StorkReview PDF

STORK
Palmoil Review
Development of Free Fatty Acids before Sterilization
Pressure and Influence of Fatty Splitting Enzymes
The effect of fat splitting enzymes (mainly lipoids) was discovered in 1901.
The presence of lipoids in, and their effect as, palm fruit is clearly demonstrated in the
table below, and published by Fickerday. It should be understood that, at the time of the
experiment mentioned here, it was usual to cut or peel the pericarp from the nuts
without sterilization prior to expressing the oil.
Oil after Peeling %FFA

A B
Test 1 - immediately 43.1 2.4
Test 4a - immediately 52.9 2.3
Test 4b - after 2 days 66.9
Test 4c - after 4 days 67.2
A = peeled fresh fruit

B = peel off after preliminary heating at 90oC or 10oC.
Conclusion
(i) Oil from unbruised fruit, which before further treatment was heated to a
temperature at which any enzymatic reaction is excluded, has a low FFA content,
often below 1%.
(ii) Oil recovered from fresh fruit, which was peeled and/or bruised, has a FFA
content of over 40%.
Thus this can only be explained by the presence of an active enzyme in the fruit, which
can start its fat splitting action if the walls of the oil bearing cells have been broken
through peeling and/or bruising.
1
Temperature at Which the Enzymatic Fat Splitting Action Can Be Stopped.
The table gives results that were collected by the researcher Barro, when he exposed
samples of palm fruit to ever increasing temperature before peeling.
Temperature oC FFA% of oil recovered

45 11.6
50 2.7
55 1.1
60 1.1
65 1.0
From these figures, which speak for themselves, it can be inferred that the enzymatic
activity cease at 55oC.
Speed of the Fat Splitting Process
It is difficult, not to say that it is impossible, to determine with reasonable accuracy the
rapidity of enzymatic hydrolysis. Several minutes will easily go into peeling a sample
of palm fruit, which, as will be shown, strongly affects the results of any analysis. The
research worker van Heurn has endeavoured to collect relevant figures and published
the results listed below:-
FFA% in oil recovered from bruised and pounded pericarp

As soon as possible 22%
After 5 minutes 33%
After 15 minutes 39%
After 30 minutes 40%
After 60 minutes 42.5%
Desasis has described the outcome of a hundred or so tests by which the FFA% was
determined are hours after the pericarp has been peeled and/or bruised.
The figure he found varied from 37% - 45% that the enzymatic fat splitting is a rapid
progress that is attains its highest value after approximately 1 hour, that probably ¾ of
this value is already attained after 15 minutes.
That any bruising of the fruit, even shortly after sterilization, will immediately lead to
an increase in acidity.
2
Role of Micro-Organisms
When bunches and detached fruit are left lying in the open air, they will often be found
to contract mould pretty soon in these places where the fruit has been damaged, as well
as at the base of the detached fruit.
Although many writers claim that mould influences acidity adversely, this is not quite
certain theoretically. Lavergne, Wilbaux & Desasis have described tests that have
shown an increase in acidity with undamaged and non-mouldy fruit and no such
increase with fruit that was bruised and mouldy.
Against the doubt that thus arisen there are 2 strong arguments, which provide a sound
basis for the opinion that microorganisms, may play a detrimental part.
Fickendey and others have pointed to the strong increase in acidity of the bruised
pericarp of sterilized fruit, if this is left in the open and contracts mould.
Wilbaux has described a fungus of the type Oospora which proved capable of raising
the FFA% of fresh fruit from 0.1 to 6.4 in 60hours.
Now, if we relate these positive observations with the fact that Loncin has succeeded in
hydrolysing palm oil with fungi of the same type, we may conclude that the
development of micro organisms on detached fruit and bunches lying stored in the open
field can cause an appreciated increase in FF content.
Remarks
Although through proper sterilization, it is possible to prevent a rise of the FFA content
resulting from enzymatic fat splitting during oil recovery process, a variety of factors
may be conducive to such an increase prior to sterilization.
1. The quality of bruised fruit and the degree of bruising; the number of detached
fruits; the type of harvest transport used and its duration; the period of storage
preceding the recovery process; all these are significant as far as the acidity is
concerned.
Many researchers have applied themselves to determining the percentage of free

fatty acid in detached fruit and fruit bunches.
Apart from the fat that they have clearly demonstrated the FFA% of oil from
detached fruit to be generally higher than the FFA% of oil from fruit in bunches.
Rough handling resulting from transport in the factory prior to sterilization will
cause bruising of the fruit and increase acidity.
3
Conclusion
• FFA will develop chiefly when the fruit is on its way from the tree to the
factory.
• Ripe and unbruised fruit will have a low FFA% (i.e. 0.5 to 0.8%)
• While in the bunch, the fruit is well-protected from bruising. Only a part of the
exterior fruit is exposed.
• Loose fruit are less effectively shielded from damage. If the y had fallen out of
the bunch and are allowed to remain in contact with the ground, they form an
ideal basis for the development of mould.
• Ripe fruit is softer and therefore has less resistance to rough handling than fruit
which is immature; it follows that although a low acidity may be due to careful
handling, it may also result from the presence of a large quantity of immature
(i.e. “hardy” fruit).
• An extension of harvesting intervals will lead to a partly ripe crop, which means
an increased amount of detached fruit and a higher acidity.
• An extension of collecting intervals will result in larger periods of temporary

storage in the plantation. This in turn will promote formation of mould on the
fruit and increase acidity.
Increase in the FFA Content of Palm Oil after Sterilization during Recovery and
Storage.
Spontaneous Autocatalytic Hydrolysis
Although the percentage of FFA in palm oil has long been known to increase in the
course of the recovery process and during storage, the real cause has remained a
mystery until few years ago. Only in 1952, Loncin has described the process,
spontaneous autocatalytic hydrolysis that is responsible for this increase in acidity.
Until then, enzymatic hydrolysis, brought about by micro-organisms was held to be the
root of the trouble and at first Loncin contested this view after his discovery.
Recent investigation have meanwhile brought to light that the latter process does not
after all play such a minor part as Loncin has assumed.
4
Loncin main conclusions were the following.
A. Presence of a certain quantity of moisture is a sire qua non for the increase in
acidity. Adequately dried palm oil is not liable to acidity.
Treatment Initial % %FFA

After keeping for 30 days at 60oC
Palm oil well dried 7.3 7.35
Non-dried 7.3 8.85
Palm oil + 20% moisture 7.3 10 - 60
B. The reactions influenced by the original FFA%, the temperature at which the
reactions takes place and the time.
K = 23/to (log a2 - log a1)
a1 = original FFA%
a2 = ultimate of FFA%
t = researcher time (in tens)
k = temperature co-eff
Temperature oC K
37 0.025
50 0.051
60 0.102 - 0.164
70 0.250 - 0.188
80 0.502
100 1.480
The moisture content of the oil is only significant as far as it renders the reaction
possible. If there is no more moisture, the reaction will cease.
In a report published by Vanneck it is stated, however, that higher moisture content has
favourable influence on the reaction, witness the under-mentioned analysis results:
Oil + 0.25% moisture Same oil + 5% of moisture

Initial FFA% 2.80 2.20
After 55days at 55oC 3.50% FFA 4.10% FFA
After 36 days at 75oC 4.60% FFA 9.65% FFA
Vanneck added a note that the moisture content of the oil (i.e. 0.25%) should have been
capable of producing an increase of about 3.5%, whereas the maximum actually found
was 1.8%. It may be due to the oil having dried in the course of storage at 55oC and
75oC, so that the reaction has been prematurely arrested.
5
Enzymatic Hydrolysis
According to Loncin, water and free fatty acids are the only substances noticeable
instrumental in acidifying of palm oil.
Vervloet on the contrary, is of the opinion that the dirt percentage of oil is a third factor
of importance in the process.
There is no need to be doubtful; of the possibility of an increase in acidity brought

about by micro-organisms for certain fungi and bacteria eg. Oospora Pseudomonas
Flurorescens & Geotrichum Condiduns have been isolated, which were found to
process a great fat splitting potency.
Enzymatic hydrolysis may occur side by side under the following circumstances;
• A temperature that is under 50oC
• The presence of a suitable basis (i.e. moisture and dirt)
• An infection by some outside source or agent.
Conclusions
• With the production methods now in use, the increase in acidity during - on
resulting from - the recovery process is but small. The comparatively short
retention in the various machines limits this increase as a rule to a few tenth %.
• FFA content may rise appreciably in the course of storage and transport,
particularly bulk transport in ships tanks, in which the oil is sometimes kept at
50oC or higher for a considerable time by means of steam coils.
• During recovery press enzymatic hydrolysis is not liable to occur, because
generally the temperatures used are so high as to prevent micro-organisms from
developing.
• It however are facts to observe the required cleanliness in the factory, cooled
down oil left in dirty interims storage tanks and/or pipelines during standstill
may increase in acidity.
• Other sources of contamination are found in the transfer of cooled oil from
transport tanks to storage tanks and vice-versa, if other equipment employed,
including that used for drawing samples, is in any way contaminated.
• Mixing dirty and moist oil having a high FFA% with oil that is clean and dry
and has low FFA content, will lead inevitably to a rise of the FFA content as a
result of the higher initial FFA content and the moisture added. Moreover, this
will create circumstances proportions to enzymatic hydrolysis.
• The moisture content of non-designated factory oil is usually too high to avoid
an increase in acidity. A moisture content of 0.25% will be enough to make the
acidity rise freely.
• The new STORK oil desiccators are capable of reducing the moisture content to
0.08 and even 0.04 so that autocatalytic fat splitting is out of the question.
6
• The knife cuts both ways, because it will at the same time deprive these micro-
organisms which may be held responsible for enzymatic hydrolysis of any
possibility to grow.
Sterilization
Purpose
1. Inactivation of the oil-splotting enzyme
2. Coagulation of the albuminous substances, which will cause trouble in the

clarification room
3. Hydrolysis of the mucilaginous matter.
4. Loosening the Fruit in the Bunch
In the sterilizer, the harvested bunch is prepared as it were, for the subsequent threshing
or stripping process.
Through the sterilization namely, a certain hydrolysis which occurs in nature as the
fruit ripens, is accelerated. The action can be compared with the hydrolysis of poly-
saccharides (starch cellulose) and their conversion into mono-saccharides. The place of
contact between the fruit and its stem is hydrolyzed by the steam and the reduction of
the poly-saccharide molecules is attended with a better solubility of the substances in
this place of contact. As a result, the fruit is loosened in the bunch (as with the natural
ripening process) and can be released.
To enable the hydrolysis to take place, the fruit should be brought into contact with the
steam. The higher the temperature and the longer the sterilization period, the more
readily the fruit can be stripped from its bunch afterwards.
Figures below show the relationship between the loss of fruit in empty bunches the
sterilization time and the temperature. The conclusions here represented are based on a
large-scale factory tests which have covered a considerable period of time. The scale
for the ordinated has not been indicated, so that no exact losses can be derived from the
graph, which may only intended to give a general idea of comparative values. For
example, it demonstrates that sterilization at 102oC for 80mins, gives the same
proportion of non-released fruits as sterilization at 115oC for 100mins.
7
CHART
We now come to the question of bunch stripping itself.
1. In the bunch stripper or thresher, varying quantities of bunch trash containing

calyx fragments are knocked out along with the fruit.
2. Inadequately stripped bunches. As longs as it is a matter of one or two stray

fruits, the bunches may be said to be empty for practical purposes, and the
losses of oil sustained on account of this will not exceed 0.04& to 0.05% in
well-run mills.
There may also be bunches however, which though stripped, contain appreciably more
than just a couple of undetached fruits and such bunches should be considered the result
of some abnormality either in the process (faulty action of the stripping machine,
inadequate sterilization) or in the bunch itself.
(i) Faulty Action of the Stripper or Thresher
(a) Imperfections in the design of the stripper and/or from incorrect operations
(uneven charging or overloading). The fruit left in the stripped bunch I soft and
will fall out immediately the bunch id dropped or shaken.
(ii) Abnormal bunches (knot-head bunches)

These are hard to deal with, because they cannot be treated very well
mechanically. Fruit in such bunches is reasonably soft but remains firmly attached
to the stem. In the stripping process, such bunches acquire a remarkably hairy and
fibrous appearance. Even after re-sterilization it will as often as not be impossible
to detach the fruit in the stripper as it has to be cut loose with a knife if any oil is
to be recovered from the fruit.
8
N.B, Any bunch which has been found impossible to strip properly of its fruit and
which therefore, should be termed a knot-head bunch should have no or hardly any fruit
at the top and a large portion of the fruit at the base.
Only, if a bunch is reaped in a completely immature state will it become an all-out

knot-head but of course this will rarely occur on a well-run estate.
(iii) Inadequate sterilization

Too low a sterilizing temperature may be responsible for improper stripping.
5. Preparation of the Fruit for Further Treatment in the Extraction Plant.
In the digester, pericarp should be loosened completely from the nuts. Through
sterilization the structure of the flesh of the palm fruit is so far weakened that this
separation can in effect be brought about during digestion.
It is generally known that bunches will shrink when subject to sterilization, so that after
the process, they are about 2.3 of their original size. Figures below shows that this
shrinkage is well on its way to completion after a period of 10mons.
Requirements
Sterilization and the quality of palm oil
Next to a low FFA content as good bleachability is among the requirements high-grade
palm oil is expected to meet. In other words, the bleachability of the oil may not be
impaired by the sterilization. The state at which the harvested fruit reaches the factory
is of paramount importance with regards to the bleaching properties of the oil. It has
been found that the palm oil obtained from over-ripe (loose) fruit is comparatively hard
to bleach and that the presence of an excessive proportion over-ripe fruit may have a
most adverse influence upon final product, (of course - this applicable to rotten fruit in
an even stronger measure). That is why fruit should be prevented from lying about on
the floor or a platform in the sterilizing station or elsewhere for a day or even part of a
day.
Furthermore, good bleachability necessitates a proper air-release prior to sterilization,

the shortest possible sterilization time and the lowest possible process temperature.
Hence a compromise between requirement for low FFB and for good bleachability had
to be reached.
9
Time and temperature limits of the sterilization process
Minimum Limit
For want of anything better, the only criteria we can apply to ascertain whether the
sterilization has been carried out correctly is the proportion of inadequately sterilized
bunches. If the temperature used has been too low and/or if the sterilizing time has been
too short, this proportion will rise considerably. There are a couple of factors governing
the minimum time and temperature limit i.e.,
1, Size of the bunches
2. Degree of ripeness of the crop
The Size of Bunches
The smaller the bunches, the more rapidly they will be sterilized, because they expose a
relatively large surface to the action of the steam whilst the heat has only to travel a
comparatively short distance. This rule has been tested on bunches from palm s of
varying ages, whereby the youngest palm trees were supposed to yields the smaller
bunches. The outcome was as follows:-
Age of Palm tree Time of Steam press inadequately

