SUGAR MILLING RESEARCH INSTITUTE

(Incorporated Association Not For Gain)

UNIVERSITY OF NATAL
KING GEORGE V AVENUE D U R B A N 4001 TELEPHONE: 359192

TECHNICAL REPORT NO. 1141

SUBJECT INVESTIGATIONS INTO OPTIMUM OPERATING

CONDITIONS OF ROTARY VACUUM SCREEN
FILTERS AT ILLOVO SUGAR MILL

PREPARED FOR Process Manager, Illovo Mill

PREPARED BY J.P. Hoareau

DATE 23rd June, 1978

INTRODUCTION

At the request of the process manager, investigations were carried

out by the SMRI during the week ending 17th June, 1978, into the
performance of the filter station at Illovo Mill.

The aim of the work was mainly to determine the optimum working
conditions which could give maximum retention and minimum pol loss.

Illovo mill has three new filters of adequate capacity for their
crushing rate. All analyses considered necessary for a good control
have been carried out. Optimum operational recommendations are made
in this report. It is hoped that they will prove to be helpful.

PROCEDURE

Each test was conducted over 2 hours. Samples of primary mud. feed
cake and filtrates were collected at 10-12 minute intervals.
Primary mud samples were taken from the mud pumps on the clarifiers.
Filter feed was sampled at the overflow of the filter feed tank.
Filter cake was taken across the whole length of the filter drum
at vacuum breaking point.

The amount of bagacillo added to the clarifier mud was weighed and
a representative sample was kept for a screen test to obtain size
distribution.

Total insoluble solids, mud solids and bagacillo ratios were

determined by screen and filtration methods on primary mud, filter
feed and cake. Muddy and clear filtrates were analysed separately
for insoluble solids, brix, purity and pH.
SMRI TECHNICAL REPORT NO. 1141

Time taken for one complete revolution of the drum was noted
for each filter. Temperature of feed and wash water were also
recorded. The quantity of wash water used on cake was calculated
from brix of feed and filtrates.

The distribution of feed across the whole length of the filter was
noted as well as the speed of the mud agitator.

Low and high vacuum, which are closely associated with good operation
of filters, were noted during investigations.

DISCUSSION

The ultimate purpose of a filter station is to recover as much sucrose

as possible from muds removed from the clarifier.

The efficiency of filtering is usually judged by the percentage of

mud solids retained in the cake (retention). We must however keep
in mind that retention must be considered in association with:

a) Sucrose content of cake, and

b) Purity drop between clarified juice and filtrate.

Pol in cake is the sucrose actually lost in the cake, while

filtrate purity reflects the recovery of sucrose extracted from the
muds into the filtrate.

Consequently low pol percent cake will show good performance only
if it is accompanied by a filtrate of reasonably high purity. Data
from analyses carried out at Illovo are listed, below in Tables 1,
2, 3, 4, 5 and 6.

Because of the high temperatures, bacterial destruction of sucrose

is unlikely. Nevertheless samples of mud and filtrate are being
analysed at the SMRI for l a c t i c a c i d content which is a measure
of bacterial activity.

TABLE 1

AVERAGE OF RESULTS OF-. CLARIFIER MUD ANALYSES

o,
Brix Purity % mud Fibre/mud "5 Total
solids ratio fibre insoluable solids

13,60 85,6 4,82 0,259 1.30 0 , X <£

SMRI TECHNICAL REPORT NO. 1 1 4 1 .3

TABLE 2

AVERAGE RESULTS OF FEED AND CAKE ANALYSIS

Feed Cake

% Fibre/ Fibre/
Brix mud mud Pol
mud Fibre
o. Solids M.solids % Solids in Filters
Feed Purity solids 'o
i
ratio in cake fibre ratio cake

2,40
10,14 85,2 4,77 3,10 0,649 11,10 8,70 0,783 1,70 2
1,40 3

1,83 Average j

TABLE 3

FILTRATE ANALYSIS

Purity Purity Mud %

Filtrates Brix Purity clear juice drop solids

Muddy 8,62 82,89 85,90 3,01

0 , 44
Clear 7,84 83,73 85,90 2,17

TABLE 4

ANALYSIS OF MIXED FILTRATES

(After lime addition to feed)

