Mechanised Harvesting of Sugarcane:

Getting the Best Outcomes in the Field and at

the Mill

Chris Norris
Chris@NorrisECT.com
 Overview
The proportion of the Brazilian and global
sugarcane crop which is machine harvested is
increasing annually.
Drivers for this change include:
– availability of and health of labour,
– environmental issues,
• soil and moisture conservation
• emissions / public health considerations
– trash utilisation opportunities
• primarily energy
Adoption of Machine Harvesting
Many major producers have achieved high
levels of mechanisation in their harvesting
operations.
Not all countries which have adopted
mechanised harvesting are doing it well:
– Most (including Australia) have significant “room for
improvement”
This presentation focuses on optimising the
harvest operation to maximise the sugar
production value chain.
 Overview: Harvesting
All harvesting systems result in some “loss of value”
between the crop in the field and the product
delivered to the mill, including:
– Loss of product
– “Dilution” of the product with extraneous components
– Deterioration of the product between harvesting and
processing
In sugarcane, harvesting systems can also impact
directly on the crop ratooning performance and
indirectly on soil health
The relative importance placed on these issues can
drive decisions relating to the harvesting strategies
adopted.
Compromise is inherent in all harvesting strategies.
Mechanisation of Sugarcane: The Compromises

The Hawaiian Industry was first Industry to

be fully mechanised.
The system developed responded to:
– The climatic system which meant that for
reliable productivity, crop age needed to be 18-
24 months at harvest, and
– The high yields which were then achieved
made manual harvesting difficult and
expensive.
 Mechanisation of Sugarcane
The Industry argued that no other harvesting
system can deliver the cane to mill as cheaply.
Total cost of harvest loading & transport to mill
< $5.00/t.
This low harvesting cost was claimed to mitigate
costs associated with low sugar recovery, and the
cost of replanting, under Hawaiian conditions.
The last mill closed in 2017
 Mechanisation of Sugarcane
The key issue was that the actual harvesting
equipment and operations must be
considered as only one part of the field to
mill value chain;
Excessive focus on the cost/efficiency of one
component of the total production system
can cause major losses in other parts of the
system.
 International Clients: Identified Losses
Much of the work undertaken by NorrisECT is for
international clients who have recently introduced
machine harvesting. Typically they have observed:
Reduced Crop Cycle Yield:
– Greater loss of stool after machine harvesting relative to
their traditional manual cutting system;
– Consequential accellerated yield decline;
Gross Yield Loss:
– Reduced t/ha delivered to mill with machine harvesting
relative to manual cutting
Reduced ts/tc in the delivered product and reduced
overall sugar production.
Reference Point: Australian Industry

The issues which have impacted on the

Australian Industry are relevant to many
other Industries as they move to full
mechanisation, and in particular chopper
harvesting.
 Mechanisation Impact: Yield Decline?
110 Data for Herbert Production region in Australia
100 Product Delivered 100% mechanisation from 1972
90 Clean cane
product delivered (t/ha)

Clean Cane
80

70

60

50

40 Average sugarcane yields in Australia have been

30 declining coinciding with the introduction of full
mechanisation.
20
1900 1920 1940 1960 1980 2000 2020
harvest year
 Crop Cycle Yield Impacts
Harvested Yield: Large African Estate
120
110
100
harvested yield (t/ha)

90
80
70
60
Pre Full Mechanisation
50
Post Full Mechanisation
40
30
20
2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

2010

2011

2012

2013
harvest commencing

Production fell >20% from 2007/2008 to 2013/2014 despite an

increase in farmed area
 Reduced Crop Cycle
Accelerated ratoon yield decline relative to hand
cutting is nominally associated with:
– Damage to cane stool from the harvester gathering, feeding
and basecutting operations, e.g
• Excessive harvesting speed, machine configuration and basecutter
blade maintenance
– Damage to cane stool from traffic on/near the row, e.g
• “uncontrolled” traffic and row spacing mismatch
– Soil compaction, e.g
• Row spacing and dramatic increase in in-field traffic reducing
available moisture storage
These problems can all be managed.
 Stool Damage by Harvester:
In the past 20 years, the number of harvesters
operating in Australia has reduced by approximately
50%* with area harvested remaining near constant.
The primary strategy to manage this change has been
to increase harvester operating speed.
One reason for continuing ratoon yield decline is
argued to be high levels of stool damage associated
with the increasing harvesting speeds during this
period.
A move to thinner stalked varieties has also occurred
because experience shows these varieties “survive”
the harvester better.
* SRA Harvest Best Practice Manual (anon 2014).
Varietal adoption for Machine Harvesting
Varietal adoption for Machine Harvesting
As Industries mechanise they generally move to
thinner varieties as these varieties “survive” machine
harvesting better.
How much can you bend a cane stalk before it breaks?
Kroes et.al.
How much can you bend a cane stalk before it breaks?
Kroes et.al.

Kroes showed that irrespective of variety or diameter, a

cane stalk failed at approximately the same level of strain in
the rind.