Yrs Sterilization kg/cm2 sterilized bunches
6 30 min. 2 29
6 30 min. 2 13
5 30 min. 2 11
4 30 min. 2 2
Degree of Ripeness of the Crop
If the bunches harvested are so ripe that much of the fruit is already loose on its stem;
the sterilizing cycle can be kept relatively short. If is the average degree of ripeness is
lower, so that the bunches are more compact, the sterilizing cycle should be prolong a
little.
Definition of under-ripe, over-ripe and ripe.
Time of which the bunches are harvested is always the result of a financial-economic
compromise made between the oil yield on one hand and the expenditure involved in
the harvesting procedure on the other.
Chief points controlling harvesting standards
(a) Fruits in a given bunch will not ripen simultaneously. At the apex fruit may
already begin to drop out, there will still be a quantity of under-ripe and even
under-developed fruit at the base.
10
(b) The fruit has its maximum oil content the moment it will become loose on its
bunch. By ‘loose’ we understand the condition which permits the fruit at the
outside of the bunch to be detached easily and by a slight press of the thumb. If a
bunch containing such fruit is harvested, the mere fact of its falling to the ground
will release a large number of fruits. A few days before this moment arrives, it
will be found possible to already break a few fruits from the bunch with a little
exertion, but in this stage there can of course be no question of loose fruits.
Data: Average content by weight of pericarp in freshly Ø ripe fruit. Which is loose in
its bunch is 50%.
5 days before this moment is reached - 43%
10 days before this moment is reached - 34%
(c) As a general rule, any increase in the percentage of detached fruit will entail a
rise of the FFA content.
(d) An increase of percentage of detached fruit is likely to impair the bleaching

properties of palm oil.
(e) Oil recovered from under-ripe fruit has a higher M.P. than oil from freshly-ripe
fruit and this sometimes is sometimes an advantage in the processing of
margarine.
(f) The fruit that has fallen to the ground is conveyed to the mill with the harvested
bunches, so far as it can be found under the trees and so far as harvesters are
prepared to collect it. There will be a slight loss, because a certain proportion of
the detached fruit lies hidden from view amid the foliage of the ground cover in
the plantation or is eaten by animals.
(g) For reasons of economy, the harvesters cannot return everyday to a given tree
whether a bunch has the correct degree of ripeness. It may frequently happen that
a harvester comes across a bunch which is slightly under-ripe but may be over-
ripe if left till the next round.
It will be clear that the crop arriving at the mill every degree of ripeness between over-
ripe and under-ripe will be represented. Now, in a well organized palm oil mill, it is a
common practice to keep a close check on the actual situation by singling out part of
the crop (i.e. a few lorries) and subjecting it to an analysis, the bunches being sorted
into different classes.
11
This is what such an analyst might look like:-
Class 00 = no detached fruit = a%

Class 0 = From 0 to 10% detached fruit = b%
Class 1 = From 10 to 25% detached fruit = c%
Class 2 = From 25 to 50% detached fruit = d%
Class 3 = From 50 to 75% detached fruit = e%
Class 4 = From 75 to 100% detached fruit = f%
Class 5 = part of interior fruit also detached = g%
Detached fruit = h%
Total 100 %
Considering the circumstance that by the fall of the bunches from the trees and as a
result of the bunch transport to the mill, a large quantity of fruit will be released before
arrival, the categories 00-0 and perhaps also category 1 may be classified as under-ripe.
The bunches from class 5 are over-ripe, while class 4 is a rather doubtful case.
To harvest a crop that is on the under-ripe side implies a lower oil content by weight of
the pericarp, a lower FFA%, better bleachability, a higher M.P. whereas with a largely
over-ripe crop, it is just the other way round.
A harvest-comprising bunches of all year-classes, which for the most part come under
the category 2 up to and including 4, can be sterilized with little or no risk of
inadequate results, if the process cycle is built up as follows:
De-aeration : 5 min
Working-up to full press : 20 min
Sterilizing : 50 - 60 min
Maximum Limit
This is controlled by a tendency of the kernel to discolour as a result of over-sterilizing.
To assess whether there has been any discolouration, one should take a kernel in its
final condition, i.e. dried and ready for dispatch, cut it through and examine the section.
This is because no signs of discolouration will ever be show at the outside.
Types of kernel oil that are obtained from palm kernels in various conditions.
1. Kernels from fruit that is sterilized for a very short period at the normal
temperature will yield oil at a slightly yellow colour although the kernels
themselves remain quite white
2. Oil recovered from normal kernels is of a light yellow colour and is easy to
bleach.
12
3. Oil recovered from kernels that have not been adequately dried is of a greyish-
yellow colour and is hard to bleach.
4. The oil recovered from light brown kernels is again light yellow and can generally
be bleached without trouble.
5. Oil recovered from kernels that have been stored wet and/or are slightly rotten is
dark red and hard to bleach.
6. The oil recovered from dark brown kernels is dark brown and cannot be bleached.
Time 100oC 110oC 120oC 130oC 140oC 150oC 160oC 170oC 180oC
15’ c c c c c c c c c
30’ c c c c c c Ø 9 }
45’ c c c c c Ø } } z
1.00’ c c c c Ø 9
1.15’ c c c c 9
1.30’ c c c c
1.45’ c c c c
2.00’ c c c
2.15’ c c c 9
2.30’ c c c }
c No visible discolouration
Slight discolouration
Ø marked discolouration
9 kernel turned to brown
} kernel turned to dark brown
z black
Effect of Sterilizing Temperature on the Oil Recovery Quotient
Use a high sterilizing temperature will improve the oil recovery quotient i.e. oil losses
in the fibre will be lower. It is obvious therefore there will be a tendency to increase the
sterilizing temperature. There is however a certain maximum limit which for reasons
explained in the preceding paragraph, it is essential to adhere to.
Reduction of Moisture Content in Fruit through Sterilization
Owing to the high temperature the fruit is exposed to during sterilization, part of the
moisture in it will evaporate spontaneously after the press is released. Loss of moisture
in this way amount to 10% of the weight of the fruit, assisting materially in the
extraction of the oil in the press system.
Optimum moisture content is 12% ± 2%.
13
Sterilizers
Horizontal Sterilizers
The vessel cylindrical steel pressure vessel, electrically welded and the diameter is
either 1700mm or 2000mm. The vessels are mounted on two or more stools according
to their length. The stool near the door is fitted rigidly, the others rest on rollers, in
order to allow free expansion and contraction of the shell under the influence of the
ever-changing temperature, it is exposed to. The sterilizers have a door which covers
the full diameters of the shell and which is suitably balanced from a swivel arm fitted
on top. This design makes it possible for the door to be turned to the side of the
sterilizer, so as to allow free access to the interior.
Air-Release
It happens sometimes that although the requirements with regards to sterilization period
and steam pressure have been fulfilled and although the bunches handled are of normal
average ripeness, one finds oneself confronted with an excessive proportion of
inadequately sterilized bunches. There is but one cause to which this can be attributed;
the temperature in the sterilizer must have been too low, i.e. it has not attained the value
corresponding to the pressure used. Such a difference finds its source in the presence of
air in the vessel.
If a sterilizer should not be provided with proper air-release (or if the air-release system
should not be used) the air would go on exerting an absolute pressure of 1 ATMOS.
Now if in such a condition, the vessel was to be connected to a steam supply line, in
which there is an absolute pressure of 3 ATMOS, this would raise the absolute pressure
in the sterilizer to 3ATMOS.
According to Dalton’s Law,
3 ATMOS pressure = Partial air-pressure (1 ATMOS) + Partial Steam Pressure 2 ATMOS
Hence temperature of steam instead of 132.9oC →119.6oC.
More important than its influence on the real; steam pressure is the circumstances that
air is very bad conductor of heat and impedes transference of heat to any fruit that
happens to be within an air pocket.
Furthermore, the oxygen in the air tends to affect adversely the bleachability of the oil.
All these points go far to show why to obtain satisfactory results, one should deaerate
the sterilizer as thoroughly as possible before proceeding to the sterilization proper.
In STORK sterilizers, the air outlets system is of a generous size and is fitted as low as
possible, whereas as the steam is admitted at the top at the highest possible point. If one
cares to supply steam at a moderate velocity to prevent any appreciable over-pressure
14
from arising in the vessel, the air will be displaced by the steam. In order to have
reasonable certainty of the dearation being carried out properly, one should continue the
procedure for a short time after steam has been noticed emerging from the air-outlet.
All in all, the dearation will take approximately 5 minutes. Only after that may one
open the steam supply more fully and increase the pressure to start the sterilization
proper.
Quantity of Steam Required For Sterilization Process
On a basis of figures obtained in practice, it maybe assured that in order to sterilize 1

ton of bunches, adequately, approximately 175kg of steam are needed, provided that
both the sterilizer and the pipelines are suitably lagged.
Corrosion and Erosion
Corrosion
The acid condensate is the main cause of the rather heavy wears which the sterilizers
are subject to. Occur in the part under the sterilizer bores. Corrosion activated by grains
of sand, which enter the sterilizer with the harvested bunches, are washed off by the
condensate and land on the bottom of the vessel together with fragments of vegetable
material, loose fruit etc... Consequently in order to prevent undue corrosion and pitting,
it is imperative that the sterilizers which for some reason or other are put out of service
for a certain time should be given a through cleaning.
A similar form of pitting may develop near the entrance to the horizontal sterilizers on a
level with the top of the sterilizer cages, for if these cages are loaded excessively or
carelessly, it may happen that part of the bunches project so far from the sides of the
cages as to be broken up by friction against the sterilizer shell. The pericarp and sand
rubbed on in these places will set up pitting, which in the end may become pretty
serious.
The horizontal sterilizers are fitted inside with liner plates, saving the purpose of
protecting the lower half of the shell from the action of the acid condensate. During
manufacture of the vessels, special care is bestowed on the gas-tightness of the welds
joining the liner plates to the shell, but in practice, it is not always possible to prevent
leaks from developing in the joints, through which condensate can penetrate between
liner plates and shell so that the protection is deprived of its efficiency. A regular and
thorough inspection of the interior of the sterilizers is therefore most essential if one
wishes to avoid a great deal of unpleasantness. Foe example, in the case the proper
maintenance of a horizontal sterilizer should be neglected, and leaking joints in the
liners plate left untended, the chances are that during the sterilizing process, a quantity
of the hot condensate will accumulate between the liner plates and the sterilizer shell.
15
When the shell is then blow down, a difference in pressure may arise between the
interior and the space between liner plates and shell, as a result of rapid evaporation
(flash). Now if the numerous small leaks in the joints are unable to eliminate this
pressure difference fast enough, the liner plate may be forced upwards jamming the
sterilizer bores, which it will then be extremely difficult to dislodge.
The acid condensate is also apt to find its way into the air outlet tubes where it will
cause a scale of rust to build up on the walls, which tends to obstruct proper air release
in quite a serious manner.
It is important therefore during routine overhauls never to skip the air release
arrangement. As a matter of course, it is the condensate discharges lines that suffer
most from the corrosive actions of the condensate. On this ground, they are as a rule
made of acid-resistant cast iron. Nevertheless, they will want replacement after a
certain period. There is no means of circumventing this necessity, nor is it possible to
evade a periodical renewal of liner plates.
Erosion
At the beginning of the sterilizing cycle, the sterilizer will consume an extremely high
quantity of steam, especially if no care is taken to make it attain full pressure gradually.
In that case, the steam boilers may as often as not prove unable to meet the excessive
demand for steam and pressure drop as well as priming are likely to occur. The steam
conducted to the sterilizer is then laden with myriads of tiny water drops. As a result of
the great difference between the pressure in the steam supply line and that in the cold
sterilizer, the rate of speed at which this steam will enter the sterilizer is about
equivalent to the critical velocity and is in the order of 500m/sec = 1800km/hr. The
droplets at which at this supersonic speed collide with the metal sterilizer shell (and/or
with the fruit), have an erosion effect which is comparable to that of a sand-blasting.
In the sterilizers manufactured by STORK, the steam inlet is designed in such a manner
as to keep the incoming steam from striking the fruit direct or at a very high speed, so
that erosion is rendered almost impossible.
For the sake of proper order, we would point out that in cases where the steam admitted
is wet and may cause erosion; there is no danger of kernel discolouration as discussed
in our preceding number, whereas conversely, no erosion by water droplets can occur if
the incoming steam is superheated.
In the early years, the recovery of palm oil gave great difficulties. Not only was there
considerable loss of oil in the fibre, but also the oily residue from the presses caused no
end of trouble in the further stages of the presses. It could not be treated effectively in
the depicarpers, with the result that the loss of nuts was greater than strictly necessary.
Moreover, an appreciable amount of oily fibre remained attached to the nuts, which
fibre formed a cake on the wall of the nut crackers so that they had to be put out of
service and cleaned very frequently. The fibre present in the cracked mixture fouled the
claybath in the kernel and shell separators, causing excessive loss of kernels. Further,
16
fibre particles were carried off with the kernels and lastly they spoiled the appearance
of the final product.
Not unnaturally, these poor results were initially inputted to some fault in the press
cage. Many experiments were carried out involving the perforations in the cage, the
nos. of press plates, conical bottom plates, the injection of water into the cages, etc, but
all the attempts to improve matters were to little practical avail. The cake pressure,
which was low originally, was greatly increased around 1930, attaining approximately
the pressure it has at the present day. Although this produces some amelioration, the
results continued to fall short of the expectations and oil losses from 18-20% a day fibre
were quite a common occurrence.
Little was known as regards to the operation and the influence of the digester. One did
know that the stirring arms of the digester should be long, otherwise there was the risk
of the fruit remaining indigested and of the mash to be discharged into the press cage
having too low a temperature, because with short arms a crust was apt to build up in the
digester wall, which impoverished the transference of the heat to the mash.
Further there was the accepted theory that a high temperature of the digested fruit mash
was conducive to a good pressing effect because it gave the crude oil a lower a lower
viscosity, but one was ignorant of the exact influence exerted by the temperature.
It was known that if digestion was continued a little beyond the established period the
results obtained were slightly better but under the circumstances prevailing in those
days, this comparatively small gain did not offer the drawback of the inevitably reduced
output and was found unacceptable. One groped, as if it were, in the dark, which in part
should be put down to lack of material for comparison: there being only a few digesters
in actual use, one know no better. It’s the number of pressing units (ie digesters plus
presses) increased, the situation changed. Not because newly installed units invariably
gave better service than there ones of earlier date.
On the contrary, sometimes, the efficiency of the new units was ever poorer than that of
the older ones, but it was now possible to start making comparisons, especially
inasmuch as certain conditions, i.e. the cake pressure used and the measurements of the
press cage remained largely the same.
The following test gave convincing proof of the digester being responsible for the
variations observed.
The stirred mash from digester A was transferred to the cage of press B and conversely,
the mash from digester B was forced into press A. In both digesters, the same type of
fruits was handled and the digesting cycles were carefully kept identical. The results
obtained were compared with those given by the original units (Digester A and Press A,
Digester B and Press B); as in the table below:
17
Fruit mash ex-digester A pressed in press cage A = 15.3% oil on dry fibre
Fruit mash ex-digester A pressed in press cage B = 14.9% oil on dry fibre
Fruit mash ex-digester B pressed in press cage B = 22.5% oil on dry fibre
Fruit mash ex-digester B pressed in press cage A = 22.0% oil on dry fibre
From this moment onwards the digester stood in the focus of attention and a
systematical enquiry was conducted into the effects of the various factors. One of the
first items on the programme was an investigation of the relationship of the pressing
time and the loss of the oil in the pericarp residue, and subsequently, the relationship
between the temperature of the digested mash and the loss of oil in the pericarp residue.
Figure 1 gives some particulars as regards to the outcome of the experiments conducted
to that end. It shows quite clearly the difference between pressing unit A and the unit B,
and the influence of the temperature of the fruit mash at the unit B under the
circumstances then obtaining. The positions of the lines that now came to be regarded
as desirables, is obvious:
(1) They should run as horizontally as possible, so that the rate of pressing ( i.e. the
capacity of the unit) has the least possible effect upon the loss of oil in the
pericarp residue.
(2) The horizontal part should start at an early point, in other words, the surviving
moment should be as near to the vertical axis as possible, so that the capacity of
the pressing unit can be increased.
The graph shows the curve of pressing unit B to have no particular point of transition
and no horizontal part at all, while the curve of the unit A is not flawless either, as
regards to the transition from horizontal to vertical.
The digestion of the fruit in the digester of unit B was extremely poor; 4-5% of the fruit
remained undigested i.e. the pericarp was still firm around the nuts and had hardly been
bruised. The discovery was made that the digester could not possibly function well,
because the contents had turn into a thick and pasty mass which was continually on the
boil. This hardly ever occurred in the pressing unit A. Therefore the digester B much
18
have too large a heating surface and/or the steam temperature (pressure) in the S.J. was
too high, which had the effect of the liquid in the digester beginning to boil and, having
been driven upwards, accumulating on the surface and acquiring a pasty consistency,
By a mere reduction in pressure to the point where boiling of the digester contends was
excluded, an appreciable improvement was brought about.
It being known that if the mesh was digester for a longer time (i.e. more intensively) the
result were much better, experiments were carried out in order to ascertain the
appropriate number of stirring arms. The figures below picture the outcome.
Sets of Temperature of % of oil

