See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/328202092

Overview of crystallization in vacuum pans in the Colombian sugar industry

Conference Paper · January 2016

CITATIONS READS
0 2,068

4 authors, including:

José D. Tascón-Vidarte Juan Gabriel Rodriguez

Synopsys Centro de Investigación de la Caña de Azúcar de Colombia
14 PUBLICATIONS   6 CITATIONS    3 PUBLICATIONS   0 CITATIONS   

SEE PROFILE SEE PROFILE

Nicolás Javier Gil

Centro de Investigación de la Caña de Azúcar de Colombia
9 PUBLICATIONS   11 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Best operating practices in sugar boiling stations View project

Floor plans from 3D reconstruction of indoor environments View project

All content following this page was uploaded by Juan Gabriel Rodriguez on 01 December 2020.

The user has requested enhancement of the downloaded file.

 OVERVIEW OF CRYSTALLIZATION IN VACUUM PANS IN THE COLOMBIAN SUGAR
INDUSTRY

J.D. TASCON, J.M. LATORRE, J.G. RODRIGUEZ, N.J. GIL

Colombian Sugarcane Research Center, (CENICAÑA)
[email protected]

Keywords: sugar crystallization, seed, vacuum pan, exhaustion.

Abstract

This paper presents an analysis of vacuum pan crystallization in Colombia. Attention to

this process is given because sucrose losses in molasses represent up to 50% of total
losses of sucrose. For this research we gathered several data of boiling stages and strikes
in Colombian sugar mills. A representative boiling scheme for two and three boiling stages
is presented with average data of brix and purity. Typical conditions of actual equipment
in vacuum pans are summarized. Additionally, we captured process data and
physicochemical properties under normal industrial conditions. Sugar crystallization was
evaluated in terms of crystal quality, sucrose recovery and energy consumption. We
propose best operational practices that aim at profitable processing of sugar even under
manual control of vacuum pans. The strategies lead to achieving higher or constant
recovery rates and a reduction in boiling time of up to 15%.

Introduction

The Colombian sugarcane industry is located along the Cauca River Valley. In 2014 the
harvested cane was 24 283 248 tonnes, producing 2 843 947 tonnes of sugar and 406 504
000 liters of ethanol (Asocaña, 2015). Fifteen sugar mills are located in this region which
include 7 refineries and 6 distilleries The general data of productivity are 119.2 ± 25.7
tons of cane per hectare, 13.8 ± 3.1 tonnes of sugar per hectare, 14.98 ± 0.51 fiber % cane,
and factory performance 11.63 ± 0.45 kg sugar / 100 kg cane at 99.7 Pol (Cenicaña, 2015).

Colombian sugar mills are classified as dual and non-dual depending on whether or not
they have an ethanol distillery. According with this classification a sugar mill will have two
or three boiling stages respectively. The types of sugar produced are: raw sugar, white
sugar and refined sugar.
Nowadays, one of the main concerns in Colombian sugar industry is related to reduction
in purity (apparent) of mixed juice. This reduction impacts directly in boiling house
recovery (BHR) as it has been seen in last eight years, with exception of 2010 when
climatic conditions affected even more sucrose recovery, (Figure 1). The Colombian sugar
industry is increasing mechanical harvesting (from 39% in 2012 to 51% in 2015) and it has
been found that it is related to increases in extraneous matter in cane (3% for handcutted
burnt cane and 10% for mechanical harvesting green cane). The Colombian Sugarcane
Research Center (Cenicaña) is developing studies towards the improvement in juice
treatments, process control, pan boiling technologies and operational practices to
mitigate this impact. In this study, we exhibit an overall view of Colombian sugar industry
pan boiling stations and specific analysis of performance in B massecuite process.