The thicker the stalk diameter the less you can

bend a stalk before it breaks
 Knockdown: Current harvesters
Kroes developed curves
at which 50% of cane
stalks would be broken 700

if originally vertical
stalks
On this basis, most 500

20
stalks fail before the 25
base of the stalk 30
reaches the basecutter
300

blade.
100

0 100 200 300 400 500 600 700 800 900 1000
 Knockdown: Current harvesters
Again, failure of 1200

most stalks will

have occurred
1000

before the 800

basecutter severs
the stalk.
600

20
25
30
400

200

0
0 100 200 300 400 500 600 700 800 900 1000
 Knockdown Damage: Summary
The design of current harvester “front end”
geometry is very aggressive, and can expect to
cause the failure of a large proportion of the cane
stalks prior to basecutting, particularly in erect
crops;
The move by Industries around the world has been
to adopt thinner varieties as they “survive” the
harvesting operation better;
Agronomic strategies such as planting depth and
depth of cover can impact on outcomes.
Harvesting Speed:
 Harvesting Speed:

In most mill areas in Australia, the location of the

harvester is monitored in “near real time” for cane
consignment purposes
Harvesting speeds can now be derived and monitored
This also allows the impact of harvesting speed to be
correlated with rate of ratoon yield decline in individual
fields and compare it with other fields.
 Effect of Harvesting Speed on Subsequent Yield
rrr
Herbert Impact of Speed 2012 on Productivity 2013
110

100
(Yield 13 / Yield 12) x 100

90

80

70
y = -0,4807x2 + 3,0175x + 98,897
Herbert Harvesters Av. Speed 2012 R² = 0,9596
60

50

40
0 2 4 6 8 10 12
Harvester Block Speed 2012
 Key Observations
Initially it was believed that the reduction in yield was
primarily caused by an increase in damage associated
with the basecutting process at higher harvesting
speeds.
Further analysis indicated that the problem was not
this simple: A number of factors were involved,
including:
• Harvesting speed
• Crop size (related also to degree of lodging)
• Pour rate = speed x crop size
• At the same harvesting speed, the larger the crop the higher pour
rate and more damage was observed even at lower speeds;
Effect of Pour Rate on Subsequent Crop Yield
Pour rate is an even more consistent
predictor of yield depression than
harvesting speed:

This implies gathering & feeding

damage by the harvester also
have major impacts on ratooning
of the crop.
Yield change by pour rate – farm level
 Current Research
On the basis of recent data and previous research,
three primary Issues were considered to impact on
stool damage during the harvest operation:
Gathering damage
• Schembri & Garson, Norris Davis etal, Davis & Norris, Davis &
Schembri
Feeding Damage
• Davis & Norris, Kroes & Harrris, Kroes
Basecutting Damage
• Kroes & Harrris, Kroes
 Current Research
A research project is currently being run by NorrisECT
and Queensland University of Technology
The fundamental hypothesis is that matching “front
end” components tip/surface speeds to groundspeed
is a necessary component of performance/design
optimisation.
– At lower groundspeeds component speeds are actually too
high in standard machine configuration
Once matched, more advanced machine cane
interactions/strategies can be used to facilitate
improvements in harvester performance.
 JD 3520 Standard
20 700
18
600
component tip speed (km/hr)

16
14 500

basecutter RPM
12
400
10
8
300

6 200
Spirals: Tip (.2m) Spirals: Top (.45m)
4 Knockdown "Finned"
Buttlifter basecutter RPM
100
2
0 0
0 2 4 6 8 10 12
groundspeed (km/hr)
 JD 3520 Standard
20 900
working speed range
18 800
component tip speed (km/hr)

16 700
14
600

basecutter RPM
12
500
10
400
8
300
6
Spirals: Tip (.2m) Spirals: Top (.45m)
4 200
Knockdown "Finned"

2 Buttlifter basecutter RPM 100

0 0
0 2 4 6 8 10 12
groundspeed (km/hr)
 Typical Trial Treatments
Treatment Harvesting
Basecutter
Speed Comment
Speed (RPM)
(km/hr)
1 Matched: low harvest speed /
450 4.5
pour rate
2 Matched: high pour rate /
800 8.0
harvest speed
3 Standard: low harvest
620 4.5
speed/pour rate
4
Standard: high harvest speed /
620 8.0
pour rate
Clearing Observation Plots
Analysis of damage
 Initial Results
Data from first years trials indicated that very high
levels of damage were being inflicted by the
harvesting operation;
Similar results from all five trial sites in a range of
crops;
– Lodged and erect crops
 Observed Stool Damage after Harvest
Mona Park : Stool Damage
100%
90%
Mode of damage (% of stumps)

80%

53.0%
53.3%

57.4%
70%

63.3%
60%
% Major Damage
50%
% Minor Damage
40% % no damage

30%
39.8%

42.3%

37.1%
32.2%
20%
10%
6.9%

5.5%
4.7%

4.5%

0%
5.5 x 620 5.5 x 550 7.5 x 620 7.5 x 850
 Q 240 Stool Damage: Childers (Petersen):
Major Damage Minor Damage No damage
100%
90%
80%
level of damage (% of stumps)

42% 47% 50% 51%

70%
60%
50%
40% 33%
36% 32% 30%
30%
20%
10% 24% 19%
17% 18%
0%
5.0 x 500 8.5 x 850 5.0 x 620 8.5 x 620
 Q 240 Stool Damage: Childers (Petersen):
Major Damage Minor Damage No damage
100%
90%
High levels of major and minor damage across all
treatments
80%
level of damage (% of stumps)

42% 47% 50% 51%

70%
60%
Lowest speed with matched “front end” had lower
damage
50%
40% This33%
translated to36%
higher shoot 32%
populations and
30%
30% higher stalk numbers at harvest
20%
10% 24% 19%
17% 18%
0%
5.0 x 500 8.5 x 850 5.0 x 620 8.5 x 620
 Population and Yield Impacts
Q 240 Pre-Harvest Stalk Count (stalks/0.5m): Childers (Petersen): Q 240 Yield: Childers (Petersen):
96
6.5
pre-harvest stalk count 95
Gross Yield
6.0 94
93
5.5

gross cane yield (t/ha)

