REVIEW ON DESIGN AND OPERATION OF CLARIFIERS IN

THE SOUTH AFRICAN SUGAR INDUSTRY FOR THE PERIOD

1975 TO 1995
lLMSA JULLIENNE AND 2G MONTOCCHIO
'Illovo Sugar Ltd, PO Box 194, Durban, 4000
"Iongaat-Huiett Sugar Ltd; PO Feliaon, 3875

Abstract on improvements due to additional gas venting at each com-

The major trends in clarifier design and operation over the partment.
past 20 years are reviewed. No important breakthrough has
been achieved despite changes in juice quality due to diffu- Table 1
sion and deterioration in cane quality. The main change has Number of compartments in clarifiers
been the advent of the trayless SRI-type clarifier which now
represents sixty per cent of the installed units. Number of compartments
Type Ratio settling/thickening
Keywords: clarifying, diffusers, design, SRI Juice Mud

Dorr/Graver/Bach 4 1 4
Introduction RapiDorr 4 2 2
Dorr 444/BMA 4 4 1
At the outset it is worth noting that there has been no major SRI 1 1 1
breakthrough in the industrial process used for clarifying
cane juice in the last twenty years. Clarification is still carried
out by liming and settling in sub siders at boiling point and the In the clarifiers with one mud boot per settling tray, the
operating technique is basically the same. It is remarkable SRI-type unit has received overwhelming preference over the
that this static situation was maintained despite of three major Dorr444. For example, the last 17 clarifiers installed locally
changes to the local conditions, namely, the advance of diffu- since the early 1980s, except one at Komati Mill, have all
sion, the decline in cane quality and the many drought years. been of the SRI type which now represents 60% of the total
Of the three, the drastic reduction in the amount of suspended number in use. It must be noted that, in the more recent instal-
solids in mixed juice due to the progressive replacement of lations, the conventional SRI model has been slightly modi-
milling trains by diffusers is certainly the most striking. On a fied by reducing the size of the underflow region and mud
quantitative basis the diffuser does more of the clarification boot (lpersonal communication) to accommodate the very
duty than the clarifier itself. For example, in the 1995 season, low mud volumes encountered locally and reduce the reten-
the average solids content of diffuser mixed juice was 0,2% tion time of the muds, which is otherwise excessive.
against 1,0% for the mills, equivalent to a 80% removal of
One of the latest additions to the local clarifier population
solids in the diffuser (Anon., 1996).
is at Komati and was installed in 1994. It is a new design in
It is arguably a poor reflection on the local technologists the sugar industry although its concept has been used for
that this particular set of conditions has not led to a clarifier many years in water and effluent treatment. It is trayless, with
of entire South African design. All the clarifiers presently in- a central feedwell and eight radial launders fitted with identi-
stalled are of foreign design, albeit with certain modifications cal diameter holes through which the clarified juice exits. It is
to suit the local conditions. felt that there are not enough performance data available, at
this stage, to express an opinion on its performance. As with
the SRI model the mud region of this clarifier is being re-
Clarifier design duced presently to accommodate the very low mud volume.
Trends in clarifier types and numbers The almost universal adoption of the SRI-type clarifier by
Over the many decades since juice clarification by settling the industry is based on many reasons, not the least of which
has been carried out on a continuous mode, the main evolu- are the markedly cheaper cost of equipment, installation and
tion in clarifier design has been in the direction of a decrease maintenance, and the simplicity of operation. In general its
in the number of compartments for settling, relative to mud clarification performance is considered equal to the multi-
thickening (Hugot, 1986) as shown in Table 1. This trend has tray units except that it is known to be sensitive to juice flow
culminated in the SRI design with only one settling compart- variations and it has been found almost essential to have auto-
ment and one mud boot. mated control of the juice flow into the clarifier for optimum
The trend in the types of clarifier installed in the industry operation. Although it has the reputation of using more floc-
over the years has followed the same pattern with very few cui ant this is not reflected in the industry figures for the 1995
Dorr, Graver, Bach and RapiDorr clarifiers still being in op- season, with 3,3 ppm on mixed juice for both the SRI-type
eration. Of these, most, if not all, have been modified to and multi-tray clarifiers (Anon., 1996).
multi-point juice off-takes for more uniform flow pattern and The change in the number and type of clarifiers operating
some have been baffled to reduce turbulence at the juice inlet in South Africa (Allan, 1973; Reid 1992) over the past 20
to each compartment. The modifications have generally led years is shown in Table 2.
to much reduced retention times as reported by Scott (1988)
and van Duyker and Tosio (1986). Scott (1988) also reports 1 Mr S North-Coombes, formerly of Smith tech Pty Ltd, Durban