92.4
y = 0.6166x + 37.255
R² = 0.8118
92.0
Boiling House Recovery (BHR)

2012 2008
91.6
2013
2009
91.2 2011
2014
90.8
2015 2010

90.4
86.5 87 87.5 88 88.5 89
purity (apparent) in mixed juice
Figure 1. Relationship between boiling house recovery and purity (apparent) in mixed juice
in Colombian sugar industry (year average values)

Materials and Methods

A generalization of operational practices and equipment in Colombian sugar mills are

presented. For this purpose we captured data from 4 sugar mills with dual operation and 3
sugar mills with only sugar production. These factories represent 58% of the Colombian
sugar industry and equivalent to 65% of the cane processed per year.
A tool for vacuum pan evaluation is the circulation hyperbole (CH) factor proposed by
Cárdenas et al. (1996). The usefulness of this factor is for comparison of vacuum pans
and/or strikes. The equation describing the circulation hyperbole is:.
ℎ
𝐶. 𝐻. = 𝐷2 ∙ 𝑑 2 ∙ Equation 1
𝐻

Where: D is the pan diameter; d is the downtake diameter; h is the calandria height; and H
is the massecuite height, the single variable parameter in ¡Error! No se encuentra el
origen de la referencia.. The fixed parameters D, d and h are related with the vacuum pan
design. Therefore, the strike height H yields different behaviors in circulation due to
operation routines. A pan with higher values of CH factor corresponds to higher
crystallization rates. Furthermore, the CH factor is proportional to circulation and
evaporation rate.

In a further section three vacuum pans are compared with CH factor. The design
specifications are given in Table 1. The 3 vacuum pans have mechanical stirrer. In addition,
three strikes of B massecuite in pan 3 are compared.

Table 1. Batch vacuum pan design specifications

Pan 1 Pan 2 Pan 3
Pan Diameter D [m] 4.86 5.15 4.9
Downtake d [m] 2.25 1.751 1.11
Calandria height [m] 0.991 0.92 1.2
Pan Volume [m ] 3 54 46 42
Massecuite A A B

In a later section massecuite under industrial conditions was evaluated in one sugar mill. A
total of 11 strikes in the same batch vacuum pan were analyzed. The vacuum pan is
processing B massecuite and corresponds to pan 3 of Table 1. We collected analytical and
process data and undertook mass and energy balances to validate the overall data. For
analytical data we sampled input and output material and massecuite each for 15
minutes. The massecuite was nutsch filtered to get mother liquor samples. Process data
was acquired every 15 seconds during strikes and recorded in a HMI reference SM100
ABB. The variables monitored were: stirrer speed, vacuum pressure, barometric
condenser temperature, massecuite temperatures (one sensor below and the other above
the calandria), massecuite level, and feeding tank levels. A summary of measurements and
technical specifications are shown in Table 2.
 Table 2. Technical specifications of analytical equipment and process instruments
Measurement Principle Reference Accuracy
Index Instruments 0.02 brix at 0 brix up to 0.04
Brix (analytical) Refractometric
TCR 15-30 brix at 95 brix (@20°C)
Angular Degrees (°)
± 0.02 °Z ± 0.01 angular
pol (analytical) Optical Rotation SacchAAr 880D
degrees (10 to 90°)
(ICUMSA)
Microscope XS-402
Crystal size + Industrial Camera 4μm according to
Optical
(analytical) Aptina CMOS crystallographic lab. provider
Sensor
Reducing sugars
Lane & Eynon - -
(analytical)
Conductimetric
Ash (analytical) - -
(ICUMSA)
Micro-Polar brix LB
Online brix Microwave  0.5 brix
565
Temperature RTD PT100  0.1 °C
Pressure Capacitive Yokogawa EJX430A 0.04% of span
Panametrics GE
Flow Ultrasonic  2 % of reading
PT878

The flow of condensate vapor in the calandria and water flow in the barometric condenser
were measured with a non-invasive flow meter with ultrasonic principle, i.e. Panametrics
GE PT878. Condensate vapor was measured in a 4’’ pipe with vertical orientation and
ascending flow. Water in the barometric condenser was measured in a horizontal 8’’ pipe
upstream of the automatic valve.