92
stalks per 0.5m plot

5.0 91
6.3 90 94.7
4.5
5.6 5.5 89
5.4
91.9
4.0 88
87 88.7
3.5 88.4
86
3.0 85
5.0 x 500 8.5 x 850 5.0 x 620 8.5 x 620 4.5 x 450 Std 8.5 x 850 Std 4.5 x 850 Std 8.5 x 450 Std
 Population and Yield Impacts
Q 240 Pre-Harvest Stalk Count (stalks/0.5m): Childers (Petersen): Q 240 Yield: Childers (Petersen):
96
6.5 The treatment with lowest damage 95
achieved highest
pre-harvest stalk count
6.0 stalk numbers and highest yield. 94
Gross Yield

5.5
This was consistent across
93 all sites

gross cane yield (t/ha)

92
stalks per 0.5m plot

5.0 91
6.3 90 94.7
4.5
5.6 5.5 89
5.4
91.9
4.0 88
87 88.7
3.5 88.4
86
3.0 85
5.0 x 500 8.5 x 850 5.0 x 620 8.5 x 620 4.5 x 450 Std 8.5 x 850 Std 4.5 x 850 Std 8.5 x 450 Std
 2018 Trials
Because of the high levels of damage observed in all
trials, an additional sub-treatment was introduced
where the plot was manually harvested leaving
stubble app 25 cm long
This stubble length did not interact with the harvester
forward feed rollers
The only damage would then be that caused by the
basecutters
Any damage greater than this in the “standard plots”
was then caused by the gathering and feeding
processes
Harvesting Pre-cut Plots
 Q 240 Stool Damage: Childers (Petersen):
Major Damage Minor Damage Undamaged
100%
90% 22% 26%
80% 36% 36%
level of damage (% of stumps)

53%
70% 60% 58%
69%
60%
43.9%
50% 51.4%
39.2%
40% 43.5%

30% 35.1%
29.3% 32.0%
20% 18.7%
34%
10% 23% 25%
19%
11% 12% 10% 13%
0%
4.5 x 450 8.5 x 850 4.5 x 850 8.5 x 450 4.5 x 450 8.5 x 850 4.5 x 850 8.5 x 450
Std Std Std Std No KD No KD No KD No KD
 Q 240 Stool Damage: Childers (Petersen):
Major Damage Minor Damage Undamaged
100%
More significantly, the level of major damage was also
90% 22%
significantly reduced. 26%
80% 36% 36%
level of damage (% of stumps)

53%
70% 60% 58%
69%
60%
43.9%
50% 51.4%
39.2%
40% 43.5%
The level of undamaged stool approximately doubled in the
30% treatment
29.3%
35.1% where the effect of the gathering and forward
32.0%
20% feed components were 18.7%removed.
34%
10% 23% 25%
19%
11% 12% 10% 13%
0%
4.5 x 450 8.5 x 850 4.5 x 850 8.5 x 450 4.5 x 450 8.5 x 850 4.5 x 850 8.5 x 450
Std Std Std Std No KD No KD No KD No KD
Summary: Harvester “Front End” Issues
Damage to the crop stool during the harvesting
operation is a major issue for mechanised sugar
Industries:
Aggressive “front end geometry” causes a
significant proportion of the damage;
To date, the Industry has pursued a “variety
development” approach, i.e thinner varieties;
This strategy has significant consequences for
other aspects of the sugar cane harvesting value
chain.
 International Issues: Cane Loss
“Where has the cane gone?”
– As estates move into machine harvesting, mills insist
on trash free cane:
– Harvester suppliers respond by recommending high
trash extractor fan speeds and short billet lengths.
Cross referencing machine harvested yields with
hand-cut yields often indicates significant
“missing” cane.
 Cane Loss at Harvest
Where is the missing cane?
The monitoring of cane stalk left in the field by
the harvester is a useful strategy for managing
losses associated with many harvester operating
parameters, e.g.
– Basecutter height
– Billet spillage etc
Monitoring Harvesting Losses
Monitoring Harvesting Losses
 Harvest Loss Log
Morning shift
H17- H15- H15- i12- i12-
Block H19 H16 H16 i13 i12-i13 i13
Mean
Harvester name H1 H4 H5 H6 H7 H8 H9 H10 t/ha
Chips (blown by primary fan) 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Base cutter losses (debris) 0.3 0.3 0.3 0.8 0.8 0.3 0.5
High base cutting 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Pick up losses (long loose cane) 0.5 0.6 0.7 1.5 1.1 1.2 0.9
Uncut stalks (attached to stool) 0.3 0.0 0.0 0.0 0.0 0.2 0.1
Sound Billets (from chopper drums) 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Sound Billets (from elevator) 0.0 0.3 0.2 0.0 0.0 0.0 0.1
Uprooted stools 0.0 0.0 0.0 0.0 0.0 0.0 0.0
TOTAL LOSSES T/ha 0.0 0.0 1.0 1.1 1.2 2.3 1.9 1.8 1.5
 Cane Loss at Harvest
Where is the missing cane?
Monitoring the billet fragments which are
identified as coming from the extractor typically
indicates extractor cane loss to be minimal, or in
some trials, to actually reduce as extractor speed
increases.
The most significant losses, billeting losses and
trash extractor losses, cannot be measured in
this way.
 “Invisible” Losses
The largest sources of losses are:
The chopper systems (billeting the cane)
– Predominantly juice loss, and
The extractor (trash extraction)
– Billets which are extracted by the extractor fans are
effectively “dissociated”, the resulting product is
difficult to identify or collect
Both are significant sources of “invisible” losses
and are often much higher than the total of the
“visible” losses found on the ground after harvest.
Look first at billet length and billeting losses.
 Billet Length / Billeting Losses
The operator can change billet length by changing
the speed of the feedtrain rollers.
Reducing billet length improves load density:
• Transport cost benefits
• Mitigate the impact of increasing trash levels on load
density
– Manufacturers now option machines for very short
billet length, based on client demand:
• 4 & 5 blades/drum now common ( 8 & 10 blade systems),
with 12 blade systems being available from aftermarket
suppliers
 Billet length: Which is best option?