Stirring arms Fruit mash oC loss on dry fibre
3 83 16.5
3 89 15.2
4 88 13.2
5 90 13.7
The conclusion was drawn that with 4 sets of stirring arms, the optimum stirring is
obtained, the provision of a 5th set of arms giving no further improvement. The test
involving a perforated sliding door at the discharge, through which a quantity of liquor
was drained from the digester in a provisional fashion, did show a marked ulterior
amelioration.
This experiment confirm the view which had been gaining ground that in order to
obtain the maximum stirring efficiency, it was necessary to increase the fiction within
the pericarp mash as brought about by discharge of the crude oil released during the
stirring process, which oil acts so to speak as a lubricant.
The digester of the type B were now provided with a perforation corresponding to the
one used in the digester type A, which perforation for that matter functioned badly in
the latter, for it proved to be a problem apart from how to make the perforation give
adequate service.
At the same time therefore attempts were made at using the perforation in the proper
way. Soon the conditions necessary to achieve this aim were found and the quantity of
liquid discharged from the digester through the perforation grew from 29kgs/hr to as
much as 300kgs/hr.
The effect was formidable. Not only the liquid thus drawn off gave an increase in
capacity because it no longer require any space in the presses (incidentally, it also cause
a relative increase in the amount of solid matter in the press cage) refer Fig 2 but also
the pressing curve became much more favourable, being pronounced by the horizontal
and having its curving moment at the point where the mash had been under H.P. for
about 2mins. (refer curve C in Fig. J). This favourable curve remained when the
number of pressing cycles was raised, so that it appeared to be possible to increase the
overall output w.r.t. the units that had so far been employed. There was an ulterior and
19
by no means negligible plus point in the owing to the improved digestion, the water
content of the press cake was reduced, which facilitated treatment of the residue ejected
from the presses.
As a matter, of course trials were also taken to find out how far any further release of
crude oil from the digester would have a favourable effect on efficiency. For that
purpose, the perforation in the digester was so far perfected as to allow ever increasing
quantities of crude oil to be drained off. It turned out that although in this way the
output of the press units could be raised to a considerable extent; there was appreciable
further decrease in the oil losses in the pericarp residue.
As ever larger quantities of non-fat were discharged with crude oil, difficulties arose in
the clarifying plant (this point will be evaluated in the number of dealings with
clarification). Although it is common practice to combine the crude oil from the
digester with that from the press cage and to treat them both as a single entity, we must
point out that to all intents and purposes there is a difference between the 2 types of
liquor, especially as regards to the proportions of the constituents’ elements. The same
difference exists between oil from digesters adapted to discharge a large amount of
crude oil and oil from digesters not so arrange. The following figures give a general
idea of the differences in composition between crude oil ex-digester and crude oil ex-
presses, the oil coming from fruit of the Deli-Dura type (i.e. 17% of oil by weight of the
bunch).
Intensive release i.e. approximately 60% of the recoverable oil discharged through the
digester perforation;
Composition of the crude from: Digester Press

Oil 57% 49%
Water 32% 44%
Non-oily solids 11% 7%
Slight release is approximately 14% of the recoverable oil discharged through the
digester perforation
20
Composition of the crude from: Digester Press
Oil 86% 60%
Water 11% 35%
Non-oily solids 31% 5%
From these figures, the inference can be drawn that the composition of the crude oil
changes in proportions as the quantity drained from the digester is larger. Not only des
this intensive release result in comparatively higher water content, but also in a larger
proportion of solid non-fat, which fact is necessary to take account of as far as
clarification is concerned.
The method of intensive release of crude oil, by which a quantity of oil corresponding
to 15-20% of the weight of the bunches is drawn off through the perforations in the
digester has been adopted early in a few mills. The advantage of the higher output of
the pressing units has in these scattered cases been regarded as outweighing the
drawback of the mere complicated clarification.
Explanation Of Various Observations Taken
There are 2 factors which account for the observations described in the preceding
section
(I) Rupturing Of Oil Bearing Cells
By the stirring action of the digester, the walls of the oil bearing cells are ruptured, so
that the oil released spontaneously or can easily be forced out from the broken cells
during the extraction process following digestion. When examining oil cells from a
press cake through a microscope, we saw that they have given up practically all their oil
and that only a few droplets have remained behind on the cell walls. The situation is
comparable to the drops of the liquid adhering to the inside of a squeezed out plastic
sachet.
Now, if the crude oil liberated from the cells by the stirring action is not removed, it
will act as a lubricant as a result which the stirring arms will lose more and more of
their grip on the pericarp and a large number of oil cells will not be broken up. During
the pressing cycle, it is not possible for the oil to leave these cells unless the cell-walls
are made to collapse under the pressure exerted. As, however, the walls of an oil cell is
exceeding elastic and as there is but little difference between the pressures within the
cells and outside it, this collapse is not easily brought about and as a rule requires
extremely high cake pressures. Even then, it is improbable that all the cells are ruptured
simultaneously, an all likelihood the breaking up of a given cell will cause the walls of
the adjacent cells to be distorted to such an extent as to render possible their subsequent
collapse, which in turn affects the walls of the neighbouring cells and so on. The release
of the oil from the cells which the digester has failed to open up therefore necessitates a
gradual time clearing procedure, during which the press contents are under a H.P., so
that the liberated oil must force its way through the narrow capillary ducts in the cake.
21
In consequence, the oil content of the press cake will decrease at a very gradual rate.
The 3 curves below may be regarded as an illustration of this thesis.
Viscosity of the Crude Oil
In case the crude oil is not released from the digester, the stirring arms will agitate the
liquid for a considerable period and emulsions will be formed.
The crude oil then consists of:
1. Non-fat: fibre and cells remnants plus soluble components

2. Watery soluble
3. Watery soluble emulsified with oil
4. Oil emulsified with water soluble
5. Free oil
The emulsified sub 3 and 4 are invariably of a high viscosity and may on that ground
impede the flow of crude through the capillaries in the press cake.
This thesis too, is reflected in the curves above. Attempts have been made to reduce the
Viscosity of the crude oil by adding (hot) water to the digester contents, but they had
little or no success. The cause of the failure has in the nature of the emulsions that have
been formed.
If we take a small quantity of crude ex-digester and remove the oil from it by
centrifugal extraction (which can be done most easily and almost quantitatively), and if
thereupon we dilute the residue left with water in a proportion of 1 to 1, this water will
absorbed colloidally in toto, and all we obtained is a double quantity of thick and
gummy pulp. The absorption of the water is consummated at lightning speeed and it
needs a fairly large additional supply of water for the gummy mass to become a bit less
thick and viscous.
22
Now, if a large quantity of hot water were to be added to the digester contents, the
necessary function within the contents would be reduced while moreover undamaged
oil cells would be detached from the pericarp and be carried off with the crude, the
intercellular cement being readily soluble in the hot water. During the clarifications,
these unruptured cells would eventually find their way into the sludge and thus the oil
they contain would be lost.
It is time that the supply of generous quantities of water to the pericarp mash gave
lower oil losses in the residue from the presses, but the oil losses in the sludge had
increased disproportionately. No distinct indication was found as to the exact influence
of the viscosity of the crude upon the oil losses in the pericarp residue, but on
theoretical grounds it may be considered certain that this influence is by no mean
negligible.
Boiling Of the Digester Contents
The assumption is justified that the 2 thesis enfolded above give a valid explanation if
the behaviour of the stirring arrangement and the perforation used in digesters, although
it would be difficult to express their relative effect in percentages. In order to account
for the adverse influence of boiling of the pericarp mash with respect to the oil losses in
the press cake, we must hark back to the same theory.
The vapour bubbles rising in the digester give the mash an airy consistency, while
moreover they push the crude oil upwards, so that the mash is abundantly lubricated, so
to speak. Consequently, the rotating stirring arms encounter little resistance and the oil
beating cells are not ruptures in adequate quantities.
Apart from this, the oil is stretched into a thin film at the percipiency of the bubbles.
When the latter burst, the film disintegrates into a multitude of tine droplets, which
strongly promotes the emulsification of the mash.
The requirement of proper digestion
To ensure satisfying results, the pericarp mash pre-treated in the digesters should come
up to the following standards:
1. It may not contain any single undigested fruit and the pericarp should be
completely loosened from the nut.
2. The pericarp mash may not have been digested too thoroughly, i.e. it may not be
over-pulpy and the fibrous texture of the fruit should remain distinctly visible.
3. The mash should be homogeneous: the nuts may not settle out.
23
4. The crude oil liberated during digestion must be released from the digester and
may not be allowed to emulsify.
5. The mash should have such a temperature as to make it suitable for the treatment
following digestion (i.e. at least 80oC).
6. During digestion the mash may not be allowed to boil.
Check On Digestion
The checks to be kept on the proper functioning of the digester are largely of a visual
character (refer to the requirement 1, 2, 3, and 6 as enumerated in the preceding
section). It is even possible to maintain a visual check on the discharge of crude from
the digester (requirement 4), while the temperature of the digester mash (requirement
5), can be measured in a simple way. Unfortunately, all these checks give us no clue as
regards to the actual performance of the machine. If we wish to know whether the
digester really fulfills its task so adequately as to give optimum results, certain
additional measures should be taken.
A regular determination of the oil losses in the fibre does not in itself provide sufficient
information. Any rise or fall in these losses may be attributing to other factors (such as
the degree of maturity of the harvested fruit and the sterilization) besides the digester
proper. Hence, if we want to exclude those factors as far as possible, the various
measures should be carried out within a brief period of time. One of the additional
methods of control we shall now proceed to describe.
The speed of the digester adapted to suit a given pressing capacity e.g. 8 pressing cycles
per hour, is changed to a given moment so as to make the machine discharge a smaller
quantity of mash sufficient for, say, 4 pressing cycles.
In other words, the press is filled every 15mins instead of ever 7.8mins but care should
be taken to make the transition so gradual as to allow the contents of the digester
(which we put at, say, 6 times those of the press cage) to pass through completely
before the lower capacity has in effect been obtained. During the periods of digestion
and it is this cake that are used for the test, which consists in determining the loss of oil
in fibre on a dry basis in each of the cakes and plotting a graph with the aid of the
figures found. The lines thus obtained show the relationship between the loss of oil in
the pericarp residue and the period of digestion.
24
Curve A shows the results obtained with a newly reconditioned digester, the stirring
arms and the perforation of which are in the best possible state. Curve B, on the
contrary, gives quite a different picture, proving as it does that the digesting period
would have to be prolonged considerably in order to arrive at the same outcome.
Now, this is a reliable indication that the stirring effect is poor, which maybe put down
to excessive wear of the stirring arms and/or to an adequate release of crude oil from
the digester.
By carrying out this test at regular intervals and comparing the outcome of the latest
test with that of the previous ones, it is possible to build up a picture of the effect of
gradually increasing wear and tear of the stirring action of the digester. The experiment
suggested, however, is rather comprehensive and requires an appreciable number of
time taking oil loss analyses, so that an easier method is often resorted to. This consists
in collecting and weighing separately the quantities of liquor can be drained from the
digester per unit of time. In order to exclude as far as possible the influence of
fluctuations in the release caused by the filling of the press cage and the changing of the
digester with fruit from the stripper, it is necessary to continue collecting crude
throughout at least 4 pressing cycles. This test is simple and can be taken quite rapidly
and with a sufficient degree of accuracy by moderately skilled operatives.
If the quantity of liquor drawn from a given digester is repeatedly found to be below the
mark, then this, too, is a sure sign of the stirring effect being poor. Before the check is
carried out, however, it s necessary to ascertain how much crude oil can be (or should
be) released from the digester in order to secure optimum results.
This figure, expressed for example is the percentage of crude oil ex-digester by weight
of bunches (otherwise called percentage of cycle released by weight of bunches) must
be found by trial and error and may vary considerably according as the oil content by
weight of the bunches is higher or lower. With fruit that has an oil content of, for
instance 22% by weight of the bunch, the amount of crude to be drawn from the
digester will be much higher than with fruit having an oil percentage of only 15% by
weight of the bunch.
25
Summary
In this issue, attention has been given to the function of the digester in the oil recovery
process. We have demonstrated that the part it plays is not to be underrated and on the
strength of the wide experience that has by now been gained, has expounded the
standards which digesters in an up-to-date palm oil mill should come up to.
Further, we have suggested a couple of methods to keep the stirring effect under
constant and effective control.
In the next issue, of this Review, we proposed to deal with the question of digester
design and will enter at some length into the conditions any digester should fulfill from
the constructional point of view, if it is to ensure good digestion.
Digester II
1. Constructional Details of Digesters
Generally speaking, all digesters resemble each other very closely. As far as there are
any differences worth noting, they are found in the size of the digester, the drive of the
stirring arrangement - which may be located at the top or at the bottom, or may even be
complete absent.
The Digester Drive

Both top drive and bottom drive have their advocates and opponents. The arguments
advanced by either category against or in favour of one of the types do not carry much
conviction; in point of fact, both types are used successfully in practice.
The Perforations
Advantages of wall perforation, which is not in general use, are as follow:
1. The perforation is visible from the outside and can be cleaned during service if
required.
2. As a rule, wall perforation permits the release of larger quantities of crude oil than
does bottom perforations, so that the output of the press is increased considerably,
while the oil losses in the pressed residue remain the same.
The crude oil thus released carries with it an increasing proportion of solids non-fat,
which entails certain consequences in regard to the classification. This point will be
discussed at length in one of the issue to come.
In the preceding number of this periodical, we have pointed out that the quantity of
crude oil that can be - or should be - drawn from the digester depends in fact upon the
26
thickness of the layer of pericarp. That is why most digesters handling (wild) fruit with
a thin layer of pericarp have no perforation, if they had, the fruit mash, which contains a
large proportion of nuts, would become too dry, the fibre would be mashed up as well
and power consumption would be inadvertably high.
The Power Consumption
Power consumption varies widely from one type of digester to the other. The effective
consumption is lowest with digesters having no perforations and highest with digesters
having wall perforations. However, contrary to what one may be inclined to expect, the
relative difference between the consumption figures is small, as is illustrated in the
table below. The comparison it gives is based on 3 different types of digesters, each
handling fruit of the Deli-Dura variety (17-18% of oil by weight of the bunch). At the
capacities stated, the losses of oil in the pericarp residue were about the same for all 3
digesters.
Without Bottom Wall

Perforation Perforation Perforation
Digester height in mms 2400 2400 1800

Digester diameter in mms 1100 1100 1000
Digester contents in litres 2280 2280 1890
Digester contents in tons of bunches K2.5 K2.5 K2.1
Press Cage height in mms 1350 1350 1182
Press Cage diameter in mms 540 540 500
Press Cage contents in litres 380 380 232
No. of pressing cycles/hr 6 8 12
Capacity in tons of bunches/hr K2.5 K3.4 K4.2
Digesting period in min. 60 45 30
Drive in H.P. K15 K22 K25
Effective power take-off
in hp/cu.m digester content K6.5 K10 K13.2
Relative power take-off
In KWh/ton of bunches handled K4.42 K4.76 K4.36
As digesters without perforation have lower effective power consumption than those
with perforation, users of the latter type sometimes feel tempted to plug the perforations
wholly or partly with a view to combining lower consumption with high capacity.
Unfortunately, this arrangement does not pay, because it invariably increases the oil
losses in the pressed residue.
The increase in question is quite appreciable, being in the order of 2% that is to say, if
under normal circumstances the oil lossed in the press cake would be 9% by weight of
the dry fibre, the stopping up of the perforation would raise this percentage to 11. If the
harvest to be processed, were to be of the following composition.
27
Shells 80%
Nuts 38%
Fruit 62% Kernels 20%
Bunch 100% Empty Bunch
Desiccation Water 36%
Pericarp 62% Oil 50%
Fibre 14%
A simple calculation shows that given a capacity of 3.4tons/hr.