Purity Mixed Mud

Brix Purity
clear Purity juice solids Retention
a
juice drop purity %

6,66 85,74 86,79 1,05 85,78 0 , 25 86

SMRI TECHNICAL REPORT NO. 1141 .4

TABLE 5

OPERATING CONDITIONS

Speed of Low High Cake Temp, wash Temp

Filters drum vacuum vacuum thickness water feed Dilution
(rev/hr) (bar) (bar) (cm) (°C) <°C) % feed

1 15,0 0,6-0,5 0, ou 0, 6-0, 5 80 93,5

2 18,0 0,6-0.5 0,82 0,5-0,4 78 93,0 15-18
3 17,2 0,6-0,5 0,81 0,5-0,3 82 92,8

TABLE 6

COMPARISON OF SIZE DISTRIBUTION OF BAGACILLO

Percentages obtained from IL Percentages considered very good

Aperture % fractions Tyler sieve & Equivalent % fractions

of sieve retained No. (mm) retained

2,36 mm 1,0
1,00 mm 7,0 14 1,17 4,0
0, 60 mm 23,8 22 0,77 22,0
0, 30 mm 41,2 44 0,33 46,0
passing thro' 27,0 passing thro 1
28,0

100.0 100,0

From the data listed in Tables 1 to 6 operation of the filter station

is considered to be satisfactory except for a few points.

The following observations were made:

1. Feeding system of mud to filters was well controlled and

continuous. Level control in the filters' troughs was good.

2. Mud solids in feed was very near the optimum (5,0 - 5,5 %) .

3. Bagacillo quality which is important for obtaining the required

filtering medium is considered good.
SMRI TECHNICAL REPORT NO. 1141 .5

4. Temperature of feed and wash water was in the good range.

5. Low and drying vacuum which are closely associated with pol
in cake and retention were within the recommended range.

6. The filters were operated at a reasonable drum speed.

7. Retention was found to be around 8 3 - 85 %.

8. Mud solids in filtrate could be a little lower but were not

abnormally high.

9. No disproportionate purity drop from clear juice to filtrate

has been found. After addition of lime to muds it was normal.

10. Thickness of cake averaged 4 - 5 mm.

The following unsatisfactory points should however be corrected in

order to run the filters with optimum efficiency:

1. Fibre to mud ratio in feed is considered to be very high. This

has certainly a detrimental effect on cake washing as well as
retention. Fibre to mud ratio in cake is consequently increased
and can only bring a higher production of cake for the same
mud output.

2. The distribution of wash water on cake was uneven (Filter 1 ) .

3. Insufficient wash water was applied on cake. Pol loss in cake

was accordingly higher than expected. A relatively high pol
mud ratio is apparent. (This is demonstrated by the calculated
dilution from feed to filtrate (15-18 % ) .

4. pH of filtrates is too low. Better floes in feed and porosity

in cake could be obtained at reasonable pH range of 7,8 -
8,0.

REC OMMENDATIONS

Because of the relatively high content of bagacillo found in primary

mud it is recommended to reduce the* amount of bagacillo added to
filter feed.

The fibre mud solids ratio in feed averages 0,649. It should be kept
around 0,400 which has proved to be the optimum at a relatively low
drum speed.

Additional banks of sprays or drip pipes should be installed on

filter No. 1.

More wash water could also be applied on cake if the filtrate pumps
were of adequate capacity.
SMRI TECHNICAL REPORT NO. 1141 .6

The amount of water required for a good washing at 160 tch will
3
be 15-16 m / h r or about 16 tons/hr.
3
Pump capacity required for filtrates will be m / h r at the above
dilution. It is also strongly recommended that lime be added to mud
before the feed tank. The use of lime saccharate will be preferable
because of better reaction time, better floe and no possible
clogging of screens.

ACKNOWLEDGEMENTS

The assistance received from Mr. P. de Robillard and the laboratory

staff of Illovo sugar mill during this work is gratefully acknowledged.

PH/DM
SMRI
23/6/78