4 blades/drum: 150-
170mm billets

2 blades/drum: 380mm billets

Côte d'Ivoire
 Billet Length v’s Load Density
450

Ramu 2002 green Burned

Average load density (kg/cuM)

400

350

300

250
15 20 25 30 35 40
Billet length (cm)
 Billet Length v’s Load Density
450

Ramu 2002 green Burned

Average load density (kg/cuM)

400

350

300 Reducing billet length is an effective strategy for increasing

the transport load density
– 5cm reduction in billet length ( 20 cm to 15cm) gives an
approximate 10% increase in load density
250
15 20 25 30 35 40
Billet length (cm)
 Billet length v’s EM:

PL LARSEN; PA PATANE; I ASAMOAH, (2017) BENCHMARKING CANE SUPPLY QUALITY IN THE HERBERT, BURDEKIN, PROSERPINE AND PLANE CREEK REGIONS
Proc of ASSCT .
 Billet length v’s EM:
Improved trash extraction is often given as a
reason the reduce billet length:
Within the “normal” billet length range
reducing billet length has no consistent effect
on final leaf EM levels in the cane supply.

PL LARSEN; PA PATANE; I ASAMOAH, (2017) BENCHMARKING CANE SUPPLY QUALITY IN THE HERBERT, BURDEKIN, PROSERPINE AND PLANE CREEK REGIONS
Proc of ASSCT .
 Billet Length v’s Load Density/Binweight
Burned cane Invicta 2002
7

6,5
bin weight (t)

5,5

4,5

4
150 160 170 180 190 200 210 220 230
billet Length (mm)
 Billet Length v’s Load Density/Binweight
Burned cane Invicta 2002
7

In samples from 62 different harvesters

6,5 delivering clean burned cane, there was no
clear relationship between billet length and
bin weight (t)

6 transport bin weight.

5,5 Billet diameter was also measured and
5
recorded (as average for bin).

4,5

4
150 160 170 180 190 200 210 220 230
billet Length (mm)
Billet Diameter v’s Load density/Binweight
Burned cane Invicta 2002
7

6,5 y = 2,675ln(x) - 2,6363

R² = 0,9509
raw bin weight (t)

5,5

4,5

4
10 15 20 25 30 35
billet diameter (mm)
Billet Diameter v’s Load density/Binweight
Burned cane Invicta 2002
7
Billet diameter explained >95% of the
6,5
variance in load density.
y = 2,675ln(x) - 2,6363
R² = 0,9509
raw bin weight (t)

5,5

4,5 The move to thinner cane stalk diameters has very

significantly impacted on load density/transport costs.
4
10 15 20 25 30 35
billet diameter (mm)
 Billet Length v’s Load Density
(corrected
Binweight V'stoLength
average billet diameter)
corrected for Diameter
1.1
binweight relative to binweight from billet diameter

1.05 y = -0.366ln(x) + 2.9153

R² = 0.745
1

0.95
After correction for billet diameter, the impact of billet length on
load density was similar to other observations:
0.9
– a 50 mm (5cm) change in billet length gives a 10% change in
load density (binweight)
0.85 – Similar to other data where billet diameter does not change

0.8
150 160 170 180 190 200 210 220 230
billet length(mm)
 Impact of Billet Length and Billet Diameter on Load
Density of Clean Billets (Parkhurst)
600

550

500
load density (kg/cu)

450

400

350 20 mm
300 30 mm
250 45 mm
Ramu green low EM
200
10 15 20 25 30
billet length (cm)
In thicker varieties there is much less advantage in shortening billet
length than in thin varieties:
Billet Length, Diameter & Binweight
Industries have moved to cane varieties with
thinner diameters
In an attempt to manage the impact on transport
load density, billet length has been reduced.
This has impacted on billeting losses
 Harvester Components
Billet length is controlled by the feedtrain (6) & chopper(5) configurations

5 3

1 6
Billet Length v’s Chopper Losses
The chopper system billets the cane, which is
presented by the harvester feedtrain
It is important to understand the process to
manage losses and damage
BSES Chopper Test Rig gave useful reference
data which is still relevant:
– Design of common chopper systems (Differential)
has not significantly changed
 Billet Length v’s Chopper Losses
Trials conducted at realistic processing rates to
represent “real operating conditions”
– 120 & 240 t/hr cane processing rates
Chopper test rig represents “ideal” conditions,
and therefore underestimates the true
magnitude of loses which occur.
BSES Chopper Test Rig
 BSES Chopper Test Rig Interchangeable
Chopper Module
on load cells

Feed
Conveyor
(400 kg cane
+ 50 kg leaf)