3400 x 0.62 x0.62 x 0.14 = 183.5kg of dry fibre would be removed with the press cake.
Consequently, an increase in the oil loss in the press cake of 2% by weight of the dry
fibre would be tantamount to 0.02 x 183.5 = 3.67kgs. of oil per hr.
Against this we may set the economy in effective power consumption which would be
22 - 15 = 7hp. The calculated relative power consumption is (15 x 0.736) ÷ 3.4 =
3.25KWh/ton of bunches, so that the reduction will be 4.76 - 3.25 = 1.51KW/ton of
bunches. The price to be paid for the reduction would correspond to 3.67 : (3.4 x 1.51)
= roughly 0.7kg of oil/KWh which needless to say is a bad bargain.
The Stirring Arms
The stirring arms, by means of which digestion is effected, are subjected to

considerable wear and tear and therefore they are made of a very wear-resistant
manganese-silicium steel. The jacket wall is protected by special liner plates fitted in
way of the stirring arms. Baffles made of angle bar and serving the purpose of keeping
the mass from rotating with the arms are set vertically against the jacket wall, between
the liner plates.
In some digesters, flat bars arranged clockwise are used to be the same end. The route
traveled by and any single fruit while it is in the digester is rather erratic; in the main it
makes a larger number of almost vertical upward and downward moments. These
movements, by which the fruits are displaced over comparatively great distances with
respect to each others and are rubbed along each other, result from the oblique position
of the stirring arms. The rubbing action of the fruit is at least as important to the stirring
effect as a direct contact of the fruit with the arms.
The stirring arms are set at such an angle as to lift the fruit mash while they rotate.
Owing to the wear that will gradually set up on the stirring arms, the distance over
which they force the fruit mash upwards will diminish, as shown below:
Stirring arms which have been worn away too far cause a considerable reduction in the
quantity of crude oil released through the perforations, while moreover the power
consumption will drop and the oil recovery quotient will be greatly impoverished.
28
Furthermore, excessive wear of the expeller blades at the bottom of the digester makes
it impossible to keep the perforations fully open, which causes the efficiency to
decrease even further.
We shall now illustrate by means of a rough calculation the financial and economic
consequences of employing excessively worn stirring arms.
Depending on the type of fruit handled, a set of arms can give from 2500 to 300 hrs of
useful service. The value of a complete stirring arms plus 1 set of expeller blades
roughly corresponds to 1500kgs of palm oil.
In this valuation, we have taken the estate as a basis, increasing the price of the stirring
arms by the cost of storage and transportation and proceeding from the sales price ex-
works as far as the quantity of oil is concerned.
With these assumptions in mind it would be possible to economise a sum equivalent to

the value of 600kgs of oil on the running expenses of the estate, by prolonging the
service of the stirring arrangement for another 1000hrs. The use of excessively worn
stirring arms, however, will entail increased oil losses in the pressed residue. It would
be far from exaggerated to assume the difference in terms of oil losses by weight of dry
fibre between new and worn stirring arms to be 1%.
Now, proceeding on the proposition that the throughput capacity of the mill is 3tons of
bunches/hr, and that the composition of the fruit is identical with that mentioned in the
preceding paragraphs, we can make a calculation to the effect that during the prolonged
service of 1000hrs a quantity of 100 x 3000 x 0.62 x 0.62 x 0.14 x 0.01 = 16200kg of
oil is lost additionally.
29
Hence the economy in upkeep cost which we estimated to be equivalent to
approximately 600kgs of oil in no way counteracts the increased oil loss in the pericarp
residue, amounting to some 16200kgs.
Though naturally the figures started should not be taken too absolutely (we only meant
to give an example), they go far to show that here again crime does not pay.
Influence of the Digester Contents on the Stirring Effect
One of the requirements of proper digestion which is most frequently neglected in

practice is the necessity of keeping the digester properly filled. It is obvious that the
height of the fruit mash column in the digester is decisive if the pressure within the
mash, as a result of its own weight and, consequently, determines the friction that is
enchantered by the stirring arms. Hence there must be a pronounced relationship
between the height of the column and the stirring effect produced. Further it will be
clear that the stirring time, too, has a certain influence upon the stirring effect.
When the digester is not kept completely filled, the stirring time will be shortened and
the pressure in the fruit mash reduced, and these 2 factors combine to impoverish
digestion.
If we assume the product of filling level and bottom pressure to be 100% with
completely filled digester, it is possible to picture the effect of inadequate filling as in
the figure below:
30
The curve h (height) x p (pressure) runs from 100% to 1%. From this diagram, it
appears that when the digester is filled to 75% of its capacity, the stirring effect is only
56% of the optimum. Now, one should not imagine that it is impossible to recover any
oil from undigested fruit during the subsequent pressing cycle; the point is that there
will be an appreciably greater loss of oil in the pressed residue. The curve designated
by the Greek letter ‘eta’ shows approximately what an increase in question will amount
to.
For instance oil loss to dry fibre sustained in a given press unit is 9.5% if the digester is
full. If the digester is ¾ filled, the oil loss would be likely to rise to:-
100/91 x 9.5 = 10.5% to dry fibre

If the disgester were filled up to 60% of its capacity, the increase could be expected to
mount to:
110/85 x 9.5 = 11.2% to dry fibre
These figures clearly prove the importance of keeping digesters quite full.
Summary
In the present number we have approached a few questions of digester design, dwelling
at some length upon the increased oil losses that may ensure. In the next issue, we
propose to broach the subject of ‘pressing’; it will deal with: The correlation between
fruit mash components and the losses of oil in dry fibre.
Comments
During visits to several palm oil mills the questioner has observed that in a given mill
the pressed residue visually contained more oil than residue in another given mill,
whereas analytical figures respecting the oil losses to dry fibre which were shown to
him did not indicate such a difference at all, or were evening favour of the residue that
was more oily to look at.
The question be posed is as follows: Is it possible that the difference observed is a

result of varying methods of analysis?
Reply
If we rule out the possibility of faulty sampling as the cause of the difference, we have
no choice left but to attribute it to varying methods of analysis.
It will be known that analysing oil losses to fibre is still a matter of great difficulty,
which requires further research. For example different extraction agents have been
found to produce varying analysis results. To illustrate the point, we quote a couple of
lines from the 7th Annual Report of the West African Institute for Oil Palm Research
(WAIFOR);
31
It was observed that consistency high oil to dry fibre figures were obtained when
Petroleum Ether of the Boiling range40-60oC was used for extraction. Similar trends
had been reported by other laboratories and companies of different solvents suggested
that the mill extraction efficiency based on oil to dry fibre decreased with the solvent
used for analysis as follows:-
Petroleum Ether 60o-80oC < Shell British Petroleum Solvent (S.B.P.) 60o-120oC <
S.B.P. 62o-88oC < Isohexane < Petroleum ether 40o-60oC.
Further we know that the Soxhlet method, with which the sample is pre-dried to a
constant weight, gives slightly lower oil loss figures than does the methods with which
the wet fibre sample is analysed at once. The latter method: Areotrope distillation -
which is used frequently in e.g. Sumatra, is somewhat less accurate, but it has the
advantage of giving quick results. That is why it is sometimes called the ‘rapid’
method.
However the differences between analytical figures resulting from the use of different
solvents and varying methods of extraction are not so great as those found between pre-
dried and non-dried samples or samples that are first kept for a certain period.
To explain this, we shall give a number of analysis results. The tests in question were
carried out simultaneously at various estates and the results were found to be analogous
in all cases.
Of a certain quantity of thoroughly mixed fresh pericarp fibre, a part was analysed
immediately according to the ‘rapid’ method. The oil loss figure found was 9.9% to dry
fibre. Another part was first dried to a constant weight in a drier and next analysed. The
oil less then found was 8.2% to dry fibre. The rest of the fibre was kept in an open
basket and from this samples were drawn after 24, 48 and 72 hrs, which samples were
also analysed according to the ‘rapid’ method. They showed a loss of 9.1%, 8.2% and
7.2% oil to dry fibre respectively.
This justifies the conclusion that apparently the oil content of the fibre decreases during
the keeping period and that besides desiccation other factors must be at play, for if the
sample is allowed to stand for a certain time the oil content decreases to a greater extent
than could be explain by desiccation alone.
Further it has been found that the manners in which samples are kept i.e. in ‘open’ or
‘closed’ containers, is responsible for considerable variations in the test results.
The figures below clearly illustrate the variations alluded to. The samples listed under
the heading ‘open’ were kept in plaited baskets without cover and those indicated under
the heading ‘closed’ were kept in closed tins. All the analyses were carried out
according to the Soxhelt Method.
32
Initial samples
Moisture Content 34.7%
Oil to dry fibre 9.3%
FFA content 2.1%
Sample kept for 6 days Open Closed

Moisture Content 13.1% 38.4%
Oil to dry fibre 7.1% 5.1%
FFA content 10.8% 30.1%
Unsaporitiable

Unsaporitiable 5.4% 7.7%

Unsaporitiable 5.4% 15.3%
Approximately though these figures may be, they show indubitably that the oil losses
found in fibre that had been kept for a certain period of time are unrealistic in that they
no longer represent the losses actually sustained in the recovery process.
Hydraulic Presses
Introductory Note
Although the loss of oil in the pressed residue depends largely on the composition of
the fruit mash to be pressed, as had been explained in the preceeding issue of this
periodical, yet there are other determining factors.
The temperature of the fruit mash, the pressure exerted, the pressing time adopted, the
design of the press used, all these are points which affect the final result. In this present
number we shall let in some light on the points mentioned, giving special attention to
the working and the effect of the automatic press.
1. Limits of the pressure, time and temperature adopted in the pressing cycle.
1.01 The pressure applied
As the industrial recovery of palm oil (I.e. fruit oil) with the aid of hydraulic presses is
much more recent than the recovery of seed oil (e.g. cotton seed oil), it is not surprising
33
that the experience gained with seed oil presses was utilized in the development of oil
palm presses. It has been found, however that one principle of seed oil recovery i.e. the
relationship existing more or less by law between cake pressure and the loss of oil in
the pressed residue does not apply to the mash of palm fruit.
In seed oil recovery it is a rule of-thumb that the loss in the press cake is inversely
proportional to the .4 power of the cake pressure. This means that if the oil loss in a
given cake is 12% at a cake pressure of 50 kgs/sq cm it would be 9.090 at a pressure of
100 kgs/sq.cm under equal circumstances and 6.88% at a pressure of 200 kgs/sq.cm.
The reader will be familiar with the fact that with palm oil presses the cake pressure
should not as a rule exceed some 80 kgs/sq.cm. It has been demonstrated by experiment
that -contrary to what one would be inclined to expect - the oil loss to fibre increases
with higher pressures. The phenomenon has been explained by researchers as for
instance L. Kehren, who has subjected the fibre to microscopic examination and found
that it was as it were impregnated with oil. He has been able to prove that the
impregnation is a result of the pressing process and is more intensive as the pressure
applied is higher.
34
1.02 The pressing time
With palm oil presses as with seed oil presses, there is a distinct relationship between
the oil loss in the pressed residue and the period during which the cake has been under
full pressure. This is clearly shown in graph nr 1, in which the curves of Fig. I Vol. I Nr
5 (indicating the correlation between the oil loss in the pressed residue and the pressing
time) have again been set out, but this time on a log-log.scale.
From these straight lines the following formula can be derived:
L =C/Tn
in which L represents the oil loss to dry fibre, C a constant, T the pressing time in
minutes under full pressure and n a number smaller than 1.
C is a constant value for a given cycle, but it may change from cycle to cycle. If the
fruit mash has been given a good preparative treatment in the digester and has the
proper temperature the value of C is approx. 10, but it may be as much as 40 if the
preparative treatment has been inadequate. As to the value n, as far as can be inferred
from the material available, this may vary from .10 to .20. The formula in question
gives ground for the useful conclusion that any continuation of the pressing cycle
beyond the point where the mash has been under full pressure for a couple of minutes is
not likely to produce much extra oil.
1.03 The pressing temperature be
We have emphasized fore that the temperature of the digested fruit mash has a
considerable effect upon the oil losses in the pressed residue. The causal relationship
between these factors lies in the viscosity of the oil extracted, which is lower as the
pressing temperature is higher.
Several research workers have attempted to establish a law defining this relationship,
but so far they have not been able to adduce any conclusive evidence. There are good
reasons to presume, however that the percentage of oil lost in dry fibre is inversely
proportional to the sixth power of the absolute temperature.
Even though this statement has but illustrative value, it is beyond doubt that the
temperature of the mash to be pressed is tied up closely with the oil losses in the
pressed residue.
35
2. The Press Cage
2.01 The perforation used
As regards the size of and the intervals between the perforations used in the press cage,
there are a great many varieties and perhaps as many points of view. This is not
unnatural if we remember that for a long time (and partly even to this clay) the opinion
has prevailed that the pressing effect depends largely on the type of perforation
employed. The perforation most commonly found in press cages at present has a
lozenge-shaped pitch of 12.5 mms and a hole diameter of 2 mms. From the outside of
the cage the holes have been bored out to 4mm over approx. half the length of the cage.
We shall abstain from offering much comment on this point and confine ourselves to
stating that by experiment we have found that the number of perforations and their
diameter to be susceptible of variation within wide limits, without giving any
appreciable difference in pressing result.
2.02 The press cage diameter
One might wonder it limits should not be set to the diameter of the press cage, for, as
the diameter is larger the route to be traveled by the liquid from the press cake to the
cage wall increases and so does the time required for the purpose. Although this idea is
certainly correct, we are unable to indicate the exact point beyond which any adverse
effect would make itself clearly felt. All we can say is that even with the largest press
cage manufactured by Stork, which cage is 540 mms in diameter, no such bad effect
has ever occurred.
36
2.03 The length of the press cage
It will be clear to anyone who has read the explanation given in the preceding issue of
this Review that the length of the press cage does have a marked influence upon the oil
losses in the pressed residue.
As a result of the wall friction, the average cake pressure will decrease, causing the oil
losses to rise proportion as the cage is longer.
3.0 Pressing methods
3.01 Applying pressure to the pericarp mash
37
It has been found that when the digested fruit mash is put under pressure rapidly, the oil
loss in fibre is higher than when pressure is allowed to build up gradually. One of the
explanations given for this phenomenon is that by a sudden pressure increase part of the
capillary ducts is squeezed shut, and oil is confined in the spaces thus blocked. This
theory may be sound, but it is just as plausible to assume that owing to the higher speed
at which the press ram moves up and down the imaginary paraboloids which enclose
the areas of equal pressure (refer to the foregoing issue of this Review) will be flatter,
i.e. not so high, in consequence of greater frictional forces. Hence the pressure is
transmitted over a shorter distance in the cake, so that a large proportion of the fruit
mash is not under adequate pressure during the cycle.
A popular illustration of what happens is found in daily life. People seldom manage to
pass their arms quickly through the sleeves of a jacket or coat. If they try to do so, the
arms get caught half way the sleeves in most cases.
With presses receiving the necessary hydraulic fluid from a pump to which they are
coupled direct, a gradual build-up of the cake pressure is ensured automatically.
The situation is different, however when a hydraulic accumulator unit is installed

