The mud recycling system set up to handle the mud stream and sets out the problems encountered in getting this system running and the alterations planned to enhance the system. The paper also looks at the diffuser losses for 2002 season compared with the combined mud and bagasse losses of previous seasons.

The bagasse bed in the diffuser process tends to trap sand and mud brought into the system more than it occurs in the milling train. Thus, another option of separating the mud solids from the juice in the clarifier underflow is recycling it back into the diffuser, and using the bed of prepared cane to trap the mud solids. The trapped mud solids are passed onto the boilers with the bagasse and ultimately removed with the boiler ash. The need for the rotary vacuum filter and its auxiliary equipment for mud filtration are eliminated. A number of South African mills have already eliminated the filter station in this way and this has provided the basis for this project.

There are many advantages of mud recycling system. With the removal of the filter station, there will be savings in maintenance and operating costs and supervision of this area is simplified considerably. Other savings result from the elimination of mill mud cartage and the need to maintain mill mud dump sites and the access roads.

The disadvantages of this processing option are the effects on the diffuser and the boiler tubes. The South African experience has been based on diffusers with two sets of lifting screws with the mud stream entering the diffuser at the first set of screws. This situation is less disruptive on the diffuser brix profile. The De Smet diffuser at Lautoka Mill has only one set of lifting screws, which is situated at the back end, where the press water from the two dewatering mills enters the system. This is where the clarifier underfeed is currently being added to the diffuser. Thus, high brix material from the clarifier is entering the diffuser in a low brix area, which may increase losses. The cane bed is sensitive to anything that causes the percolation rate of juice through the bed to slow down and may ultimately cause flooding.

Mud recycle system for eliminate solids from bagasse

Figure 1 shows the flow diagram of the mud recycle system. The diffuser at Lautoka Mill is a ‘De Smet’ design diffuser, rated at 360 tonnes cane per hour. The draft juice is pumped through the primary and secondary juice heaters into the SRI clarifier. The mud is withdrawn from the bottom of the clarifier into the clarifier underflow tank.

Fig. 1—Mud recycle system crushing season.

The clarifier underflow is pumped from the underflow tank into the press water circuit, which enters the diffuser at the lifting screws. The clarifier underflow rate to the diffuser is controlled via a recycle line back to the underflow tank. The flow rate set point is set by the operator and depends
on the mud level within the clarifier and operations at the diffuser. Some interlocks are also provided from the diffuser. If the diffuser conveyor stops, the mud system goes into a recirculation mode and pumping to the diffuser stops. The totalised amount of underflow transferred to the diffuser is recorded each shift. Underflow is sampled and analysed on a daily basis. These results are used to provide feedback on the amount of mud solids being transferred to the diffuser.

Initial operational problems

The mud transfer system has operated well but problems occur if the underflow becomes too heavy. A small amount of water is then added to the mud tank to control the consistency.

In the initial trials, the mud stream entered the diffuser at the stage 3 juice collection trough and distributed across the diffuser bed via its juice recirculating pump, about six metres or so before the lifting screws. It was soon obvious this was causing flooding. The entry point was therefore changed to the press water tank, where the mud stream mixed with the press water and was distributed across the bed just before the lifting screws. However, intermittent flooding was still occurring due to the following reasons:

  • Low bed height, due to new operators becoming familiar with the new automated operation of the diffuser.
  • Lifting screws had been shortened in previous seasons.
  • The bed was unevenly shaped. It was humped in the middle due to poor distribution of cane across the diffuser and the juice tended to run off towards the sides.
  • Poor distribution of press water and underflow mixture across the width of the bed.
  • Overuse of polyelectrolyte.

Most of these problems have been rectified and, visually, positive results obtained. However, it is still evident that the addition of mud solids in the back end area is affecting the percolation rates. One obvious problem is that not as much imbibition water can be added to this area as in previous years, and this is affecting extraction to some extent.

The addition of high brix material to the back end of the diffuser is obviously not the best practice, since it distorts the brix gradient and ultimately the overall extraction. Figure 2 shows the brix gradient for a typical shift during week #6 compared with a typical shift for the 2001 crushing season before mud recycling was practised. Typically, stage 1 brix should be around 2.0 and it should rise progressively across the diffuser stages to around 15.0 for draft juice.


Figure – 2 Brix gradient along the length of the diffuser


Table 1 shows bagasse and mud sugar loss results for this season and the previous five seasons. Mud and bagasse pol loss have been calculated from analysis of mud and bagasse every shift during operations. We have looked at the average losses measured as pol loss % pol in cane, over the five years prior to mud recycling change and compared this with figures obtained from 2002 season. We have also removed the effect of the conventional milling train in previous seasons, so that direct comparison can be made regarding pol loss in bagasse from diffuser only. The final bagasse loss for this year is higher when compared with the combined bagasse and mud losses for the previous years. This result includes the effect of major changes to the cane extraction operation for 2002 season, which have taken some time to settle down. The changes that took place were:

  • Removal of the conventional milling train and modifications carried out to enable diffuser to take the full rate.
  • Automation of the feeding and diffuser stations.
  • Installation of a new final two roller dewatering mill (C2).
    These changes brought about the following effects:
  • Higher rate through the diffuser. As a result, disruptions to continuous crushing were encountered throughout the season as C1 dewatering mill is under capacity.
  • Disruptive operations during the early part of the season due to automation adjustments.
  • Problems associated with the performance of the new two-roller mill.