High-speed
Cine Motion
Feedtrain roller Camera
speed adjustable
 Billet Length v’s Chopper Losses
Test Rig Operation:
– Feed-rate controlled by speed of feed conveyor;
– Chopper speed was fixed but different feedtrain roller
speed settings were used to achieve different billet
lengths.
– Only mass loss was measured, not deterioration effects;
• 400 kg cane stalk on conveyor
• 390 kg billets and visible pieces sorted from the billeted cane and
trash
• 2.5% loss of cane mass
A total of 127 tests were undertaken.
 Results: Mass Loss 3 blade/drum
Mass Loss during Billeting
10.0

9.0 Billet quality deteriorated as billet length 3 blades/drum

8.0 reduced, and variability in billet length

7.0
also increased.
mass loss ( % )

6.0

5.0

4.0

3.0

2.0

1.0

0.0
10 15 20 25 30
billet length (mm)
 Billeting Loss Results: General
Billeting mass loss is:
– loss per cut x number of cuts/m length of stalk
The minimum loss per cut was achieved at the
maximum billet length setting used for any
chopper system,
– Losses increases as billet length is shortened
Billet damage lowest at maximum billet length
setting
– Damage increased as billet length was shortened
 Billet Mass Losses: Other Configurations
12

10
Billeting Loss (%)

2
2 Blade Drum 3 Blade Drum 4 Blade Drum 5 Blade Drum
0
10 15 20 25 30 35 40
Billet Length (cm)
 Optimising Chopper Performance
Further analysis showed that for all chopper systems
tested:
Billeting losses, billet damage and power consumption
were all minimised when the tip speed of the rollers
was 60-65% of the tip speed of the chopper blades.
At this ratio, the blades were not applying any tension
to the cane bundle as they severed the stalks
At lower ratios, the blades generated high tension in
the cane bundle as they attempted to pull the cane
stalk through the feedtrain, increasing losses and
damage.
 Optimising Chopper Performance
38

36

34 A O
C P
32 M C
A T
E
billet length (cm)

30 R I
P M
2 blades/drum G T
28 I U
3 blades.drum A M
N B
26
4 blades/drum A L
24 L E

22

20

18

16

14
0.35 0.4 0.45 0.5 0.55 0.6 0.65

Chopper/Feedtrain tipspeed ratio

Blunt /poorly adjusted chopper blades
had losses 200-300% of losses for sharp
blades
Juice + Soil = Mud
 Billet Length Field Trials
Field trials in association with Chopper Test Rig
program indicated that total recoverable sucrose
losses were approximately twice the mass loss
indicated by the test rig*.
– Higher actual losses in field v’s test rig
– Higher proportions of damaged & mutilated billets
– Increased rates of deterioration
Further trials in Nicaragua (December 2014)** gave
similar results.
– Significantly reduced CCS% with reduced billet length

* James, M (2003) FINAL REPORT – SRDC PROJECT MCB001 LIFTING THE VIABILITY OF THE
MOSSMAN SUGAR INDUSTRY BY IMPROVING THE CANE SUPPLY
**NorrisECT Visit Report to Client
 Billet Length Field Trials
A series of four field trials in association with
Chopper Test Rig used billet length adjustment in
harvester to give long and shorter billets

Large trial also conducted in Nicaragua in

December 2014, comparing harvesters with 3
blades/drum and 4 blades/drum
 Billet Length Field Trials:
Billet Lengths
TCH TCH Field Test Rig Losses
(mm)
% loss % %
Short Long Short Long % Change
loss
Trial 1 140 175 100.9 103.2 2.28% 7.5 4.2 3.3
Trial 2 170 200 80.75 85.4 5.76% 6.25 4.7 1.55
Trial 3 174 204 56 58.3 4.11% 6.2 4.65 1.55
Trial 4 150 180 1.30% 7 4.2 2.8
Change in crop yield
Average 159 190 3.36% with change from2.30%
159-190mm = 3.36%
6 8
Nicaragua
blade blade
170 230 estimated 1.90% 5 3.8 1.2%
 Billet Length Field Trials
Billet
TCH TCH Field Test Rig Loss
Lengths
(mm) Short Long Difference
% Change % loss % %
Short Long Short Long
loss
Trial 1 140 175 100.9 103.2 2.28% 7.5 4.2 3.3
Trial 2 170 200 80.75 85.4 5.76% 6.25 4.7 1.55
Trial 3 174 204 56 58.3 4.11% 6.2 4.65 1.55
(from report)
Trial 4 150 180 1.30% 7 4.2 2.8

Average 159 190 3.36% 2.30%

Nicaragua 170 230 estimated 1.90% 5 3.8 1.2%

 Billet Losses (120 tph pour rate)
12
2 Blade Drum 3 Blade Drum 4 Blade Drum 5 Blade Drum

10

Nicaragua
Billeting Loss (%)

NQ
2

0
10 15 20 25 30
Billet Length (cm)
 Billet Length Field Trials: CCS
CCS
Short Long % Change
Trial 1 13.9 14.1 1.6%
Trial 2 13.5 13.2 -2.2%
Trial 3 14.8 15.2 2.7%
Trial 4 NA 2.90%
Average 1.24%

Nicaragua 12.3 12.6 2.4%

 Tonnes Sugar/ha: Long v’s Short
CCS TC/ha TS/ha
Change Change Change
Trial 1 1.6% 2.3% 103.9%
Trial 2 -2.2% 5.8% 103.4%
Trial 3 2.7% 4.1% 106.9%
Trial 4 2.9% 1.3% 104.2%
Average 1.2% 3.4% 104.6%
30mm change in billet length setting changed recovery by almost 5%