between the pump and the press (or presses), for of no special precautions are taken, the
fluid in the high pressure or low pressure accumulator is apt to flow to the press
cylinder in a very short time. There are many devices available which preclude a
precipitate flow of high pressure or low pressure fluid from the battery. These devices
have a common feature in that they force the liquid to pass through a narrow aperture,
as a result of which it is checked in its course. As long as the press ram moves, there is
a considerable difference in pressure between the spaces before and behind the
aperture. When the ram comes to stand-still, the run of the fluid stops and the pressure
before the aperture becomes the same as that behind. The change in pressure difference
is used to operate the valves which admit fluid of a higher accumulator after the
possibilities of the low pressure accumulator have been utilized to the full.
3.02 The floating press cage
As far as we can ascertain, the floating press cage which has found widespread use in
the seed oil industry, has never been applied in the industrial recovery of palm oil.
Nevertheless, it is useful in the compass of this article to discuss the principles of the
f1oating cage, because it eliminates to a great extent the drawbacks consequent upon
wall friction.
In a press having a floating cage, the cage is closed at the top by a die, which just fits
into its inner circumference
As the cage is not retained, it will be pushed upwards along with the rising press ram,
as soon as the frictional forces along the wall exceed the weight of the cage.
38
The mash to be pressed being held back by the counter-die, adversative frictional forces
will arise owing to the relative movement of the press cage with regard to the mash. As
a result, the end of the stroke will find the cage in a position where it has risen over a
distance tat is slightly less than half the stroke of the press ram.
It is clear that at the end of the stroke the counter-die absorbs a force corresponding to
the one exerted by the ram, less the weight of the press cage and the cake.
With regards to the force exerted by the ram, the weights of press cage and cage are
neglible, however, so that we may state that when a floating cage is used both the top
and the bottom of the mash are put under maximum pressure, the zone of the lowest
pressure being in the middle of the cake (refer to Fig. 2).
39
3.03 Pressing by pulsation
Experience has taught that the oil loss in the pressed residue is reduced when the
pressure on the cake is released and then applied again. There two ways to explain this
fact, viz:
a. Owing to the pressure release, the cake expands, by which the frictional forces
between the cage and the cage wall are largely eliminated. When pressure is
applied subsequently, a smaller proportion of the force exerted by the press ram
will be lost in wall friction, which results in a higher average cake pressure and,
consequently, in lower loss of oil.
b. During the pressure release most of the internal frictional forces are eliminated as
well, so that when pressure is applied subsequently, the solid particles I the press
cake assume a slightly modified position with respect to each other. This exposes
the capillary ducts and hollow spots in the cake to further constriction and
consequently, it is possible to exert on the pressure which is required to expel it.
The method of pulsation here described is of little practical use in regard to standard
hydraulic presses, because, comparatively speaking, much time is involved in releasing
small quantity of oil recovered additionally would be excessive.
As regards the Stork automatic palm oil press, this is so designed that the principle of
pulsatory pressing is taken full advantage of. Pressure on the mash in the cage of this
automatic press is released and reapplied 4 to 5 times, and thus a substantial
contribution are made towards the reduction of oil losses in the pericarp residue.
40
41
4.0 Various types of Stork palm oil presses
4.01 The hand operated hydraulic press

The hand operated hydraulic press as pictured in Fig. 3 has been designed for the
special purpose of handling small quantities of fruit. The press cage has it diameter of
350mm and is 400 mm high. At normal working speed it is possible to carry out 5
cycles per hour, in which approx. 200 kgs of cooked and digested fruit are treated.
The principal feature of this simple 50 tons press is the design and the position of the
press ram. This ram is hollow (refer to Fig. 4) and moves downwards instead of
upwards as with the standard presses.
On completion of a pressing stroke the ram is drawn back to its starting position by
means helical springs.
The max. fluid pressure of 70 kgs/sq.cm results in a cake pressure of approx. 50

kgs/sq.cm.
The fruit that is handled in this type of press is sterilized in an open drum and then
crushed in a wooden mortar. As the preparative treatment of fruit in such a mortar
cannot be as efficient as in a digester, while moreover the mash cools down during the
crushing process, the oil loss in dry pericarp fibre is bound to be higher than with
presses of the standard type. In spite of this, the average oil recovery quotient obtained
with the hand operated press is over 90%, which is much higher than the percentage
obtained with screw spindle presses. The West African institute for Oil Palm Research
(WAIFOR), near Benin City (Nigeria) has tried out the hand press in its experimental
station and has arrived at the following results:
Pressing with fruit sterilized in open fired drum and pounded in wooden mortar
(peasant extraction):
% oil to dry fiber fruit % extraction
Fruit: matte:
407.0 35.0 91.4
315.5 30.3 90.4
279.6 24.0 91.4
413.3 23.3 94.4
510.5 30.7 94.0
388.0 18.6 95.2
360.7 31.3 91.3
567.5 32.2 94.3
Average oil to dry fibre of extracted matte: 28.2%

Average % extraction: 92.8%
42
43
4.02 The standard press – 1.5 ton
The cage of this press has an inside diameter of 430 mms and a height of 1040 mms,
which corresponds to an effective content of 150 litres. The max. fluid pressure of 110
tons results in a max. cake pressure of 75 kgs/sq.cm.
At 6 to 7 cycles per hour the capacity of the press is approx. 1000 kgs of fruit per hour,
which is equivalent to a throughput capacity of approx. 1.5 ton of bunches/hour.
The main characteristic of the press is its square housing, which s constructed from
electrically welded mild steel segments. A picture of the press is given in Fig. 5.
4.03 The standard press – 3 tons

The first press of this type, of which some 270 are now in every day use, was delivered
in 1928 and is in full service at the present day. The max. fluid pressure is 172 tons,
which implies a max. cake pressure of 75 kgs/sq.cm.
The inside diameter of the press cage is 540 mms and its height is 1346 mms, the
effective content being 300 litres. If the number of pressing cycles is 6 or 7 pet hour,
the capacity of the press is 2000 kgs of fruit per hour. Expressed in terms of harvested
bunches this is tantamount to 3 tons per hour. The press is shown in Fig. 6.
4.04 The automatic press
The automatic press, of which Fig. 7 gives a sectional drawing and Fig. 8 a photograph,
is provided with a press cage identical with the cage of the 3 tons standard press.
Furthermore, the digester with which the press forms one unit is the same as that used
in conjunction with the 3 tons press and the hydraulic pump serving the one type of
press only requires a minor modification in order to be suitable for the other. These are
all the points which the two types of presses have in common.
The characteristic difference between the automatic press and the standard press is that
the action of the former is completely automatic and is not controlled by any
manipulation from outside, while the pressing method applied varies widely from the
usual one.
44
As is indicated in Figures7 and 8, the press is in direct communication with the
digester. In Fig. 7 the press ram, arranged above the press cage, is in its highest
position, which means that the stirred fruit mash can pass freely from the digester into
the press cage.
The cage contains 4 cakes of residue prom the previous pressing cycles (not drawn in
Fig. 7). The cone of the cake breaker blocks the cage at the bottom. When the press ram
comes down it first closes the digester outlet and next puts the mash in the cage under
45
increasing pressure for some time, the cone of the cake breaker moves downwards,
releasing the opening at the bottom of the cage (refer to Fig. 9).
Now the press ram passes further down the press cage, forcing out a part of the pressed
residue. When it has reached its lowest position, it returns to its starting point at an
accelerated pace, unblocking the digester outlet, so that a fresh quantity of fruit mash is
allowed to enter the cage on top of the newly produced cake.
The cake breaker cone rises almost simultaneously, and by so doing breaks up the
pressed residue is discharged subsequently.
At the end of its upward motion the cake breaker cone pushes the non-expelled cake
slightly back into the press cage. The frictional forces thus produced along the cage
walls are opposite to those arising from the stroke of the ram. The very useful purpose
served by the counter-active forces will be perfectly clear to the reader after the
explanation we have given concerning the adverse influence of wall friction upon the
pressing effect. As the cake breaker cone need not - and does not - close off the press
cage from bottom entirely to prevent any pressed residue from being forced out, no nuts
will be cracked between cage rim and cone during the pressing process.
The operation of the press is effected with the aid of valves, which are controlled by
cam discs. It is possible to shift the position of the various cams with respect to one
another, while moreover the speed of the shafts on which the cams are fitted is
infinitely variable. In consequence, the programme of any given cycles, as well as the
number of cycles per hour can be varied within wide limits. All the control gear is
mounted in a separate box (Refer to Fig. 10).
The automatic press has been designed for a throughput capacity of 4.5 tons of bunches
per hour, but when tested in practice it has been found capable of handling 6.5 tons of
bunches per hour.
As a result of the pressing method used, every part of the press cake is put under high
pressure 4 to 5 times at various points in the press cage, prior to being expelled. If in
the press cage there should be any zone where the prevailing pressure is higher than in
other parts, it is almost certain that every portion of the cake press has been subjected to
the higher pressure some time during the successive cycles.
Owing to the automatic control and the accelerated reverse motion of the press ram,
there is hardly any loss of time in a given cycle, so that the total period during which
every part of the press cake is under maximum pressure is practically the same as that
in a press of the standard type.
Inasmuch as neither the fruit mash nor the press cages are in contact with the
atmosphere, the high temperature of the fruit mash is maintained better throughout the
process.
46
All these factors have a favourable effect with regard to the pressing efficiency, as we
have been able to demonstrate in practice.
Fig. 11 give a comparison of the pressing effects between the automatic press and a
standard press, the line A representing the results produced by the standard press and
the line B those obtained with the automatic press. Both presses are tried out
simultaneously, on the same type of fruit. The oil losses were determined by the same
method and by one and the same analyst.
Now, if one does not pay attention in the first place to the effective value of the oil
losses to dry fibre, but to the relative differences, it will be clear not only the automatic
press has a considerably greater output than a standard press of the same size, but that it
is much efficient.
Clarification I
Introductory note
In the clarifying station the crude oil is seperated into pure oil and sludge, allowance
being made for the double necessity of restricting the amount of impurities in the oil to
the greatest possible extent and reducing the loss of oil in sludge to a minimum.
As with nearly all technical processes, the effect of the clarifying process is not 100%,
although the clarifying station is of an efficient set-up this optimum can be approached
fairly closely. For no good reason the clarifying process has always been shrouded in a
veil of mystery. In the present article we propose to demonstrate that it is subject to
simple natural laws and that once we are familiar with these laws it is possible to
improve the clarifying effect by locating and removing any faults in the clarifying
station.
1. The composition of the crude oil
If we consider the way in which the crude oil is recovered from the pericarp in the
pressing station, we find that the crude oil has three main components, i.e.
1.01 A mixture of oil and water
This mixture can be sub-divided into:
1.01.01 Practically non-dispersed oil in water
The oil in such a mixture originates for the most part from the digesters and from the
press cages during the period when the mash is not under high pressure.
If one allows this mixture to stand for a short time, the separation will be effected
almost completely. The oil from the mixture is generally designated in palm oil mills as
47
“free oil”, which means that this oil has no affinity whatever for the water surrounding
it.
1.01.02 Highly dispersed oil in water
This mainly comprises the liquid that is expelled from the press cage when the well-
stirred fruit mash is put under high pressure. The crude oil is then forced through the
narrow openings and capillary ducts in the press cake, so that it is homogenized, as it
were, in the watery medium.
1.02 Oil in water emulsions
If the oil is highly dispersed in the water, it is possible to speak of true homogenization
and more or less stable emulsions are formed.
Next to this, however, there are emulsions in which the dispersion has not advanced to
the stage of homogenization. In such cases the emulsion is formed by the presence oil
so-called emulsifiers, which have the effect of increasing the stability of the suspension.
Fatty acid—mucilages—finely divided fibre remnants etc. etc. may have the function of
emulsifier or stabilizer.
1.03 Water in oil emulsions
As far as these emulsions play a part at all in the palm oil trade, they do not present any
special difficulty in the oil recovery process, so long as the correct temperature is
maintained during the clarification (i.e. 80-100° C.).
2. Methods of clarification
Of the various systems of clarification that are used individually or in combination in

the palm oil trade, i.e. static clarification, static clarification followed by sludge
treatment in centrifugals or static clarification followed by a biological treatment of the
sludge, the system of static clarification is far and away the most important. The
principles underlying static clarification are the corner stone of clarification itself and
they are also largely applicable to the other methods mentioned.
3. Basic principles of static clarification
Inasmuch as oil is essentially immiscible with water (sludge), a mixture of oil and
sludge always tends to separate into two layers, in such a way that the liquid of the
lowest specific gravity—in this case the oil—is at the top. The force separating the two
elements is the difference in specific gravity between them.
48
An oil droplet of the volume V in sludge undergoes a vertical upward force
corresponding to V x Sd in which formula Sd represents the specific gravity of the
sludge. The weight of the droplet itself is V x So, in which So stands for the specific
gravity of the oil. The force pressing the droplet upward is the difference between the
vertical pressure and the weight i.e. V x Sd - V x So = V (Sd - So).
Now the difference in specific gravity between sludge and oil is practically constant
throughout the temperature range from 40°C to 100°C. In other words if we take the
value Sd-So as equal to 0.1 the error involved is negligible. Thus, V (Sd - So) passes
into 0.1 V. If the diameter of the droplet is d, V will be 1/6 πd3, so that the force
pressing the droplet upward is equivalent to:
0.1 x 1/6πd3 = 1/60πd3 (1)
Force engenders acceleration; hence the droplet, if not checked, will rise with
accelerated motion, which in fact it does initially. It is at once checked in its course,
however, owing to the resistance of the sludge, which resistance increases as the speed
of the droplet is higher.
At a given moment the speed is so high that the restraining force becomes equivalent to
the upward pressure and the droplet no longer rises with accelerated but with uniform
velocity.
According to Stokes' law the magnitude of the resistance corresponds to
3. πη .d.s (2)
In this formula the symbols have the following significance:
η = dynamic viscosity of die sludge