Operational Practices in Numbers

The slurry (alcoholic suspension with milled sugar) is prepared by operators. The most
common mill used is the Slurry Mill™ which is a vertical drum with rotating blades that
contains an average of 2500 pellets of ¼" diameter. The sugar milled is refined sugar
produced locally. The mill is operated at constant speed. Milling times vary widely,
depending on the type of mill. For rotating blade ball mills1, milling time varies between
1.5 and 4 hours. The installations with drum ball mills2 presented longer times, i.e.
between 5 and 21 hours. Reduced milling times with drum ball mills were associated with
higher CV (60 %). Rahiman et al. (2013) reported CV values higher than 50 % when milling
times for ball mills were lower than 24 h. In previous studies Gil (2001) found a strong
statistical significance between milling time and CV.

The crystal size target in slurry is 10 μm average and CV maximum of 45 %. The medium
for preparing the slurry is anhydrous ethanol. Normally, 6 liters are used for milling and 2
liter for rinsing the mill. The required amount of slurry for seed strike ranges from 26 to
156 mL/m3 of B and C massecuites, respectively. These values are higher compared with
those reported by SASTA, where 14 and 28 mL/m3 of B and C massecuite are used
respectively (Ninela & Rajoo, 2006). However the difference could be explained by the
crystal size in the slurry or the crystal size target in the massecuite among others.

The crystallization process starts with seed preparation for C or B (for dual sugar mills)
massecuites. In general, the boiling schemes are double magma. The seed is made in a
batch vacuum pan in two stages. In the first stage, the slurry is seeded in mother liquor
which was previously made by mixing A and B molasses to achieve 75 brix and 70 - 72
purity. In this stage control of purity in molasses is the main concern to guarantee high
crystal size growing rates. The crystals grow from 10 μm up to 100 μm. Once the vacuum
pan is full, half of its volume is cut and then feeding continues. The cut massecuite volume
is processed in the same way. In the second stage, crystals are expected to grow from 100
μm up to 200 μm.

Table 3 summarizes the crystal sizes achieved in Colombian sugar mills. The greater
difference is presented in massecuite B. This lower value contrasts with lower strike times
for B massecuites. Table 4 presents the time in hours for massecuite strikes with other
recommended times. In general, B massecuite times in Colombia are slightly shorter.
Additionally, we observe the shorter times in C massecuites and if seed preparation time
is added this time would be greater.

Table 3. Final crystal size for massecuites in millimeters

Massecuite Colombia Hugot (1986) Chen and Chou (1993)
A 0.5 – 0.6 0.55 – 0.65 0.5 – 0.7

1 This mills possesses a vertical drum and shaft with blades rotating at fixed speeds to move the balls
contained in the drum
2
This mill possesses an horizontal drum, with balls inside, rotating over bands at fixed speed
 B 0.35 – 0.4 0.40 – 0.45 0.4 – 0.45
C 0.22 – 0.3 0.25 – 0.30 0.28 – 0.32

Table 4. Operation times in hours for massecuites in the boiling house.

Massecuite Colombia Rein (2007) Hugot (1986) South Africa Fives Cail – Babcock
Rein (2007) Hugot (1986)
First stage 2 – 2.5 - - - -
seed
Second 1.5 – 2 - - - -
stage seed
C 6 9 4–8 9 8
B 3-4 4.5 3–4 6 4.5
A 2.5 – 3.5 3.75 2.5 – 3 4.5 3.3

Figure 2 shows the basic scheme of two strikes, which corresponds to A and B
massecuites, including the average and standard deviation for the materials involved in
Colombian sugar industry. This process occurs in dual sugar mills which feed B molasses to
the distillery. The B massecuite is a full seeding crystallization process. In this scheme the
seed is prepared with slurry and A molasses. From B massecuite, B magma is obtained (92
± 2 brix, 94 ± 2 % purity). For crystallization of A massecuite, B magma and syrup (62 ± 3
brix, 87 ± 2 % purity) are used. The main variation in this scheme is B molasses purity with
58 ± 4 %. This is related to the distillery requirements and the availability of B molasses;
some dual sugar mills with less production of B molasses and/or bigger distilleries will
need B molasses with higher purity or even add A molasses.