Proc S Afr Sug Technol Ass (1996) 70 277

Review on Design and Operation of Clarifiers for the period 1975 to 1995 LMSA Jullienne and G Montocchio

Table 2 Clarifier operation

Type and number of clarifiers in the sugar industry As already mentioned there has been little change over the
past 20 years in the way the clarifiers are operated. The only
1975 1995
noteworthy industrial trends are the running of clarifiers with
Ratio 4 Clarifiers muds at lower levels and thinner concentrations than in the
(Dorr/Graver/Bach) 21 1 past and the shift from mechanically stirred tanks to in-line
static mixers for lime mixing.
Ratio 2 Clarifiers
(RapiDorr) 25 7 Liming
Ratio 1 Clarifiers " All the mills, except one, use a straight hot liming system
for their clarification. The exception is Gledhow mill which,
Dorr 444 2 1
SRI-type 0 16 at times of high turbidity and colour, practices saccharate
BMA 1 1 liming. Saccharate liming has been tried at different places
Envirotech 0 1 (North Coombes et al., 1981 and Scott, 1988) but has never
been adopted" the opinion being that the process difficulties
Total 49 27 associated with it outweigh the marginal benefits. The quan-
tity of lime used is about 0,6 kg CaO per ton cane and does
not vary much between mills and seasons. The addition of
In 1975 there were 20 mills and 49 clarifiers in the indus- lime is automated by a pH controller and the pH of the limed
try, corresponding to an average of 2,5 units per mill, which juice (range of 7,3 - 8,1 measured at room temperature) is set
has reduced to 1,7 in 1995 (27 clarifiers in 16 mills). In this to obtain a clear juice at 6,9 - 7,l.
connection it is interesting to note that in the last three new Liming is generally effected in-line, upstream of a static
installations, two (Komati and Eston) installed only one unit mixer and before flashing. Five mills are still liming after the
each whereas Felixton installed three units. Within the indus- flash tank using a stirred tank.
try eight mills (50%) operate with only one clarifier, the other The concentration of the lime slurry varies between 4 and
half having two or three units. 16 degrees Beaume, the average being around 10. In general,
With the mills running more than one unit, the proper split- the milk of lime is prepared with hot condensate.
ting of the juice out of the flash tank into the clarifiers, often
of different capacity, is an important operational aspect. The Phosphate addition
most commonly used system consists of a distribution box About half the mills monitor the phosphate content of their
with weirs. mixed juice regularly and add phosphoric acid when the P20 S
level drops below 150 ppm in order to bring the concentration
Capacity of clarifiers back to about 200 ppm.
In the past the capacity of a subsider was usually expressed
in terms of the settling area of the trays. However, it is con- Flocculant usage
sidered more practical to compare capacity trends based on The consumption of flocculant in the industry has not
subsider volumes, for the simple reason that the most widely changed much over the past 20 years, considering the advent
used SRI-type clarifiers are trayless. As can be expected, with of the short retention time clarifiers during that period. The
the progressive replacement of old units with the short reten- industry average was 2,7 and 3,4 ppm on mixed juice in 1975
tion SRI-type, the installed volume of clarifiers in the indus- and 1995 respectively. Contrary to 1975 when the popular
try has decreased considerably in spite of a marked increase flocculants were Separan 273 and Talosep A3, those most
in crush rate, as shown in Table 3. widely used now are Magnofloc D and LT 27, Floccotan
R300, Primco A115 and Rheofloc 430.
Table 3 A few years ago there was a shift in the industry to liquid
Installed volume of clarifiers and crush rate
flocculants which, for various reasons, was discontinued and
all the mills are now back to powder form.
1975 1995 Under normal conditions these flocculants are used on their
own although most mills prefer a cocktail of two or even three
Industry overall TCH 3500 4640 flocculants when handling refractory juices. Two thirds of the
Total clarifier volume (m') 11700 7200
mills use cold potable water for flocculant preparation, the
Capacity in m3 per TCH 3,3 1,6
remainder using condensate cooled to about 50°C. Most in-
stallations allow four to six hours retention time for hydroly-
sis and the dosing concentration is normally 0,05%. The point
For the six mills using only SRI-type units the average in- of application of the flocculant to the juice is invariably at a
stalled capacity is 1,1 m3/TCH (range 0,9 to 1,2 m3/TCH) spot of turbulence just before entering the clarifier.
while it is 2,0 (range of 1,4 to 2,3) for the eight mills operat-
ing multi-tray units only. Juice quality and monitoring
Based on the above figures for installed capacities the aver- With the availability of two distinct types of clarifiers in
age nominal retention times in the clarifiers are 1,5 h for the industry it is interesting to make a comparison of the trend
multi-tray clarifiers and 0,8 h for SRI-type clarifiers. The in purity and pH of the juice from the multi-tray and short
largest units installed in the industry are three multi-tray retention time clarifiers. Based on the reported data for 1995
clarifiers of 11,0 m diameter at Malelane mill (RapiDorr), (Anon., 1996 and 'personal communication), there is no
Darnall mill (BMA) and Umzimkulu mill (Dorr444) and an marked difference in both parameters between the two types,
SRI-type of 10,7 m diameter at Eston mill. as shown in Table 4.