d = diameter of the oil droplet
s = the constant rate of rise of the oil droplet.
If we take the two forces (1) and (2) as equal, we find the relationship between the rates
of raise, the size of the droplet and the viscosity of the sludge:
3πη doss = 1/60 πd3, or s = d2/180η
In this form however the formula is of little practical value the magnitude of the various
factors being unknown. In order to find this magnitude, the following experiment has
been carried out. In a trough having a depth of sedimentation of 1 metre, sludge was
allowed to settle for two hours. Next sludge samples were examined under a
microscope. This revealed the presence of oil droplets of sizes varying from less than 1
micron to 30 microns ( 1 micron is 1/1000 mm). Of droplets measuring 30 microns there
were but few; of smaller droplets great numbers were observed. There were no droplets
exceeding 30 microns.
49
The droplets of 30 microns must have had a rate of rise of over 1 metre in 2 hours =>
140 microns/sec. As has been said there were a multitude of droplets of under 30
microns which had not attained the velocity of 140 microns/sec. indicated. Hence it is
evident that the particles of 30 microns of which only a few were found, must have just
been able to reach the rate of 140 microns/sec.
Given a rate of rise = 140 microns/sec. and a droplet diameter d = 30 microns, it is

possible to calculate the dynamic viscosity η. Making allowance for the units, we arrive
at .036 deci-poises for the value η. When we consider that water of 90° to 100° C has a
dynamic viscosity of approx. .03 deci-poises, it appears that the viscosity of the sludge
examined exceeds that of water by 20%, which is in keeping with the results of direct
viscosity tests. After this preliminary calculation, which has made clear the order of
magnitude of the various factors, it is possible to draw up a table covering the
calculated rate of rise in micron/sec. for oil droplets measuring from 1 to 40 microns in
watery sludge at viscosities varying from .01 to .25 deci-poises.
Viscosity in deci-poises
0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10 0.15 0.20 0.25
1 0.56 0.28 0.19 0.14 0.11 0.09 0.08 0.07 0.06 0.06 0.04 0.03 0.02
5 14 7 4.6 3.5 2.8 2.3 2.0 1.8 1.6 1.4 0.9 0.7 0.56
10 56 28 19 14 11 9.3 8.0 7.0 6.2 5.6 3.7 2.8 2.2
15 125 63 42 31 25 21 18 16 14 13 8.3 6.3 5.0
20 222 111 74 56 44 37 32 28 25 22 15 11 9
30 500 250 167 125 100 85 71 63 56 50 33 25 20
40 890 445 296 222 178 148 127 111 99 89 59 44 33
The values in this table have also been set out in a graph (Fig.1). It gives the
relationship between the rate of rise and the size of the droplets at varying viscosity on
a log-log scale. The lines are therefore lines or constant viscosity.
3.01 The effect of the droplet size upon the oil losses in sludge
Graph 1 is self explanatory. The main conclusion to be drawn from it is that in practice
there will always be some loss. Humanly speaking, it is impossible to reduce the oil
loss to zero. Particles smaller than say, 15 microns will always get lost, whether wholly
or in part.
Hence the oil loss in sludge depends among other things upon the size of those small
droplets. As far as they are present in the palm fruit there is little to be done about it; for
the rest one should attempt as far as possible to prevent them from being formed in the
process. Among the factors contributing to their existence we mention protracted and/or
over vigorous stirring in the digesters as well as protracted and/or heavy boiling of the
crude oil in the pre-heaters.
50
51
3.01.01 Effects of the digesters
It can be said that as a general rule the fruit mash should not be stirred longer than is
rendered necessary by the combined oil losses in pericarp fibre and sludge.
In this connection we see some use in referring once more to the matter of water
injection into the digester. Large-scale experiments carried out in Sumatra have clearly
demonstrated that the oil losses in sludge resulting from water injection into the
digester were far higher that those sustained when no such injection was used. Below
we are giving a few figures illustrating the point. They are a comparison between the
percentages of oil to dry matter in sludge with and without water injection into the
digester after equal periods of sedimentation:
Oil % to dry
Test matter in sludge Remarks
1 37.9
2 37.4 with water injection
3 37.3
5 23.9
6 25.8 without water injection
7 25.4
The difference between the test results can be explained in three ways:
First of all, it is easy to realize that with water injection more solids-non-oil are milled
(or washed) out, so that the viscosity of the sludge is increased and the rising of the oil
droplets retarded.
The second factor, which has been discussed previously in the chapter on digesters, is
that the warm water added tends to dissolve the intercelular cement between the oil
bearing cells, causing non-ruptured cells to find their way into the crude oil.
The third point—which is the most important in the compass of this paper—is that by
the addition of water the oil in the digester is more liable to be broken down into minute
particles.
If the stirring arms of the digester rotate through a medium of pure oil, it is easy to see
that there can be no disintegration, for the simple reason that no other liquid will
intermingle with the broken down particles, so that they fuse together again. The
situation is quite different, however, as soon as water is injected into the fruit mash.
True, the mash does contain some water of itself, but most of this is not free, so that in
general the mash in the digester behaves as a medium of pure oil. The more water is
added to the mash from outside, the greater will be the risk of disintegration of the free
oil.
52
Next to the danger of unopened cells being washed away, this is a^l^iotl^ier reason
why we must set down the method of water injection as unsuitable.
3.01.02 Effects of heavy boiling
It has been clearly demonstrated in practice that heavy boiling of the crude oil has an
adverse effect upon the static clarification. The results obtained with experiments
carried out to establish this effect have proved gentle boiling to give the following
advantages:
1. The oil recovered is very clear to look at.

2. The amount of oil in sludge is smaller.
3 The sludge produced on gentle boiling has been found to release its oil more readily
when the biological method is applied.
4 The period of sedimentation may be shortened considerably and yet give the same
result.
The points mentioned sub 2, 3 and 4 are attributable purely to a larger average size of
the oil particles.
The sludge used in the tests mentioned was of a low viscosity because of the low
percentage of snow. The case is quite different however, if large quantities of oil are
released from the digester. Owing to the comparatively high percentage of solids-nonfat
then contained hi the sludge, its composition will be such as to render favourble the
influence of intensive boiling upon the viscosity, which influence is clearly visible.
The viscosity of unboiled or insufficiently boiled sludge having proportion of s.n.o.

undeniably gives the impression of being very high.
Sludge of this nature must be subjected to heavy boiling until all foam has disappeared
from the surface of the liquid. The procedure in question, however, will always be at
the expense of the size of the oil droplets.
Now, in order to minimize the risk of disintegration one should deprive such oil as will
separately readily from the crude of every opportunity to break up as a result of heavy
boiling, by discharging it prior to boiling the sludge. It is for this very purpose that the
continuous first clarifier is included in the clarilying plant. ^J
3.01.03 Effects of the mill capacity upon the oil lower in sludge
It has been found repeatedly that the oil losses in sludge increase with the output of the
mill, that is to say, in a given palm oil the oil losses in sludge at a high rate of milling
will exceed those suffered at a low rate of milling.
53
If, for example, we assume the normal time of sedimentation for sludge to be two
hours, the viscosity of the sludge being .04 deci-poises and it being our aim to recover
all the oil particles of 30 microns upwards, we can calculate how deep should be the
settling tank if it is to answer the purpose. From graph 1 we can infer that an oil droplet
of 30 microns in sludge having a viscosity of .04 deci-poises has a rate of rise of 125
microns/sec. Consequently, in two hours the droplet is capable of covering a distance of
2 x 3600 x 125 = 900.000 microns. If therefore, we give the tank a max. height of 90
cms, we are able also to recover those particles of 30 microns which will be on the
bottom initially.
Of the particles smaller than 30 microns a large proportion will not be recovered at all.
Graph 1 shows that the particles of say, 20 microns have a rate of rise of 56
microns/second. In other words it is possible to recover these particles down to a depth
of 2 x 3600 x 56 microns = 40.1cms).
Now, when the rate of milling increases, and with it, the production of crude oil and
sludge, the residence time in the settling tank must of necessity be shortened. For
example, if the residence time in the tank were to be reduced to half the original period,
the reader will appreciate that the oil droplets 30 microns, which at the beginning of the
settling process are more than 45 cms removed from the surface would not be
recovered. For particles of 30 microns this distance would be 20 cms. Hence, whereas
in the former case allowing for a period of sedimentation of 2 hours, all the particles of
30 microns and 44% of the particles of 20 microns would be recovered, in the latter
case, which pre-supposes a reduction of the settling period by half, this would be only
50%, of the particles of 30 microns and 22% of the particles of 20 microns,
It follows that the higher the rate of production and the shorter the periods of
sedimentation the greater will be the losses of oil in sludge and vice versa.
3.02 Effect of the viscosity
Graph 1 also shows the enormous influence exerted by the viscosity upon the rate of
rise.
If one requires a rate of rise of 125 microns/second, the viscosity may be .071 deci-
poises for particles of 40 microns. For droplets of 30 microns, however, the viscosity
should be reduced to .04 deci-poises. If one should wish to go even beyond that by
including particles of 20 microns, the viscosity would have to be reduced further, i.e. to
.018 deci-poises, but this is outside the scope of practical possibility, because it would
imply going below the viscosity of pure water at 100° C, which is 0.0282 deci-poies.
Hence, if we take .03 dp as the extreme limit for the viscosity of sludge, the particles
have their maximum rate of speed at that point.
According to the graph this is 74 microns/second for droplets of 20 microns, 167

microns/second for droplets of 30 microns and 296 microns/second for droplets of 40
54
microns; which explains once again why particles of 15 microns with their max.
velocity of 42 microns/second is not susceptible of effectual recovery in practice.
The foregoing implies that one should always attempt to reduce the viscosity of the
sludge as far as possible. For this purpose one may resort to various expedients such as:
1. raising the temperature of the liquid.

2. strong dilution.
3. affecting the liquid by chemical and/or colloid chemical reactions.
In order to go in an insight upon the relative importance of the changes in viscosity

brought about by a modification of the temperature or the concentration of solids-non-
oil, an experiment has been conducted in Sumatra based upon the viscosty of the non-
diluted sludge.
The determinations were carried out by Ir. van der Linden with the help of an Ostwald
viscosimeter which was calibrated with water of 50° C. The outflow period was 17.2
seconds, while water of 50° C. has a viscosity of .055 dp. Consequently, the constant of
the viscosimeter was:
0.055/17.2 =0.000323
The sludge was of the following composition:

dry matter = 11.80%
fat = 1.90%
non-fat = 12.90%
The observations taken were thrown into a tabular form, as follows:

Outflow Viscosity η of
Temperature periods of the sludge in Viscosity of
in oC sludge in secs. deci-poises water in dp
50 84.7 0.274 0.055
60 62.1 0.201 0.047
70 52.4 0.169 0.041
80 44.3 0.143 0.036
90 38.5 0.124 0.031
100 34.6 0.112 0.028
With the help of the formula log η = log a + x log b, in which a and b are constant
values and x represents the concentration of s.n.o. in the sludge, and basing ourselves
on the observations as given above, we can now calculate all the possible values of η.
x is expressed in per cent non-fat in sludge

a is equivalent to the viscosity of water, which can be ascertained easily; if we take
x as equal to 0, log η must be log a or η = a.
55
Now, if in the foregoing formula we let η represent the viscosity corresponding to the
temperature t, a the viscosity of water at the same temperature and x the analysed value
12.9% non-fat, it is possible to find the value b. This constant appears to vary slightly
with the temperature.
Temperature Constant value b

50 oC 1.133
60 oC 1.119
70 oC 1.116
80 oC 1.113
90 oC 1.113
100 oC 1.113
We now know all the constant values in the formula, so that we can vary the value x
and calculate the corresponding viscosities. This has been done in the following table:
Viscosity in deci-poises at:
% non-fat 50° 60° 70° 80° 90° 100°

0 0.055 0.047 0.041 0.036 0.031 0.028
1 0.062 0.053 0.046 0.041 0.035 0.031
2 0.071 0.059 0.052 0.045 0.039 0.035
3 0.080 0.066 0.057 0.050 0.043 0.039
4 0.091 0.074 0.064 0.055 0.048 0.043
5 0.103 0.083 0.071 0.061 0.053 0.048
6 0.116 0.092 0.079 0.068 0.059 0.053
7 0.132 0.103 0.088 0.076 0.066 0.059
8 0.149 0.116 0.099 0.085 0.073 0.066
9 0.169 0.130 0.110 0.094 0.082 0.074
10 0.191 0.145 0.123 0.105 0.091 0.082
11 0.217 0.162 0.137 0.117 0.101 0.091
12 0.245 0.182 0.153 0.130 0.113 0.102
129 0.274 0.201 0.169 0.143 0.124 0.112
The figures in this table have also been set out in a graph (Fig. 2) which indicates on a
semi-log scale the relationship between the percentage of s.n.o. in sludge and the
viscosity. Consequently, the lines are lines of constant temperature.
3.02.01 The effect of the temperature
The graph is quite self-explanatory. For optimum results the temperature should be
raised as far as possible and should be maintained throughout the period of settling.
This points to the desirability of insulating the settling tanks and covering the top. In
many oil mills, however, the top of the tanks is left uncovered and the enormous clouds
of vapour which are sometimes seen hanging about the clarifying tanks in such mills
56
are indicative of the large amoums of heat energy which are wasted on the surface of
the liquid.
Let us suppose that owing to circumstances only a temperature of 70° C is reached in

lieu of the 95° C required. Such may be the case if the crude oil pre-heater is unable to
give the liquor the temperature it should have, because the crude oil has been diluted
with too much and too cold water, or because the crude oil has cooled down too far
and/or the quantity of crude oil fed to the tank is excessive and/or there is not enough
steam and/or the steam trap is out of order etc. etc.
Let us assume, furthermore, that the concentration of the solids-non-oil in sludge is 4%.
The graph 2 now indicates that the viscosity of the sludge increases from .045 to .064
deci-poises, while graph 1 shows what effect this has on the rate of rise of the various
particles:
the speed is reduced or in