Figure 2. Typical configuration with two boiling stages

Figure 3 shows the typical scheme with three boiling stages and double magma, in sugar
mills producing only sugar. The seed is prepared with slurry and A or B molasses. From C
massecuite, C magma is obtained (93 ± 1 brix, 79 ± 5 % purity) and final molasses (89 ± 2
brix, 33 ± 3 % purity). The A and B massecuite are elaborated similar to the two boiling
stages. However, the brix and purity values are slightly different.

Figure 3. Typical configuration with three boiling stages

Equipment in Colombian Sugar Mills

The Colombian industry has batch and continuous vacuum pans. Sugar mills with milling
rates greater than 400 tonnes of cane per hour tend to have continuous vacuum pans for
A and B massecuite, while older batch pan are still used for seed preparation. On the
other hand, small sugar mills tend to have only batch vacuum pans, where 2 or 3 of 7 are
highly instrumented including brix measurement while the others just have basics analog
instruments to measure: massecuite temperature, vacuum and calandria pressure.

Batch vacuum pans operate with vapor 1 (pressure 0.6-0.75 bar) for A and B massecuites
and vapor 2 for C massecuites (pressure 0.35-0.55 bar). The capacities of batch pans range
from 20 to 30 m3 for seed preparation and between 30 and 50 m3 for A and B massecuites
respectively. Continuous vacuum pans usually operates with vapor 2 or 3 (pressure 0-0.15
bar). These pans are mostly horizontal with 12 cells and average capacity of 100 m3,
operating for A or B massecuite The installation of continuous pans has continued as part
of strategies to increase processing capacity and energy integration (using steam of lower
pressure than batch pans); nevertheless, this equipment requires more stability in seed
characteristics (CV, purity, brix) and feed flows.

Most batch vacuum pans have mechanical agitation to improve circulation, pressure
indicators for vapor and vacuum and massecuite temperature. The general trend in
automation is to have simple PID loops for controlling vacuum pressure with condenser
valve and brix control through vapor valve. Continuous vacuum pan are highly automated
instead, because manufacturer offerings.

Evaluation of B massecuite boiling

The boiling process for B massecuite in a batch vacuum pan performed at a Colombian
sugar mill is described as follows. The vacuum pan capacity is 42 m3. Initially, the vacuum
pan which is clean and free of any previous material is taken to vacuum condition at
0.235- 0.25 bar. Then it is fed with 5.5 m3 of C magma and 7.2m3 of A molasses at 60 °C,
completing 12.7 m3 of footing volume. In the middle of this process, the stirrer is started
and set to run at 53 rpm. Later, the operator proceeds to open the calandria valve (vapor
1 is used with nominal value 0.62 bar), adjusting the pressure of vapor in the calandria to
0.07 bar (between 0.035 to 0.135 bar).

The crystal growth stage initiates when the temperature and concentration conditions are
stable. On average a total of 34m3 A molasses is fed to the pan. The feeding is done with
manual control trying to maintain a constant value of 88.5 brix (microwave sensor) and a
temperature of massecuite between 61 and 64 °C. When the feeding is completed the
tightening phase follows until the massecuite reaches 93.8 brix. Finally, it is necessary to
open the vacuum breaker valve and the massecuite is discharged. When the download is
finished the operator proceeds to turn off the stirrer and the residual material in the pan
is cleaned using steam.

