Q and R separation efficiencies for trial 1
2.6
Discussion of results
Figure 11 shows the mass balance for the trial. This was calculated from the total weights
collected during the trial and the compositions measured by hand sorting samples of each of
the products.
The light and middle fractions consisted mainly of unidentifiable fines and plastic with only
18% of the copper lost from the feed in these fractions. The heavy fraction is a mixture of
plastic, copper wires, PVC wires, stone/glass, other metals and fines.
Previous work by Axion has shown that the coated wire fraction typically contains 50%
copper by weight. The hand sort categories that are treated as the target metal fraction in
the analysis of sorting efficiency are therefore ‘copper’, ‘coated wire’ (assuming 50% of this
is metal) and ‘other metal’.
The concentration of metal in the heavy fraction, at 38%, is an increase from the feed
concentration of 8%. The heavy material is still unlikely to be attractive to copper smelters
at this metal concentration. Furthermore, the concentration of plastic, wire coating, rubber
and wood in the heavy fraction is 32%, which is well above the target of 5% combustible
material required by the smelters. Of this 32%, 11% is contributed by the plastic
component coated wire and is therefore impossible to separate from the copper without
further size reduction. The heavy fraction may need to be reworked by a different process
in order to reduce the combustible concentration below 5%.
Table 4 shows the Q and R separation efficiencies. The product separation efficiency, Q,
for metals is good at 83%.
The reject separation efficiency, R, is 93%. This needs to be higher because too much
combustible material ended up in the heavy fraction.
2.6.1 Throughput
The throughput measured during the trial was 50kg/hr. This throughput is too low to be
commercially viable. It appeared that the settling separation itself was not overloaded but
that the throughput of the machine was limited by the ability of the product off-take system
to convey heavy material away without blocking. The system uses water jets with venturis
to entrain the material as it lands at the base of the separator and convey it in flexible pipes
up to the dewatering sieves. When the feed to the machine was increased above about
50Kg/hr the pipes carrying heavy material started to block with particles.
2.7
Conclusions from trial
The machine achieved a reasonable separation of copper from the plastic with 66% of the
metals recovered to the heavy fraction. However, the heavy fraction contained only 38%
metals and the combustible component exceeded the 5% limit expected by copper smelters.
17 WRAP MDD018 WEEE Separation Techniques Delft Kinetic Gravity Separator Trial Report This material contained a high proportion of relatively thin coated copper wire particles.
Around half of these were separated into the middle fraction rather than the heavies.
18 WRAP MDD018 WEEE Separation Techniques Delft Kinetic Gravity Separator Trial Report 3.0 Trial Number 2 - Recovering heavy copper and metal from
a plastic mix
3.1
Trial objective
The main objective of the trial is to use the kinetic gravity separator to separate metals from
other components of mixtures derived from mixed WEEE, and produce a saleable metal
fraction.
3.2
Feed material
The feed material for trial 2 was a mixture of heavy copper, other metals and plastic which
was coarser than the trial 1 material, and is illustrated in Figure 12 below. This material was
derived from large WEEE items and was collected at a different stage in Axion’s WEEE
separation process than the first trial sample.
Figure 12: Photograph of heavy copper plastic mix feed material
19 WRAP MDD018 WEEE Separation Techniques Delft Kinetic Gravity Separator Trial Report 3.3
Trial 2
For this material, the kinetic gravity separator was operated with a faster rotation speed of
22 seconds per revolution. This is because the particle size of the feed material was larger
than the previous trial so the settling velocities would be higher.
Table 5 shows the initial results collected at the end of the trial. The gain in weight is due
to absorption of water by the samples.
Trial
2
Material
Heavy cooper plastic mix Rotation
Outlet position
Feed
Start time
End time
Throughput
sec/rev
1
2
3
4
5
6
kg
22
full
full
full
full
full
1/2 open
7.1
14:20
14:26
kg/hr
71
Heavies
4.76
Product fraction
weights Middle
2.08
Total
Gain/Loss
Lights
Floaters
kg
kg
0.47
0.5
7.81
0.71
Comments
very good copper
& glass
plastic & few pieces
of cable
plastic only
Table 5: Results collected during trial 2
A large quantity of thick copper wires can be seen in the heavy product fraction in Figure
13. Figure 14 shows the middle product fraction and there are pieces of PCV coated wire
visible but no exposed copper wires. The majority of the fraction is plastic. The light
product fraction is shown in Figure 15 and there are some PVC coated wires visible again
along with the plastic. Figure 16 is the floating fraction which is plastic and wood.
20 WRAP MDD018 WEEE Separation Techniques Delft Kinetic Gravity Separator Trial Report 3.4
Photographs of product samples
Figure 13: Trial 2 heavy product fraction
Figure 14: Trial 2 middle product fraction
21 WRAP MDD018 WEEE Separation Techniques Delft Kinetic Gravity Separator Trial Report Figure 15: Trial 2 light product fraction
Figure 16: Trial 2 floaters product fraction
22 WRAP MDD018 WEEE Separation Techniques Delft Kinetic Gravity Separator Trial Report 3.5
Analysis of product samples
The results of the hand sorting of the samples from the product fractions is shown in Table 8.