If d = from: to: percent:
40 micr. 200 micr./sec. 136 micr./sec. ~ 32%
30 micr. 110 micr./sec. 75 micr./sec. ~ 32%
20 micr. 50 micr./sec 34 micr./sec. ~ 32%
The influence of the reduced velocity or rate of rise upon the oil loss in sludge will be
quite clear to the reader, if he will kindly refer to what has been written on this point in
par. 3.01.03, so that we shall abstain from offering any further comment.
If it should be feasible to raise the temperature of the sludge to over 100oC, for example
by settling under pressure and in a closed tank, the reduced viscosity would permit a
proportionally greater exhaustion (de-oiling) of the sludge.
The correctness of this theory has been tested in practice by allowing exhausted sludge
ex clarifying station to settle in a sterilizer at 130°C. The sludge was then in effect
found to deliver some oil.
3.02.02 The effect of the concentration
As is indicated clearly in graph 2, the viscosity will become appreciably higher when
the concentration of the solids-non-oil increases. In the graph the concentration is
expressed in per cent non-fat, while no allowance has been made for the relation
between and the nature of the solids
This is debatable on theoretical grounds, inasmuch as the s.n.o. consist of soluble and
non-soluble or suspended particles and the viscosity of the sludge will vary with the
proportion these components bear to each other.
57
If we assume, however, that the relationship concerned is constant, which will be more
or less the case in practice, we cannot make a significant error in coupling the viscosity
of the sludge to the concentration of the entire group of solids-non-oil. Anyway, this
concentration should be kept as low as possible; in other words, the sludge should be
diluted to the maximum extent.
In diluting sludge, however, one is tied to certain limits.
First of all, the stronger the sludge is diluted the larger should be the size of the
clarifying tanks etc., and of course this cannot go on ad infinitum. Considerations of
finance and economy alone will keep in check any tendency to employ excessively
large equipment.
In the second place, the sludge contains suspended solids-non-oil, the specific gravity
of which, like that of oil, is lower than that of water. Hence these particles have a
tendency to rise to the surface when given the opportunity to do so.
Consequently, under certain circumstances the notorious ‘third layer’ may form
between the oil and the sludge. Once it is there, this layer will absorb much oil and will
impede the clarification. It is only possible to remove it by draining the clarifying tank,
which implies waste of oil, so that everything should be done to prevent it from being
formed. Nothing definite can be said as regards the viscosity limit beyond which the
third layer is apt to occur. If crude oil having a high percentage of s.n,o. is diluted with
water, the first faint tendency of the third layer to develop is noticeable at a
concentration of 4 to 5% s.n.o. in the sludge, this tendency becoming quite a nuisance
at a concentration of 3%.
If we set the limit of concentration at 3.5% non-fat, we can calculate the maximum
dilution of the sludge on the basis of the concentration of s.n.o, in the crude oil, at
which max. dilution optimum results will be obtained.
If one were to depart from the optimum by diluting the crude oil with a smaller quantity
of water, one would be at once confronted with increasing oil losses in sludge. The
concentration of the s.n.o. in the crude oil is a factor closely related to the operation of
the digester. When digestion is intensive and much oil is released, the oil loss in
pericarp fibre will be low and the concentration of s.n.o. in the crude oil or undiluted
sludge will be high. In the palm oil industry the concentration of the s.n.o. in the
undiluted sludge is sometimes designated as the Non-Fat Pressing Quotient (NFPQ).
This quotient indicates the total percentage of s.n.o. in the pericarp which has found its
way into the crude oil or undiluted sludge. The concentration of s.n.o. in undiluted
sludge and the Non-Fat Pressing Quotient are not fully identical, because both the
percentage of sludge in crude oil and the percentage of s.n.o. in the pericarp are
variables, which depend on the type of fruit handled. There is a relationship, however,
to the effect that the higher the Non-Fat Pressing Quotient the more s.n.o. there will be
in the undiluted sludge.
58
When the quantity of water used to dilute the sludge is kept constant, an increase in the
Non-Fat Pressing Quotient will imply an increased concentration of s.n o. in the sludge.
We have explained already that this will cause the oil losses in sludge to rise. In other
words, there will be an increase in the oil loss per unit non-fat or, to put it in a different
way, a higher percentage of oil to dry matter in sludge. To keep the losses as small as
possible, the crude oil should be diluted to the concentration permissible.
59
It will be clear that in this case the oil losses in sludge calculated by weight of the fruit
bunches are directly proportional to the ratio of s.n.o. in pericarp to bunch weight and
the Non-Fat Pressing Quotient. We shall illustrate the relationship between these
factors on the basis of figures collected in a palm oil mill in Sumatra:
% s.n.o in
sludge by
weight of oil loses in
% s.n.o. by non-fat bunches = sludge by
weight of pressing axb weight of
bunches = a quotient = b 100 bunches in %
9.01 13.2 1.19 0.27

8.99 13.5 1.21 0.28
9.12 13.5 1.23 0.30
8.85 15.0 1.33 0.36
9.44 14.6 1.38 0.40
9.10 16.9 1.54 0.53
9.50 19.8 1.88 0.63
9.36 20.8 1.95 0.71
We do not think these figures need any further comment, the relationship between them
being conspicuous enough. The only conclusion to be drawn from the table above is
that in order to recover the optimum quantity of oil from the sludge, the NFPQ must be
kept as low as possible. In other words, the release of s.n.o. in the digester must be
reduced to the strictest minimum. This is in diametrical conflict with what has been
written in the chapter on Digesters, and thus we find ourselves confronted with the
necessity of making one of those compromises which are so frequent in the palm oil
industry.
Now, the one essential point in considering the merits and demerits of a given method
of oil recovery is the total loss to s.n.o. in pericarp, for a part of the s.n.o. leaves the
processing plant in the form of pericarp fibre in the pressed residue, while the
remainder is discharged as s.n.o. in sludge.
Consequently, it is always necessary to test the soundness of any precaution that should
lead to lower oil losses in the pericarp fibre by verifying their effect upon the oil losses
in sludge. We shall give an example to illustrate this:
Of a given fruit variety the percentage non-fat in pericarp is 9.5% by bunch weight. The
fruit is processed in an installation and the resultant NFPQ is 12%. The oil loss in
pericarp fibre is 10% to dry matter non-fat. The oil loss in sludge is equally 10% to dry
matter non-fat.
60
One day the installation used is replaced by a new processing plant, in which the NFPQ
increases to 35%, the oil loss in pericarp fibre decreases to 8%, and the oil loss in
sludge increases to 20% to dry matter non-fat.
In the first plant the total loss as oil to bunch weight amounted to:
Loss in pericarp fibre: 0.88 x 9.5 x .10 = 0.836%

Loss in sludge: 0.12 x 9.i x .10 = 0·114%
Total loss 0.95 %
In the second plant the total loss is:
Loss in pericarp fibre: 0.70 x 9.5 x .08 = 0.582%

Loss in sludge: 0.30 x 9.5 x 0.20 = 0.57 %
Total loss 1.10%
Hence, although the new plant gives a smaller oil loss in pericarp fibre, this advantage
is more than offset by the higher loss in sludge sustained.
Summary
In the present issue we have discussed the natural laws which play a part in the process
of static clarification. It has been made clear that such expedients as heating and
dilution which are used to secure optimum clarification derive their importance from
those natural laws. In the next issue we shall deal with the question of sludge treatment
in centrifugals and of sludge treatment by the biological method.
Clarification II
Introductory note
In the preceding number we have gone at some length into the method of static
clarification and the factors affecting it. The present article will be devoted to a
discussion of clarification by the biological method and sludge treatment in
centrifugals. Furthermore, we shall give a brief description of the main components of a
clarifying unit.
1. The biological method
This method, which is based on the treatment of sludge from a clarifying station using
sludge tanks or conical tanks, derives its name from the fact that bacteria play an active
part in the recovery of the tiny oil droplets which are still present in the sludge, because
in the clarifying tanks they have not been able to reach the oil layer.
The main season why biological sludge treatment has gradually fallen into disuse is that
the quality of the oil recovered is not very good, while the expense involved is
61
comparatively high. Although with centrifugals it is not possible to achieve the same
degree of exhaustion of the sludge as with the biological method, the financial and
economic results obtained with them are such that they have come to supersede the
latter system.
1.01 The fat pit
The fat pit that is employed in biological sludge treatment consists of a large number of
wide, shallow concrete or brick-work basins or troughs which are mostly sunk into the
ground. The sludge runs through the troughs very slowly passing under the partitions or
over the thresholds which separate the successive compartments. Figure 1 gives a
longitudinal section of a series of fat troughs. The oil coming to the surface during the
process is retained by the partitions.
1.02 Oil recovery in the fat pit
It has been found that under normal circumstances some 70% of the oil introduced with
the sludge is recovered in the fat pit, and that there is a more or less constant loss of
30%. In consequence, the eventual loss in sludge is decided by the oil content of the
sludge coming from the settling station. This explains why with this method efforts
have always been made to keep the oil content of the sludge led to the pit as low as
possible.
62
Over 80% of the oil obtained in the fat pit is recovered in the first compartment.
As in this early stage there can be no question of any biological action, it is to static
clarification that the recovery concerned must be attributed. The troughs are wide and
shallow and therefore lend themselves ‘par excellence' to the recovery of smaller
droplets having a low rate of rise.
Even in the open gutter running from the mill to the fat pit the wonders worked by
shallow vessels are very clearly demonstrated, because oil will commence to separate
visibly already here.
Hence the sludge in the first section of the fat pit is subject to exactly the same rules
and phenomena as the sludge in ordinary static clarification. This includes the danger of
the so-called ‘third layer’ which may develop even more readily than otherwise, owing
to the particular shape of the troughs. If the sludge is strongly diluted, the layer of dirt is
apt to build up too rapidly and become so thick as to prevent effective oil recovery. In
the case the troughs should be drained and cleaned, which operation entails loss of
recoverable oil.
The remaining 20% of the oil recovered in the fat pit are obtained by combined settling
and biological action. Bacteria present in the sludge bring about a reduction of the
viscosity by fermenting starch and sugars. The resultant alcohols cause the necessary
decrease in surface tension so that the particles especially the smaller ones, have less
difficulty in fusing together.
As regards the biological action we shall be brief and confine ourselves to giving a
summary, which epitomizes the results of many large-scale experiments carried out by
a few researchers in Sumatra:
1. The layer of dirt appeared to have a very favourable effect, inasmuch as its
removal caused an appreciable deterioration of the oil recovery.
2. The layer of dirt was found to contain bacteria which upon examination, proved to
have the properties of producing acid awl forming gas up to a temperature of 50°
C. At 60° C no activity was apparent, from which the conclusion was drawn that
the bacteria were not thermophilous. Moreover, they were found to be anaerobe,
because aeration made the gas and acid production cease.
3. Acid and gas were also produced in the fat pit and this production could be
increased by lowering the temperatute in the pit. It was found that in other to
ensure a satisfactory growth of the bacteria especially the upper layer should be
cool.
4. The optimum oil recovery proved to be contingent on a maximum production of

gas and acid.
63
5. On determination it appeared that in all probability the bacteria in question were
of the type ‘Clostridium Butyricum’, and this made it possible to infer various
properties. Thus sugar and polysaccharides (as starch) appeared to be
indispensable nutrients. On examination the sludge was in effect found to contain
both sugar and polysaccharides, so that no poor bacterial growth for lack of a
suitable medium need be feared.
1.03 Drawbacks of the biological method
There are several drawbacks connected with biological sludge treatment, from a
financial and economic point of view as well as technologically. A great technological
disadvantage is that one is never quite sure of one’s ground. For substances having a
toxic effect upon the bacteria and inactivating them for as long as a couple of days may
quite easily enter open fat pit.
Moreover, the bacteria may be the cause of their own destruction. Clostridium
Butyricum is very wide spread and may even be found on palm fruit bunches,
especially in sultry, rainy weather. They penetrate into the fruit and form an anti-toxin,
with which the bacteria agglutinize themselves. Thus a gelatinized deposit will build
up. In all probability this explains why in rainy weather the contents of the fat pit often
coagulate, which tendency is even noticeable in the clarifying station. The fat pit
contents do not coagulate gradually but of a sudden and there is little or nothing to
prevent it. A fat pit in this state has a lifeless appearance: the production of gas has
stopped and the familiar foam bubbles on the surface have vanished. If under those
circumstances one does not want to run the risk of a permanent increase in oil losses it
will be necessary to drain the fat pit contents completely.
2. Sludge treatment in centrifugals
In the nineteen thirties many, if not all, palm oil mills were equipped with fat pits. With
the steady increase in throughput capacities the fat pit layouts became ever larger. In
some mills they even had the size of small lagoons or lakes, which were attended by a
large number of labourers. By and by they grew into cumbersome and impractical
affairs, and it was only logical that attempts were made to replace them by centrifuges
to separate oil from the sludge.
Although several engineering works have endeavoured to design suitable centrifuges

for the purpose, only very few have succeeded in contriving machines that have earned
themselves a lasting place in the palm oil industry.
64
2.01 Basic principles
Before proceeding to give a short description of the main types of centtriluges used we
would discuss the basic principles underlying the separation of oil from the sludge by
centrifugeing.
As is known, the radial acceleration in centrifuges is
V2/R in which V stands for the tangential speed and R for the distance to the axis of
rotation.
For the tangential speed we may also write V = n / 60 x 2 π R = nπR / 30
in which n represents the number of revolutions per minute.
Hence the radial acceleration V2/R may be written as n2 π2 R / 900
If we express R in cms we find the acceleration in cms/sec2. The acceleration of the

force of gravity is 981 cms/sec2. Consequently, the radial acceleration in the centrifuge
exceeds that of gravitation by (n2 π2 R) / (900 x 981) times
In order to give the reader an idea as to the magnitude of this factor, we have calculated
its value at various numbers of resolutions and radiuses, taking
π2 / 981 as equal to 10 / 1000 for convenience sake
65
Arranged in a table, the figures are as follows:
R in cms
n = r.p.m 10 20 30 40 50
800 71 142 213 284 355
1000 111 222 333 444 555
1200 160 320 480 640 800
1400 218 436 654 872 1090
1600 285 570 755 1140 1425
1800 360 720 1080 1440 1800
2000 444 888 1332 1776 2220
2500 694 1388 2082 2776 3470
3000 1000 2000 3000 4000 5000
What these figures signify will be clear without any lengthy explanation, now that by
the publications respecting space navigation almost everybody has become familiar
with such formulae as, for instance ‘acceleration 6 g’. This implies that an oil particle
whirling round in a centrifuge at, say, 20 cms from the axis of rotation at 1400 r.p.m.
has become 436 times heavier, which is also applicable to the sludge surrounding the
oil droplet and to the difference in weight between the two liquids. In other words, the
separating force between oil and sludge under these circumstances has come exceed
that prevailing in static clarification (1g) by 136 times. Armed with this knowledge, we
can now form a picture of the selectivity of a sludge centrifuge by making a rough
calculation.
We proceed from the assumption that the rotor of a sludge centrifuge is a star-shaped
body, with four conical tips the inside diameter f the tips being 325 mms at a distance
of 20 cms from the axis of rotation and 95 mms at 40 cms front the axis of rotation
(refer to Fig. 2).
Furthermore we suppose the speed of the rotor to be 1450 rev. per minute and the
viscosity of the sludge to be .04 deci-poises. Through the nozzles at the end of the tips
3000 lures of sludge are discharged every hour. On the basis of these data we arrive at
the following figures:
1. The radial acceleration at 20 cms from the axis is approx. 480 g.
2. The radial acceleration at 40 cms from the axis is approx. 960 g.
3. The flow surface in each conical tip is 82958 sq.mm at 20 cms from the axis.
4. The flow surface is 7088 sq.mm at 40 cms from the axis.
5. The quantity of liquor passing through each conical tip amounts to 750 litres/h or
208333 cu.mm/sec.
66
6. Consequently, the average flow velocity of the sludge in each conical tip is 2511
microns/ second at 20 cms from the axis.
7. The average flow velocity is 29392 microns/ second at 40 cms from the axis.
8. If, for convenience sake, we consider Stokes' law to be applicable to the case
under discussion, it follows from graph 1 in the preceding issue that an oil droplet
of 40 microns in sludge of .04 deci-poises has a rate of rise (or, under the
conditions prevailing in the centrifuge, a rate of speed directed towards the axis)
of 480 x 222 = 106560 microns/second at a distance of 20 cms from the axis of
rotation. A droplet of 20 microns has a speed of 480 x 56 = 26880
microns/second. When we compare these velocities with the average flow
velocity of the sludge at the point indicated (which is 2611 microns/second) it is
clear that the oil particles will tend to move towards the axis of rotation.
9. Similarly, we can calculate that an oil particle of 40microns, which, through