The physicochemical conditions are summarized in Table 5. We observe that the average
purity of A molasses is 75.5 % and standard deviation 1.44. The average reducing sugars /
ash ratio (RS/Ash) is 1.93 and standard deviation 0.115. The coefficients of variation are 2
% and 6 % for purity of A molasses and RS/Ash respectively, which are considered low
under industrial conditions. Therefore, the strikes comparison is feasible due to similar
physicochemical initial conditions in the materials.
 Table 5. Statistical physicochemical properties of input and output materials in several B
strikes
Brix Purity [%] Pol [%] RS/Ash
Material Std. Std. Std. Std.
Avg. Avg. Avg. Avg.
Dev. Dev. Dev. Dev.
C magma 88.4 0.71 83.0 2.97 73.4 2.77 1.87 0.077
A molasses 73.4 0.81 75.5 1.44 55.5 1.35 1.93 0.115
B massecuite 93.8 0.33 76.7 1.29 71.9 1.04 - -
B molasses
86.5 0.73 47.5 1.69 41.1 1.57 1.66 0.265
(Nutsche Filtration)

Process variables such as temperature and brix for a B massecuite strike are shown in
Figure 4a. taking into account the brix and purity of the mother liquor the supersaturation
can be estimated using polynomial approximations. We use the Vavrinecz equation with
ICUMSA adaptation (both equations are given in Bubnik et al. 1995) and we validate
supersaturation values with Thieme tables. Several authors have discussed Thieme’s data,
however for the range of purities and temperatures in this study, the data are fully
covered by Thieme in comparison with Saska (2002) or Ziegler (1979) data. Although, the
calculated supersaturation in Figure 4a. is not accurate due to the use of apparent purity
and Thieme supersaturation data, it outlines the supersaturation behavior due to the
temperature variations. In the initial stage, the massecuite temperature varies from 55 °C
to 70 °C. The instability in temperature is a consequence of manual operation and
different temperature in the feeding molasses.

When supersaturation goes close to the lower limit the crystallization rate decreases. We
note this effect in mother liquor purity evolution (Figure 4b), where purity instead of
decreasing continuously remains constant (the operation between 15 and 45 minutes, i.e.
after strike footing). These variations result in longer strikes times when the process is
under non-controlled conditions (i.e. strikes 1 and 3). As previously mentioned, the
crystallization conditions are similar in terms of material properties. However there are
differences in purity exhaustion along the strikes. Exhaustion of overall strikes is
summarized in Table 6. The coefficient of variation for A molasses purity is 1.9 % and
increases for B molasses up to 3.5 %. This increment reflects that manual operation causes
high variations in pan performance in terms of purity drop.

Figure 4c and Figure 4d present crystal size and coefficient of variation (C.V.) respectively.
The B crystals start at about 205 µm and finish at about 250 µm on average. The average
growing rate is 13.5 µm/h, lower than typical reported by Broadfoot (2005) for B
massecuites in continuous pans with adequate conditions of supersaturation and crystal
content (80 µm/h). This difference can be explained as a result of manual operation in
vacuum pans under study that causes variations in supersaturation and high crystal
content estimated as 51 ± 2 % compared with 30 to 40 % recommended by Hugot (1986).
Crystal content is calculated with equation 33.46 presented by Hugot (1986). The
exhaustion is adequate with purity drops of 28 % on average. The C.V. is stable at around
50% during the strike, as we expected in a process with no water addition. However, the
C.V. can be improved from previous stages (seed preparation and C massecuite) with the
right operational practices of slurry preparation (milling time) and supersaturation control
during strikes, such as those exposed by Ninela (2006) and Morgenroth (2013)
respectively.

(a) (b)

(c) (d)
Figure 4. Illustration of physicochemical and process variables during B massecuite strikes.
(a) Temporal response of process variables in a single B strike (brix, temperature and
supersaturation), sampling time 15 seconds. (b) Purity curves of molasses after nutsch
filtration during B strikes, sampling time 30 minutes. (c) Crystal size behavior for B strikes.
(d) Crystal size coefficient of variance (C.V.) for B strikes, sampling time 15 minutes.