278 Proc S Afr Sug Technol Ass (1996) 70

LMSA Jullienne and G Montocchio Review on Design and Operation of Clarifiers for the period 1975 to 1995

Table 4 Precautionary measures for long stoppages

Purity and pH changes in clarifiers -1995 season Of the 16 mills in the industry, 10 increase the limed juice
pH to 8,0 - 8,5 for about two hours before the stoppage. Four
Type
Purity drop of the 10 will also reduce the juice temperature to just below
Short retention Multi-tray boiling point. Six mills dose biocide at rates of 40-150 ppm
half an hour before stopping. Three mills empty all the mud
Clear - mixed juice -0,40 -0,61 from their clarifier/s over the stoppage and one mill liquidates
Clear J - filtrate 1,44 1,58 its clarifier every week.
pH drop
Conclusions
Limed - clear juice 0,6 0,7 In conclusion, the main trends in clarifier design and op-
Clear juice - filtrate 0,3 0,3 eration in the past 20 years can be summarised as follows:
• the marked reduction in the solids content of mixed juice
due to diffusion.
Despite of the recent development of a juice turbidity me- • the advent and popularity of the trayless SRI-type rapid
ter by the Sugar Milling Research Institute (Stone, 1994) clarifier and the progressive elimination of the multi-tray
which is reported to be working successfully on trial at two units.
mills, the clear juice clarity is still assessed visually in the
industry by means of the clarity wedge (Allied Colloid Pat- • the lower installed volume of clarifiers and shorter nominal
ent), with readings in the range 8-14 being normally retention times.
achieved. • the mixing of lime by means of in-line static mixers rather
Fifty per cent of the mills measure the clear juice colour by than mechanically stirred tanks.
spectrophotometer. The normal values in 1995 were in the • the operation of clarifiers at low mud levels and the pro-
range of 20 000 to 26 000 ICUMSA units. duction of thin muds.
It is worth noting that the filtration of clear juice is prac-
tised, albeit crudely, at only one of the sixteen mills, by REFERENCES
means of static screens. Allan, GN (1973). Developments in South African sugar factory plant installations
1969-1973. Proc S Afr Sug Technol Ass 47: 83-89.
Anon. (1996). Monthly summary of laboratory reports for Southern African sugar
factories 1995/1996 Season (Final 1996). Sugar Milling Research Institute, 8 pp.
Mud handling Hugot, E (1986). Handbook of CaneSugarEngineering: Elsevier, Amsterdam. 872
pp.
Since the mid-1980s the trend in the industry has been to Lionnet, GRE (1984). Mud conditioning for good filter operation. Proc S Afr Sug
operate the clarifiers at much lower mud levels and at a solids Technol Ass 58: 39-41.
content below 4%. This change in practice followed the rec- North-Coombes S, Taylor, K and Koster, KC (1981). The practical development
ommendations of the Sugar Milling Research Institute and application of saccharate liming at the Pongola sugar mill. Proc S Afr Sug
TechnolAss 55: 71-74.
(Lionnet, 1984) whose objectives were to reduce the retention Reid, MJ (1992). South African sugar factory plant installations 1992. Communica-
time of the clarifier mud and improve its filterability. In most tions from the Sugar Milling Research Institute No. 156: 15.
of the mills the mud extraction from clarifiers is done by Scott, RP (1988). Modification to and experiences with RAPIDORR clarifiers in-
gravity while the others use pumps, mainly of the diaphragm cluding saccharate liming at Amatikulu. Proc S Afr Sug TechnolAss 62: 32-35.
Stone, VC (1994). The development of an instrument to measure the optical density
type. In general the rate of mud extraction is controlled by an of juice from a clarifier. Proc S Afr Sug Technol Ass 68: 83-86.
on-off automatic system based on the mud level in the mud van Duyker, Wand Tosio, C (1986). Modifications to and uprating of clarification
mixer feeding the filters. station at Simunye. Proc S Afr Sug Technol Ass 60: 40-42.

Proc S Afr Sug TechnolAss (1996) 70 279