Nicaragua 2.43% 1.90% 104.4%

60mm change with different choppers (each at maximum billet length setting) gave a
similar change in recovery
 Binweight
Binweight Effect
Short Long
Change
(160 mm) (190 mm)
Trial 1 11.80 11.30 -4.2%
Trial 2 (more trashy) 10.60 9.10 -14.2%
Trial 3 9.78 8.82 -9.8%
Trial 4 NA NA -4.7%

Average -8.2%
Nicaragua (170 &
230mm)
42.0 37.0 -12%
Billet length reduction of 30mm increased binweight by 8.9% (12% Nicaragua for
60mm change)
 Billet Length Field Trials
The loss of recoverable sucrose associated with a moderate
change in billet length (30mm) is in the order of 5%, with a
similar difference by changing from “6 blade” to “8 blade”
chopper systems.
The reduction in transport costs associated with the two
changes were 8% (billet length adjustor on harvester) to 12%
(changing chopper drums).
Depending on relative sugar and transport costs, the loss in
sugar recovery typically significantly outweighs reductions in
transport costs gained by shorter billets*.
– By factor of x4 to x8.
Short billets can be an un-recognised but very significant
source of loss for many Industries .

* Value Chain Modelling for Clients, NorrisECT

 Harvester Cleaning Systems:
•Size and weight issues constrain the size of the
separation system;

•Feed of material from the choppers is always

variable, making efficient separation difficult;

•Proximity to fan of billet trajectory is because of

height restrictions;

•Air velocity profiles are variable across the

chamber.
Harvester Performance: Leaf Extraction
Harvester extractor system aims to separate trash
from billets.
100% separation efficiency is never possible, there
will always be:
– Trash left in the cane , and
– Cane extracted with the trash
Harvesters have relatively poor selectivity due to
design constraints.
Best demonstrated in trials under controlled
conditions:
– “Workshop trials”.
 Monitoring Extraction Performance

C. Whiteing, R.J. Davis, E.J. Schmidt. Evaluation Of Cane Loss Monitoring Systems. (2004)
Proc Aust. Soc. Sugar Cane Technol., Vol. 26, 2004.
Collection & weighing of all material
Inputs - Weighing Cane Before Test
• 2 varieties
•Cane stripped & weighed
•Weighed trash added to get
10% & 15% leaf EM levels
• Cane and trash recombined
on conveyor which fed
material into harvester throat.
•Pour rate (80 tph & 140 tph)
controlled by amount of
cane/trash on conveyor (typ
400 kg) and conveyor feed
rate
•120 tests run
Conveyor system into harvester

•6-8 seconds
(effective) /run
Collection & weighing of all material
Results – Cane Loss Trends 3 blade fan

Fan Speed Cane Loss

RPM (old)
80 TPH 140 TPH Mean
(RPM modern)
1000 1.9% 2.9% 2.4%
(675)
1200 4.6% 4.2% 4.4%
(810)
1400 10.4% 7.8% 9.1%
(950)
 Results – Cane Loss Trends
Cane loss increases substantially as fanspeed
increases from 1000rpm to 1400rpm
For 17 test runs, cane billets, pieces and tiny
fragments were separated from the material
extracted by the fan as in a field “tarp test”
On average, < 25% of cane lost could be found,
which explains why “tarp tests” in the field
significantly underestimate cane loss
 Results – Cane loss mechanism
In an effort to better understand the interaction
between airflow and cane/trash in the cleaning
chamber different coloured cane stalks were placed
in the top and bottom layers of the conveyor
8.5% of cane on top layer was lost through the fan
and only 3.1% of cane on bottom layer. This also
supports the case that trash flow is the mechanism
by which billets are carried out the extractor.
Going further, any increase in trash flow creates an
increase in cane flow out the extractor through the
same mechanism.
– Reducing trash flow by topping can be expected to reduce
cane loss
Results – Extraneous Matter

Pour rate had a significant effect on EM levels

Extraneous Matter
Fan Speed 80 TPH 140 TPH
(rpm)
1000 3.1% 7.6%
1200 2.3% 5.5%
1400 1.1% 3.1%
 Results – Extraneous Matter
At 80 TPH, increasing fan speed from 1000rpm to 1400 rpm
only reduced EM by 2%, but at the same time cane loss
increased by 8%.
This supports field trial results which indicated there were
significant losses at high fan speeds with only marginal
improvements in quality.
The data also indicates that a significant proportion of the
material being ejected by the extractor is sugarcane,
reflecting high sugar levels in trash recorded by other
researchers.
Cane Loss Trends: Modern Machines
Larger diameter and/or more aggressive fan
blade design
– Increased airflow at lower extractor fanspeeds
Cane loss/fanspeed and EM / fanspeed
characteristics remain similar, however
High cane loss can now be achieved at lower
extractor fanspeeds.
Cane Loss Measurement: Field Trials
 Mass Balance: Protocol

Objectives:
Measure losses, by comparisson between net TCH or TSH,
delivered at different RPM, where 500 RPM (low fan) is considered
“losses = very low” situation.
 Mass Balance: Protocol
1200 RPM