Table 6. Statistical results of pol exhaustion for analyzed B strikes

Average Std. Dev.
Exhaustion (ΔPurity) [%] 27.98 2.11
Strike time [min] 215.2 17.2
Exhaustion/Time [%/min] 0.131 0.0156
SJM [%] 75.94 4.63
Purity of B molasses [%] (Nutsch) 47.55 1.69
Purity of B molasses [%] (Centrifuge) 47.91 2.54

The statistical results of thermal energy in the batch vacuum pan of B massecuite are
presented in Table 7. The vapor / evaporated water ratio is on average 1.22, which
corresponds with a typical vacuum pan expected ratio. The evaporation rate is 6.98
kg/m2.h on average.

We propose an index of energy per tonne of evaporated water, in this case it was of 2577
MJ/tonne. Similar results were obtained in Cenicaña for vacuum pans on A massecuite
where its average ratio was 2500MJ/tonne. The net thermal energy consumed was 6.67
MWh. The average quantity of water consumed by the barometric condenser was 612 m3,
this being similar to the expected theoretical value of 602 m3 with approach of 8 °C. The
approach value is inside the desired range according with Hugot (1986).

Table 7. Statistical results of thermal energy for analyzed B strikes

Average Std. Dev.
3
Evaporated Water [m ] 9.367 0.0571
Calandria Condensed [m3] 11.558 0.355
Vapor Pan Factor 1.22 0.055
Evaporation Rate [kg/m2.h] 6.98 0.56
Specific Energy [MJ/tonne] 2577.65 21.121
Total Energy [MWh] 6.707 0.0524
Water Use in Barom. Cond. [m3] 612.559 18.578
kg Vapor/tonne Water 2.643 0.05

The circulation hyperboles are calculated using Equation 1 for pans detailed in Table 1. We
draw the circulation hyperboles over time in Figure 5.a using real massecuite heights. Pan
1 and 2 are similar in design and have similar profiles or circulation. Pan 3 has a smaller
downtake diameter and therefore circulation is slower. Average evaporation rate (Table 7)
in this pan was lower than expected and just near to end of boiling rate (6.2 kg/m 2 h)
referred by Rein (2007).

If we compare the CH factor with the circulation ratio defined in Rouillard (1987) as the
sectional area of tubes/downtake, pan 3 is higher (6.2 m2/m2) than for other pans in
comparison (2.1 m2/m2 for pan 1). These results can be explained in accordance with
Rouillard (1987); where pans with high circulation ratio it is expected to find lower
evaporation rates. Additionally pan 3 presented smaller tube diameters (0.076 m) that are
not favorable for circulation and in consequence reduce evaporation rate even more.
Rouillard (1987) estimated that small tubes diameter reduces evaporation rate as well.

The comparisons of three strikes of B massecuite in pan 3 are shown in Figure 5.b. Two
facts are presented: First we can ensure that if the feeding of A molasses is too slow as in
strike 3, the circulation is no effective due to massecuite brix and viscosity increase, as
well as the slow feeding itself can produces longer strike times. Second, if the feeding is
too fast like in strike 1, the vacuum pan evaporates at a slower rate because there is a
large amount of material to be heated. Therefore, we can conclude that an ideal
circulation hyperbole (strike 2) is consequent with higher evaporation rates (7.76 kg/m2.h)
and in consequence a higher exhaustion rates (0.149 %/min) obtained for this particular
strike compared to averages (Table 6 and Table 7). If exhaustion rates are higher it is
inferred that crystallization rates are higher as well.

(a) (b)
Figure 5. (a)Circulation Hyperboles for 3 batch vacuum pans (b) Circulation Hyperboles
(dotted lines) compared with evaporated water (continuous lines) for some strikes of
massecuite B
In order to reduce the residence time of B massecuite we propose the following
operational practices. First, the feeding molasses tanks must have temperature control.
This ensures low temperature differences between input material and massecuite being
processed. Second, supply the pan operators with a chart with the recommended levels of
vacuum at specific times with the purpose of reproducing an ideal circulation hyperbole to
favor circulation, evaporation and crystallization rates. A specific circulation hyperbole is
more easily ensured in an automatic system under stable conditions of massecuite
temperature and vacuum pressure. These strategies should lead to achieve higher or
constant recovery rates and a reduction in boiling time of up of to 15% even under manual
operation. The strategies are already being tested in a sugar mill, and the residence time
with the B massecuite going from 215 ± 17 minutes to 185 ± 15 minutes with similar
exhaustion results.