Trial
Material
Fraction
Wood
kg
2
Heavy Copper
Plastic Mix
g
Weight of sample for
hand sorting
Composition Total Weight
of Fraction
Plastic
Rubber
Copper wire
%
PVC wires
g
%
Stone/
Glass
Circuit Boards
Other metals
g
g
g
g
%
g
%
g
%
g
36.2
6%
1.0
0.2%
204.8
32% 19.4
3%
110.4 17%
0.4 0.1% 277.3 43%
649.5
7%
31.6
12%
3.9
1%
0.9
0.3%
268.1
0.0% 0.0
0%
0.8
0.3%
246.8
0%
0.0
0%
120.0
Heavy
4.76 4760
0.0
0%
Middle
2.08 2080
0.4
0.1%
187.2 70%
3.5
1%
23.0
9%
17.6
235.8 96%
0.0
0%
4.2
2%
3.6
1%
0.1
0.9
1%
0.0
0
0.1
0.1%
0.0
Light
0.47
470
2.3
1%
Floaters
0.5
500
10.0
8%
Total
7.81
109
91%
%
0%
0.0
%
%
g
Table 6: Results of hand sorting of trial 2 samples
23 WRAP MDD018 WEEE Separation Techniques Delft Kinetic Gravity Separator Trial Report Floaters
kg
%
Total
0.5
7%
Wood
0.04
8%
Plastic
0.45
91%
Rubber
0.00
1%
Lights
kg
%
PVC Wires
0.00
0.1%
Total
0.47
7%
Wood
0.00
1%
Plastic
0.45
96%
Rubber
0.00
0%
2%
Feed
kg
%
Total
7.10
100
Copper wires
0.01
Wood
0.05
1%
PVC wires
0.01
1%
Stone/Glass
0.00
0.04%
PCB's
0.00
0.0%
Middle
Other metals
0.00
0.3%
Total
Plastic
2.62
45%
Rubber
0.04
1%
Copper wires
1.69
29%
KGS
Trial 2
Heavy Plastic Copper Mix
kg
%
2.08
29%
PVC wires
0.29
5%
Wood
0.00
0.1%
Stone/Glass
1.05
18%
Plastic
1.45
70%
PCB's
0.03
1%
Rubber
0.03
1%
Other metals
2.04
35%
Copper wires
0.18
9%
Gain
0.71kg PVC wires
0.14
7%
Throughput
70 kg/hr
Stone/Glass
0.25
12%
Heavy
kg
%
Total
4.76
67%
Wood
0.00
0%
Plastic
0.27
6%
Rubber
0.01
0.2%
Copper wires
1.50
32%
PVC wires
0.14
3%
Stone/Glass
0.81
17%
PCB's
0.00
0.1%
Other metals
2.05
43%
PCB's
0.03
1%
Other metals
0.01
0.3%
Figure 17: Schematic of the trial 2 results
24 WRAP MDD018 WEEE Separation Techniques Delft Kinetic Gravity Separator Trial Report Copper and other
metals
Q
89%
93%
R
39%
64%
Table 7: Q and R separation efficiencies for trial 2
Copper only
3.6
Discussion of results
The results from the hand sorting have been scaled up to the mass of material used in the
trial and are given in Figure 17 shows the mass balance for the trial. This was calculated
from the total weights collected during the trial and the compositions measured by hand
sorting samples of each of the products.
The floating fraction contained 91% plastic and 8% wood. The lights fraction contained
96% plastic but also contained 2% fine copper wires. The middle fraction contained plastic,
copper wires, PVC wires and stone/glass. 89% of the copper and 93% of all the metals in
the feed were recovered in the heavy fraction.
The heavy fraction contained copper wires, other metals and a small quantity of stone/glass.
The copper concentration was 32% and the total metal concentration was 75%. The
content of combustible materials was 8%. By adjusting settings, such as the rotation speed,
to ensure more copper and metal is collected in the heavy fraction and more plastic rejected
in the middle fraction it is likely that the plastic composition could be reduced below the
target of 5%.
Table 7 shows the Q and R separation efficiencies for this trial.
The table shows two separation efficiencies in each case; one for the copper fraction alone
and one for all the metals in the feed material.
The coated wire fraction is assumed to contain 50% plastic and 50% metal.
The product separation efficiency, Q, is the probability that either copper alone or all metals
are correctly separated into the heavy product fraction.
The reject separation efficiency, R, is the probability that either all non copper materials or
all non-metals are correctly separated into the floaters, light or middle fractions.
For copper alone the product separation efficiency, Q, is 89% which good. The product
separation efficiency for all metals is even better at 93%, because the material contained a
high proportion of compact non-wire metal fragments which settled rapidly and were easily
captured in the heavy fraction.