900 RPM

500 RPM

singleM
Losses = “very low”
 Mass Balance: Protocol
1200 RPM
T1

900 RPM
T2

500 RPM
T3

1 Self tipping haulout

Row lenght was
measured
 Mass Balance: Protocol
1200 RPM
T1

900 RPM
T2

500 RPM
T3
 Mass Balance: Protocol
1200 RPM
T1

900 RPM
T2

500 RPM
T3
 Mass Balance: Protocol
1200 RPM
T1

900 RPM
T2

500 RPM
T3
 Mass Balance: Protocol
1200 RPM
T1

900 RPM
T2

500 RPM
T3
 Mass Balance: Protocol
1200 RPM
T1

900 RPM
T2

500 RPM
T3
 Mass Balance: Protocol
1200 RPM
T1

900 RPM
T2

500 RPM
T3
 Mass Balance: Protocol
1200 RPM
T1 R6 R5 R4 R3 R2 R1

900 RPM
T2 R6 R5 R4 R3 R2 R1

500 RPM
T3 R6 R5 R4 R3 R2 R1

Wagons are weighted (by pairs of bins)

%Trash and CCS (Hidraulic Press) are determined
 Mass Balance: Protocol
1200 RPM
T1 R6 R5 R4 R3 R2 R1

900 RPM
T2 R6 R5 R4 R3 R2 R1

500 RPM
T3 R6 R5 R4 R3 R2 R1

Calculations :
Net Cane = Gross Cane - Trash
TCH : Wagon Weight / (row lenght x row spacing)
TAH : TCH x Recov Sugar
 Cane Loss Trials: New Guinea May 2010
120
trash
clean cane
100
product delivered (t/ha)

80

R579, free trashing, 90 t/hr pour rate;

60

Increasing harvester fan-speed reduced trash

40 delivered to mill from 9t/ha to 1.5 t/ha, but reduced
clean cane recovery by 20 t/ha;
20

< 2.5 t/ha of cane scrap found from extractor!

0
650 880 880+SE 1150 1150+SE
harvester setting
Cane loss Trial: Visible v’s Invisible Losses
Cane loss Trial: Visible v’s Invisible Losses
Visible losses v's Actualane Loss Case 8800 Zimbabwe
9
visible loss and missing cane (t/ha)

8 8.50

7
visible loss
6
missing cane
5

1
0.62 0 0.36 0.94 0.83
-
600 750 900
fanspeed ( RPM)
 Factors Impacting on Cane Loss
Key factors impacting on cane loss for a particular
harvest event include:
Presentation of crop:
– Lodged v’s erect
Level of trash and harvesting conditions
– Variety
– Damp v’s dry
Billet Length & Billet Diameter
 Nicaragua (Lodged) / Sth Africa (Erect)
Results N36
180 160
EM (t/ha)
leaf cane
170 150
component yield delivered to mill (t/ha)

13.51 clean cane

160
(t/ha) 140
150 154.50
130

product (t/ha)
125
140
10.07 120
7.98 120 118
130 115 117
129.82 129.71 8.02

120 122.59 6.33

110
108
116.66
110 100
38 t/ha of cane lost to extract 7 t/ha 17 t/ha of cane lost to extract 18
100 90 t/ha of trash
of trash
90
80
80 2.0 X 2.25 x 3.3 x 2.25 x 3.3 x 3.3 x
2.5x 650 3 x 800 4 x 800 3 x 950 4 x 950 550 750 750 950 950 1250
harvester setting (km/hr x extractor RPM) harvester setting
 Cane Loss: Different Conditions & Harvesters, Nicaragua

A 650-800 RPM: 24 t/ha (3520) 650-800 RPM: 13 t/ha (3520)

Bb Bb Bb c a b b b c

Trial 3: CH 570 Trial 4: CH 570

160 EM (t/ha)
200 EM (t/ha)
clean cane (t/ha) 180 clean cane (t/ha)
140 40
30 34 160

component yield (t/ha)

component yield (t/ha)

120 26 140
35 31
100 a a 120 A
b 25
100
B B 30
80
c C
60 80 156
117 115 104 60 117 121
40 75 92
40
20 650-800 RPM: 13 t/ha (570) 650-800 RPM: 35 t/ha (570)
20
0 0
4x 650 4x700 4x800 4x950 4x 650 4x700 4x800 4x950
 EM V’s Field Conditions
Field conditions dictate EM levels not fanspeed
18.0
16.0
WET CONDITIONS, SPRAWLED Q117
14.0
Extraneous Matter %

12.0

10.0

8.0
DRY CONDITIONS, SEMI-LODGED Q117
6.0

4.0
'IDEAL' CONDITIONS, DRY, ERECT Q117
2.0

0.0
900 950 1000 1050 1100 1150 1200 1250 1300 1350 1400
Extractor Fanspeed (rpm)
Cane Loss: Field Conditions (Tully 2013)

* Damp conditions,
wet trash in morning
** Same field being
harvested in afternoon.
Billet length V’s Loss Potential

Lower terminal velocity as billet length

reduces increases losses at any fanspeed
 Billet Diameter v’ Loss Potential
Terminal Velocity v's Diameter
(Clayton etal)
30
Terminal velocity (m/sec)

25 y = -0.3571x2 + 5.8929x + 5.75

20

15

10

0
0 1 2 3 4 5
billet diameter ((mm)
cm )
Sugar Loss: Billet Length and Diameter
In an analysis of 45 cane loss trials, SRA
researchers assessed the impact of :
– Long (> 180mm) v’s short (< 180mm) billets
– Thick (> 21 mm diameter) v’s (< 21 mm diameter)
Over the trial program, the combination of
short billets and thin diameter had, on
average 50% higher cane loss than longer
billets and thicker stalk diameter, at the same
harvester settings.
 CANE LOSS TRIAL: Ivory Coast
120
CH 570: R579 20/2/19
Leaf Tops
100
Mutilated /Pieces Clean Cane
11,6

80 9,4
component yield

5,2

60

40 80,9
70,6 68,1

20

0
550 x 3 750 x 5 930 x 5
harvester setting
 CANE LOSS TRIAL: Ivory Coast
120
CH 570: R579 20/2/19
• Increasing fan speed from
Leaf 550 toTops930 reduced
100
leaf levels by 6.3 t/ha but
11,6
Mutilatedreduced
/Pieces Cleanclean
Cane cane
80 delivery by 12.8 t/ha 9,4
component yield

5,2

60• This is lower than many other trials but is within

expectations:
40 80,9

• Long billet length and thicker cane stalk both

70,6 68,1

moderated actual levels of loss.