Table 1—Pol loss data for the 2002 season and the previous five years for comparison

Year final bagasse loss(diffuser) mud loss mud + bagasse (diffuser) loss
1997 4.20 0.36 4.56
1998 2.89 0.42 3.31
1999 3.10 0.30 3.40
2000 3.91 0.48 4.39
2001 3.84 0.40 4.24
Average 3.59 0.39 3.98
2002 4.56 0.00 4.56

One of the potentially negative impacts of mud recycling is the continued pumping of juice to the diffuser when the mud level in the clarifier is low. Recycle of juice would result in a low mud solids % clarifier underflow but Figure 3 shows that acceptable mud solids levels have been maintained. This is largely due to high level of operator involvement. A continuous check on the consistency of the mud withdrawn from the clarifier is being done. A means of automatically maintaining good underflow consistency would be an advantage, as it would simplify supervision in this area. Eventually, manning can be reduced to have a single operator for the filter and evaporator stations.

Fig. 3—Trend of weekly average mud solids % clarifier underflow for weeks 14 to 30.

As mentioned previously, the overuse of polyelectrolyte has been cited as a contributor to bed blinding. A concerted effort was made to reduce the polyelectrolyte usage. The polyelectrolyte usage for Lautoka Mill has been traditionally high. Table 2 shows that a significant reduction was achieved. Although no definitive proof exists, it was thought that this assisted in reducing bed blinding.

Table 2—Polyelectrolyte usage for the years 1997 to 2002.

year Crush(tonnes) Poly usage (kg) Poly usage (ppm)
1997 1160512 4495 3.9
1998 626348 1680 2.6
1999 1433730 4945 3.1
2000 1301698 5170 3.6
2001 907244 3450 3.4
2002 979016 1912 2.0

Figure 4 shows that the clarifier handled less mud solids and therefore the requirement for polyelectrolyte was also reduced. The mud solids % cane over the past six seasons is shown in Figure 4. The 2002 season figure is calculated from the underflow totaliser readings and the daily analysis while the previous seasons were calculated from the filter cake analysis and output. The reduction in mud loading is in line with the greater dirt retention characteristic of the diffuser. The seasons 1997–2001 all represent a combination of diffuser and milling train operation.

Fig. 4—Trend of mud solids % cane for the period 1997 to 2002.

Clarifier underflow consistency control

As discussed earlier, a high level of operator input is required for underflow withdrawal from the clarifier. Thus, a control measure is required to assist the operator. In some South African mills, a standpipe is used to measure the underflow consistency. It works on the principle that the time taken to fill the standpipe (at a constant head and temperature) is proportional to the underflow consistency as determined through the trials carried out at Maidstone Factory in South Africa (Jensen, 2001). A standpipe was included in the original Lautoka design of the withdrawal system but it did not provide the expected control. An alternative or an enhancement of its operation is currently being investigated.

Benefits of mud recycling system

The advantages of mud recycling to the diffuser for Lautoka Mill may be summarised as follows:

1. Reduced maintenance cost with the elimination of the rotary vacuum filters, bagacillo supply system and mud conveyor system.
2. Reduced physical loss of sucrose through the filter cake.
3. Reduction in condenser water usage as the filter vacuum system is eliminated.
4. Elimination of the need for a mud contractor and the costs associated with upgrading the mud dump site and the access road.

The financial benefit of mud recycling system advantages is shown in Table 3.

Table 3—Estimated cost savings in 2002 for mud recycling system.

Items Saving
Removal of mud contractor (5 year average) $ 35 120.00
Reduced maintenance cost (5 year average) $ 82 309.00
Savings from upgrade of mud dump site and access road (5 year average) $ 11 182.00
Reduced physical losses (calculated at 979,000 tonnes cane, 12.17% pol in cane and 81.9% boiling house recovery with 30% proceeds to F.S.C.) $ 34 823.00
TOTAL $ 163 434.00

An increase in bagasse loss has been experienced, although it is believed that this full loss cannot be attributed to the mud recycling. The operation of the diffuser and the dewatering mills was not ideal, particularly in the early part of the season, and some portion of the extra bagasse loss was due to this.

We are also aware that boiler maintenance may increase due to tube erosion and the amount will become evident over the next few seasons. Any increase in ash cartage would be absorbed by the current payment structure where the contractor is paid a standard sum per week and is independent of the amount of ash discharged.

We presume that ash tonnage will increase as a result of clarifier underflow recycle, but so far there is no indication of any increase in ash production during this season as compared to previous seasons. The costs incurred are shown in Table 4.

Table 4—Estimated cost incurred in 2002 from mud recycling.

Items Cost
Installation cost (instrumentation, mechanical and programming : one-off cost) $ 70,000.00
Decrease in sugar income through bagasse loss (calculated at 979,000 tonnes cane, 12.17% pol in cane and 81.9% boiling house recovery with 30% proceeds to F.S.C.) $ 86,789.00
TOTAL $156,789.00

We have obtained the loss of raw sugar as the difference in tonnes of pol converted to raw sugar using the 2002 season boiling house recovery and converted this to revenue at world market price of sugar. We have then determined the amount of money that the mill would have lost or gained from this (30%) since this is what will affect the mill economics and the mill had to put up the full amount to fund the project.

The benefits of mud recycling system have already been satisfactorily proven for diffuser factories. Some issues remain to be resolved for entirely satisfactory recycling at milling factories. Some ideas have been proposed for mitigating the problems of additional sand into boilers, for the problems caused by mud return to milling trains and for the loss of filter-cake for fields fertilisation.