The reject separation efficiency, R, was 39% for copper alone because the feed contained a
large proportion of other heavy materials (metals, glass and stone) which joined the copper
in the heavy fraction.
The reject separation efficiency, R, for all metals was higher at 64%. It was reduced by the
high proportion of stone and glass in the feed, which mostly collected in the heavy fraction.
25 WRAP MDD018 WEEE Separation Techniques Delft Kinetic Gravity Separator Trial Report Note that some of the other metals may be of little value to the copper smelters so could be
treated as a contaminant.
3.6.1 Throughput
The throughput measured for this material was approximately 70kg/hr. This is too low to be
acceptable for industrial/commercial use. However, again the constraint appeared to be the
product off-take system rather than the settling velocity separation itself.
3.7
Conclusions from trial
The trial objective, to separate a saleable metal fraction containing less than 5%
combustible material was almost achieved. A metal recovery of 93% (89% for copper) was
achieved. The combustible concentration in the heavy fraction was 8% which is close to the
target of 5%. If minor adjustments were made to the settings of the machine to fine tune
the separation it is likely that the specification could be achieved.
4.0 Economic assessment of the technique via a payback
calculation
As the machine showed technical potential, an assessment of the economic feasibility of the
separation was made using a simple pay back calculation, shown in Table 8.
26 WRAP MDD018 WEEE Separation Techniques Delft Kinetic Gravity Separator Trial Report Trial
Equipment
Delft
Kinetic Gravity Separator
Capacity
Cost of unit
te/hr
£
1
200000
Basis of operation
Overall Equipment Effectiveness OEE
Plant Input
hr/yr
%
te/yr
3000
70%
2100
Operating Costs
Water
Consumption
Cost (assuming £2/te)
kg/hr
£/hr
100
0.20
Power
Consumption
Cost (assuming 10p/kW hr)
kW £/hr
15
1.5
£/te of feed
£/yr
1.70
3570
Water and Power costs
Water and Power costs
Labour costs (assuming £15/hr job cost)
45000
Annual process licence costs to Delft
10000
Total Operating Costs
58570
Revenue
Assuming 15% of feed is separated as product
Product extracted
te/yr
Value of product
£/te
315
Assuming only 70% of product is useful metal with a value of £1000/te
700
£/yr
220500
£/yr
161930
months
15
Margin
Payback time
Table 8: Economic assessment by a payback calculation
27 WRAP MDD018 WEEE Separation Techniques Delft Kinetic Gravity Separator Trial Report This assessment assumes a 2m diameter unit with the capacity of 1 tonne per hour and a
build cost for the unit of £100,000 and an installed cost of £200,000.
The basis for operation of 12 hours per day, 5 days a week, 50 weeks a year is assumed,
giving 3000 hours of operation per year.
The water usage has been based on the assumption that 10% of the fresh water is lost per
tonne of material sent to landfill. Therefore the fresh water required is 100kg/hr, with an
assumed water cost of £2 per tonne.
The power cost estimate assumes demand of 15kW for the rotor drive and water pumping
with a power cost of 10p/kW hr.
The calculation assumes a overall equipment effectiveness (OEE) for the separator system
of 70% where the OEE is defined as follows:
OEE  capacity rate x quality rate x availability
actual throughput
capacity rate 
rated throughput
quality rate  % of on specification product
actual run hours
availability 
available run hours
Total job cost for 1 operator is estimated to be £45000 per year.
It is assumed that a technology license from Delft University is required to operate the
machine, estimated at £10,000 per year.
The revenue estimate is based on the amount of non-ferrous metal extracted. An overall
recovery of 15% metal rich material from the feed has been estimated, with a useful metal
content of 70%.
These assumptions produce an estimated operating margin of £161,930/yr. This gives a
payback time of 15 months for the estimated installed equipment cost of £200,000.
28 WRAP MDD018 WEEE Separation Techniques Delft Kinetic Gravity Separator Trial Report 4.0 Overall final conclusion of trial
The results of the kinetic gravity separator trial were promising, particularly for the larger
sized feed material.
The results for the second trial were better than for the first trial as more metal was
recovered at a higher concentration.
For the first sample of material the trial objective was only part met. The machine
recovered 66% of the metal to the heavy fraction. However, the heavy fraction contained
only 38% metal and the combustibles exceeded the 5% limit, which meant that the product
did not meet the target specification.
In trial 2 a metal recovery of 93% at a concentration of 76% was achieved. The plastic
concentration in the heavy fraction was 8% which was much closer to the target of 5%. If
minor adjustments were made to the settings of the machine to fine tune the separation it is
thought that the specification could be achieved.
Throughput of the trial machine was too low to be commercially viable because the product
collection system tended to block at very low throughputs. A production version of the
machine would require a redesign of this system. It appeared during the trial that the
settling velocity separation section itself could handle a throughput of at least 1te/hr.
29