20

0
550 x 3 750 x 5 930 x 5
harvester setting
 Cane loss trial:
• Increasing fan speed from 550 to 930 reduced
leaf levels by 6.3 t/ha but reduced clean cane
delivery by 12.8 t/ha
• This is lower than many other trials but is
within expectations:
• Long billet length and thicker cane stalk both
moderated actual levels of loss.
 Harvester Performance: Cane Loss & Trash
Extraction
Cane Loss is real, and is mainly “invisible”.
As fan speed and pour rate increase, extraction
systems become less selective with respect to
trash extraction and cane loss increases
dramatically.
– Typically, up to 500-600 RPM cane loss is low, but
maximum 50% trash extraction
– After 500-600 RPM, each additional tonne of trash
removed by the harvester takes increasing amounts of
cane with it.
• Up to 5 tonnes cane/tonne of leafy trash
– At maximum fan speeds, cane loss can exceed 30%.
Poor harvesting conditions and high pour rates
increase both EM and cane loss
 Impact of Increasing Trash Levels

Increasing trashy EM levels

Reduce transport density
– Transport costs
– Unloading system capacity

Increased fibre and impurities:

– Reduced mill extraction, milling rate and
increased boiling house losses
• Reduced sugar quality
 Impact of leaf % on load density:
(Ramu Fan speed trial May 2010, 2 harvesters)

Machines 39 & 40
400

380

360
load density (kg/cuM)

340

320

300

280

260

240

220

200
0 2 4 6 8 10 12
leaf (%) of total
 Yield v’s transport payload: Mauritius
Load Weignt V's Clean Cane /Trip

160

140

120
clean cane yield (t/ha)

100

80
y = -0,0052x + 163,7

60

40
Increasing fan speed
20

0
6.000 7.000 8.000 9.000 10.000 11.000 12.000 13.000 14.000 15.000
haulage unit payload (kg)
Impact of Trash on Mill Performance:
Work at Mossman & Condong Mills by Kent [1] indicated:
“The results showed sugar recovery reduced by 0.9
units for each 1.0 unit increase in cane fibre content
caused by increased trash content.
The sugar analysis showed increased ash and colour,
and some evidence of decreased filterability.
The results indicated that the reduced sugar recovery
was caused by greater Pol losses in molasses,
bagasse and mud, in that order of importance.”

[1] Kent,G.A., Moller, D.J., Scroope, P.D., Broadfoot, R., (2010) The Effect of
Whole Crop Processing on Sugar Recovery and Sugar Quality. Proc of Aust
Soc Sugarcane Technol V32, pp 559-572.
 EM Levels v’s Sugar Recovery
Viator demonstrate that the reduction in cane loss can be negated
by increased milling losses associated with higher trash levels.
Harvester fanspeed 650 850 1050

Measured cane loss Control 6.1 t/ha 16.4 t/ha

ERC of delivered cane Control + 0.7 units + 1.4 units

Recovered Sugar 10.6 ts/ha 10.6 ts/ha 10.0 ts/ha

“Less than 15% of known cane loss could be found”
Viator, R.P. Richard, E.P., Viator, B.J. Jackson, W., Waugespack, H.L., Birkett, H.S. (2007).
“Sugarcane Chopper Harvester Extractor Fan and Groundspeed Effects on Sugar Yield, Cane
Quality and Field Losses”. Applied Engineering in Agriculture. Vol 23 No 1.
 Maximising Sucrose Recovery:
Because of the negative impact of increasing
leafy EM on sucrose recovery, the optimum
operating point still occurs at concerning levels
of cane loss (typically >5 t/ha).
The implementation of strategies such as post
harvest cleaning can facilitate further significant
increases in total sucrose recovery, by allowing
the harvester to operate at low cane loss settings
whilst then supplying the mill with very clean
cane.
 Review of Losses
Tangible costs drive many decisions.
– Harvesting speed and pour rate
– Maximise harvester output
– Billet length & extractor speed
– Transport costs and milling concerns

Most Industries are unaware of the

destruction in Industry value which is
occurring, because of the lack of
knowledge of magnitude of losses
 Maximising Sugar & Value Recovery
Value Destruction is minimised by:
Minimising losses on the harvester
– Reduced extractor fan speed of “fans off”
Minimising losses during the milling process
– Post harvest cane cleaning
Additional income can be achieved by utilisation of the
proportion of the trash resource which comes to the mill
with the cane.
– Additional transport costs more than covered by increased
sucrose recovery
– Trash available as a “no cost” resource for further value adding.
Alternative strategy is “field edge” cane cleaning.
“Field Edge” Cane Cleaning
 Cleaner
Operating
Trash exiting trash chute
Cleaned cane
in elevator of
cleaner
“Field Edge” Cane Cleaning
“Field Edge” Cane Cleaning
Questions &
Comments