Conclusions
 Colombian sugar mills with drum ball mills and shorter milling times presented
higher CV values (60 %), which confirm the relevance of this parameter, for good
quality slurry (low CV), found by previous studies.
 Manual operation of batch vacuum pans enhances instabilities in supersaturation
related to temperature and molasses feeding. Some mills in Colombian sugar
industry need to improve boiling equipment to implement control strategies in
order to achieve reliable automation systems.
 Strike boiling time in batch pans can be reduced with adequate rate of molasses
feed to achieve reductions up to 13 %.
 Crystal size C.V. is constant over massecuites, in particular we observe this in B
massecuite. If a lower crystal C.V. is desired, it is required to be improved from
previous stages of crystallization (C and seed strikes) or assuring slurry preparation.
 Geometric relationships (circulation ratio and tubes diameter) impact in
evaporation rate as well as crystallization rate. In the analyzed strikes when the
evaporation rates went down or up the crystallization rates also did.

References

Asocaña. 2015. Balance azucarero colombiano Asocaña 2000 - 2015. (Colombian sugar
balance 2000 – 2015). Url: http://www.asocana.com.co/modules/documentos/5528.aspx.
Consulted in February 2016.
Broadfoot R. 2005. Design and Operating Criteria for Maximising the Benefit of
Continuous Vacuum Pans. Proceedings of the International Society of Sugar Cane
Technologists 25: 31-39.

Bubnik Z, Kadlec P, Urban D, Bruhns M. 1995. Sugar technologists manual. Bartens,

Berlin, 155.

Cárdenas, G, Paz, D, Diez, O. 1996. Efecto de la cosecha de caña verde en fábrica y

simulación de las estaciones de calentamiento, evaporación y cocimientos. Curso. Estación
experimental Agroindustrial Obispo Colombres Provincia de Tucumán. Convenio Asocaña-
SENA-Cenicaña. San Antonio de los Caballeros. Colombia.

Chen, J.C. and Chou, C.C., 1993. Cane sugar handbook: a manual for cane sugar
manufacturers and their chemists. 12th Ed. John Wiley and Sons, Inc..

Cenicaña (2015). Annual statistics of sugar and ethanol factories. Internal report. Url:
http://www.cenicana.org/web/areas-investigacion. Consulted in January 2016.

Gil NJ, Moreno CA, Alvarez AM. 2001. Metodologías para la caracterización y ajuste de la
suspension (slurry) para semillamiento de templas. Cali. Work document 506.

Hugot, E. (1986). Handbook of sugar cane engineering. Saint-Denis (Reunion).

Morgenroth B, Max H, Mogalle M, Singh H. 2013. Batch pan automation and its impact on
energyconservation in cane sugar factories. Proceedings of the International Society of
Sugar Cane Technologists 28: 1-14.

Ninela M, Rajoo N. 2006. Practical steps taken at Tongaat-Hulett sugar factories to achieve
low target purity differences. Proceedings of South African Sugar Technologies Association
80: 448-461.

Rahiman SN, Madho S, Davis SB. 2013. Assessing the quality of ball-milled slurries.
International Sugar Journal 115: 43-52.

Rein P. 2007. Cane Sugar Engineering. Verlag Dr. Albert Bartens KG Luckhoffstr, Berlin,
Germany.

Rouillard EEA. 1987. Some ideas on the design of batch and continuous pans. Proceedings
of South African Sugar Technologies Association 76-82.

Saska, M. (2002). Boiling point elevation of technical sugarcane solutions and its use in
automatic pan boiling. International sugar journal, 104, 500-508.
 Ziegler, J. (1979). Sugar boiling the syrups in the vacuum pans. The Sugar Journal, 42, 27.

View publication stats