some cause or other has arrived at 40 cms from the axis, has a speed of 960 x 222
= 213120 microns/second directed towards that axis while for a particle of 20
microns this is 960 x 56 = 53760 microns/second. (Again, these velocities are
much higher than the average flow velocity of' the sludge at this point which is
29392 microns/second, so that the particles in question will not be discharged
with the sludge through the nozzles at the extremity of the conical tips.
67
From these figures it would follow that the conditions of sludge treatment as prevailing
in a centrifuge promise far better results than are obtained with ordinary static
clarification at 1 g. That this expectation is not completely fulfilled in practice has two
reasons, the first being that the law of Stokes no longer applies at the comparatively
high rates of rise or speed as indicated. The real rates of rise are lower than those we
found in our calculation.
The second reason is that the sludge does not run evenly through the conical fins of the
star-shaped bowl, but is in a state of turbulence, so that in places the flow velocity is
higher than calculated and the risk of oil particles being carried along with the sludge is
slightly greater. Experience has taught however, that with a well-designed centrifuge
the final oil losses is sludge are but slightly higher than those suffered in static
clarification followed by biological sludge treatment in a fat pit, while on the other
hand the quality of the oil recovered is appreciably better.
2.02 Types of sludge centrifugals
The sludge centrifugals available are of various makes and can be sub-divided into two
classes i.e. centrifuges separating oil from the sludge in the purest condition possible,
which oil added to the oil coming from the clarification tank is handled further without
any additional purifying treatment, and centrifuges separating sludge exhausted as far
as possible from a crude oil fraction that may go on containing a certain amount of
sludge.
After passing through the centrifuge, this crude oil is led back to the clarification tank
together with the crude oil from the oil recovery station. The centrifuges of the former
category are provided with control gear, because the separating surface of oil and
sludge should be kept at one point between the axis of rotation and the outlet nozzles.
For if more oil should be emitted from the centrifuge than is fed to it along with the
sludge, the separating surface would shift towards the axis and sludge would flow out
along with the oily phase at a given moment. Conversely, if less oil were to be
discharged than is contained in the sludge fed to the centrifuge, the separating surface
would shift in the direction of the nozzles and recoverable oil would pass out with the
sludge. Often the control gear used is of a very ingenious design. There are systems
utilizing the circumstance that sludge is a good conductor of electricity and oil a bad
one. With such systems two rings are placed in the bowl of the centrifuge, between
which a certain voltage is maintained.
The separating surface of oil and sludge is between the rings. When this surface has a
tendency to move to the axis, because too much sludge is led to the system or the
nozzles are jammed with dirt the resistance in the electric circuit will drop and the
intensity of current will increase. The impulse is used to narrow the passage of a valve
in the feeder line of the oily sludge, etc. etc
The sludge centrifuges of the other class, which include those manufactured by Stork,
do not require any control equipment. The nozzle diameter is so chosen that the
68
capacity of the centrifuge slightly exceeds the quantity of sludge to be handled. The
liquid level in the tank from which the centrifuge receives its sludge is kept constant by
means of hot water admitted through a float controlled valve. Hence the quantity al
sludge fed to the centrifuge is always exactly in keeping with the throughput capacity.
The discharge of crude oil — which oil constitutes a fraction of the total amount of
liquor admitted — is regulated with the aid of a simple control valve in the discharge
pipe, in such a way that a visible quantity of sludge is led of with the oil. Once the
centrifuge has been set in the manner described, both the amount of oily sludge
produced and its composition may vary within wide limits without the performance of
the machine being affected in any may.
As a matter of principle the latter type of centrifuge should be preferred to the other, on
the following grounds:
1. Popularly speaking, the machine only serves one master, i.e. it has the function of
separating sludge that has been exhausted as far as possible. The oil recovered is
kept within the plant, its actual separation being left to other equipment, namely
the clarification tank.
2. Through the absence of sensitive mechanical control gear the centrifuge gains in
simplicity and, above all, in dependability.
3. The equipment used in the clarifying station
In the foregoing chapters we have given attention to the factors affecting the efficiency
at the clarifying process (time of sedimentation, temperature of the liquid, dilution),
which factors account for part of the standards to be set as regards the clarifying
equipment employed. Next to this there are certain requirements connected with the
quality of the oil produced. For this point we would refer to what has been said in the
first Stork Palm Oil Review, Vol. 1 Nr. 1, page5, par. 2.01concerning Spontaneous
Autocatalytic Hyclrolysis.
Inasmuch as the crude oil to be clarified contains water in sufficient measure, while the
temperature of the crude oil is high, there is ever risk of the F.F.A. content increasing
during the clarifying process. Consequently, the equipment used should allow the
clarification to proceed as rapidly as possible. The same requirement should be imposed
in view of the colour and bleachability of the oil.
From the same paragraph we can furthermore infer the condition that the oil should be
dried to a moisture content of .08%, in order to prevent the F.F.A. percentage from
rising during storage and transport, as far as possible.
With all these requirements in mind, we shall now discuss briefly the machinery and
equipment used in the clarifying plant.
69
3.01 Vibrating screens
The crude oil coming from the recovery unit is first fed to a vibrating screen provided
with tightly stretched gauze of 16 mesh. The material retained by the screen and
composed mainly or fibre and small fragments of pericarp is led back to the digesters.
When a sludge centrifuge is incorporated in the clarifying unit, the oily sludge
discharged from the continuous clarification tanks is passed over a second vibrator
fitted with gauze of 45 mesh. By means of hot water sprayed on to the screen the oil
adhering to the matter retained is removed for the greater part, whereupon the
70
remaining dirt and fibre are ousted from the production process. Page 1 gives a picture
of a double vibrator for crude oil placed on top of a crude oil receiving tank, which also
serves the function of sand collector i.e. the sand passing through the mesh of the
vibrator along with the crude oil is allowed to settle down in this tank.
3.02 Crude oil pumps
To ensure adequate performance of the clarification tanks it is necessary that the flow
of crude oil admitted to them should be as even as possible. That is why double plunger
pumps are generally given preference over any other type.
3.03 Clarification tanks
The clarification tanks serve the purpose of separating from the sludge the free oil and
the coarse oil droplets which rise to the surface fairly quickly. Thus it is rendered
possible to handle rapidly the greater part of the oil recovered, for the residence time of
the oil in the tanks need not be long, as is illustrated by the results of tile following
experiment:
A given quantity of undiluted crude oil from fruit of the Deli-Dura type (extraction
ratio: approx. 17% oil to bunch weight) extracted in presses (non-fat pressing quotient:
approx. 17%) was clarified by the static method. Every 10 minutes a representative
71
sample was drawn from the sludge, its oil content determined anti expressed in % oil to
original bunch weight. The figures found are indicated in the graph of Fig. 4. If we
assume that the final oil content of the sludge was .3% oil to bunch weight, so that the
oil content of the crude oil must have been 17.3% to bunch weight, it can be calculated
that already after a settling period of
20 minutes 17.3 - 1 = 95.9%

17
of the recoverable oil had risen to the surface. After 40 minutes this would have been
17.3 – 0.7 = 97.6%,

17
after 60 minutes
17.3 - 0.58 = 98.3%

17
and after 2 hours
17.3 - 0.42 = 99.3%

17
The reader wi1l agree with us in that these figures are most eloquent. Fig. 5 gives a
diagrammatic view of a continuous clarifier as manufactured by Stork. The crude oil
pump forces an even flow of liquid to a heat exchanger and thence to the continuous
clarifier. The liquid lowly enters the tank through the central intake A, which is of
generous size and separates into oil and sludge. The oil collects in the layer X and
leaves the tank through the discharge pipe B. which is lifted with a foam trap.
The sludge collects ill the layer Y and is siphoned through the uptake C. In the position
sketched the two liquid fractions are in balance, i.e. the height of the oil layer X
multiplied by the specific gravity of the oil plus the height of the sludge layer Y
multiplied by the sludge gravity of the sludge corresponds to the height Z of the sludge
in the uptake multiplied by the specific gravity of the sludge. In other words the two
liquid fractions are in communicating equilibrium with the sludge in the uptake C.
This implies: a) that the amounts of oil and sludge emerging from the tank are the same
as the amount of crude oil fed in it and b) that the thickness of the oil layer X is
determined by the difference in height between the discharge openings for oil and
sludge (X + Y) - Z.
The difference in specific gravity between oil and sludge being about .1, it follows that
the thickness of the oil layer is 10 times the difference in height between the discharge
openings for oil and sludge. Thus X = 10[(X+Y)—Z].
72
The quantity of oil fed to the clarification tank and the volume of the oil layer together
determine the residence time of the oil in the tank. If the residence time is too short, the
said difference in height should be increased by lengthening the pipe B; if it is too long,
the pipe B should be shortens.
When no more crude oil is fed to the tank, the discharge of oil and sludge will stop at
the same time. If the clarifier is not in continuous service this may result in a
comparatively large amount of oil being kept at a high temperature for rather a long
period. To obviate this, one of the valves in the sludge uptake should be shut and water
admitted to the tank through the pipe D. The oil layer will then be pressed upward and
can be drained off almost completely. Next the valve in the sludge uptake C should be
opened again so that the liquid level in the clarifier will drop down to the discharge
opening for sludge. When after that the clarifier is put into operation again only sludge
will be separated initially until the oil layer has acquired the proper thickness.
Through the valve F in the conical bottom of the clarifier, sand and other foreign matter
should be drawn off periodically.
Further, the clarifier is equipped with a heater element in order to raise the temperature
of the tank contents whenever a prolonged period of standstill has caused them to cool
down too far. Another type of continuous clarifying unit used in palm oil mills consists
of a long rectangular tank divided into several compartments by partitions fitted
lengthwise. The partitions are so arranged that the liquor passes in turn beneath and
over them. The oil that is separated as the liquor runs through rises to the surface and is
retained by the partitions projecting above the surface. Sand and other heavy particles
will settle in the pyramid shaped bottoms and should be drained away from time to
time.
The crude oil is admitted at one end through a gutter with overflow rim. The sludge
leaves the clarifier at the other end. The oil recovered is skimmed off by means of
either adjustable or float controlled chutes. Figure 6 shows a diagram of this type of
clarifier.
3.04 Sludge tanks
If the continuous clarifier is not linked with a sludge centrifuge, the oily sludge is
passed to two or more sludge tanks. In these tanks, which are filled by turns, the sludge
may be boiled intensively, whereupon it is allowed to settle for a considerable period of
time, so as to enable the small oil particles to rise to the surface, After a certain period
the oil layer is skimmed off and the skimmed sludge drained away or led to a fat pit.
The tank is then ready to receive the next charge.
73
3.05 Sludge centrifuges
Figure 7 is a sectional drawing of the Stork sludge centrifuge, the working of which has
been explained in the foregoing chapter. Through the valve B the oily sludge is
admitted to the hollow shaft and enters the star-shaped rotor. The valve A serves the
function of filling the rotor completely with water before the centrifuge is started. Once
the machine has the proper speed, the valve A is shut and the sludge feed B opened,
after which the machine is ready for service and does not need any further supervision
or attendance. The oil recovered, which is mixed with some fibreless sludge, leaves the
centrifuge through the central outlet C. The sludge is whirled into the centrifuge
housing through the 4 nozzles at the extremity of the rotor, and is discharged through
the outlet E. The machine has a capacity of approx. 3000 litres of sludge/hour. Practice
has shown the final oil loss in the skimmed sludge to be 7% oil to dry matter non-fat.
3.06 Oil drying equipment
The oil that is recovered in the clarification tanks and /or the settling plant contains a
small quantity of moisture, which for the most part is present in tiny droplets as ‘free
water’, while a minor proportion is dissolved in the oil.
74
The observation taken to the effect that if oil is left to stand in storage tanks for a
considerable time a certain amount of moisture will settle out, clearly demonstrates the
presence of this ‘free water’. The wet oil has a moisture content which as a rule does
not exceed .5%. The standard adopted by the palm oil trade is a max. moisture content
of .20%. Since it has been found, however, that a moisture content of .20% plus the dirt
that is dissolved in the water are mainly responsible for any increase in acidity during
storage and transport, the general tendency is to reduce moisture to a strict minimum.
Most probably the lowest moisture content to be reached in practice is some where
around .06%.
Although successful attempts have been made to reduce the percentage even beyond
that limit, it is to be feared that the oil will be slightly hygroscopic whenever the
moisture content drops below it. In other words, the oil will attract moisture from the
atmosphere until a balance has been attained, which is in the neighbourhood of .06%.
Of the many types of drying equipment that have found application in the industrial
recovery of palm oil, we shall now pass a number in brief review:
Atmospheric boiling pans
These are large bulky vessels having a content of approx. 1650 litres and provided with
a bronze heater element for steam of 2 to 3 atmosphere. The operation is discontinuous.
Charging and discharging take some 5 minutes and the drying (or boiling) period is
approx. 30 minutes. The final moisture content obtained is .10 to .l l%.
Vacuum boiling pans
These are also large and cumbrous vessels for discontinuous service. The hot, wet oil is
admitted to the pan, whereupon the vapour pressure above the oil is reduced to approx.
.3 atm. absolute pressure in some 30 minutes, the liquor being stirred constantly all the
while. Next the vacuum is eliminated and the oil discharged. The final moisture content
is .08 to .09%.
Continuous vacuum dryers
Dryers of this type, which are manufactured by Werkspoor Amsterdam, feature steam
jet vacuum pumps and down-flow condensers. They project above the roof of the palm
oil mill and in practice reduce the moisture content to .07 or .09%.
Continuous vaccum film dryers
In these units a thin film runs along a steam heated cylindrical body. The vacuum is
maintained by rotary vacuums pumps.
The comparatively small dryers, which are of compact build, give the oil a final
moisture content of .08 to .09%.
75
Centrifuges
As a rule, the final moisture content of oil dried in centrifuges is not lower than .18 to
.20%.
In spite of the great variety of oil drying equipment available, the last word has not yet
been said on the subject. Viewed in the light of the latest developments, the method of
drying oil in one of the apparatuses enumerated above might even be characterized as
needlessly complicated. In point of fact, it has been found that the free water in the oil
will evaporate quite easily if the oil is finely distributed and a medium is used that will
absorb and remove the vapor formed.
The following observations made in practice have led to the discovery of this principle:
— When in a mill using ‘Alva Laval’ separators the gauge glass was removed from the
separators, because it was continually misted over, clouds of vapour were seen to
escape from the opening made and the moisture content dropped from .22% to .14%,
for the sole reason that the oil, which was whirled into the separator housing as a thin
film, was exposed to the atmosphere for a few moments.
— A tentative experiment has been conducted, in which hot oil was sprayed out as a
fine rain into a weak, warm current of air. This reduced the moisture content from .30
to .05%.
On careful and repeated examination of several samples no change could be observed

either in the quality or the characteristics of the oil thus dried.
On the basis of the new principle Stork have evolved a simple continuous oil dryer, in
which the oil is raised quickly to 125/130°C by a heat exchanger. Next it is passed over
a heated cascade, where it is drawn out into thin films hanging down freely. No figures
are as yet available concerning the results obtained in practice with this new apparatus,
but we expect the final moisture content to be.06 to.08%.
3.07 Filtration
The final process the oil is subject to consists in removing the small amount of
suspended dirt it contains after clarification.
This can be done either in a filter press or in an oil centrifuge (simultaneously with the
drying process). In both cases the results obtained are excellent. Provided the right type
of filter cloth is used, it is possible to reach lower figures (down to .005%) with filter
presses than with centrifuges.
76
The Oil Extraction Efficiency
Composition of the fruit
The composition of the fruit bunch is not only determined by the variety of oil palm,
but also by climatic conditions, the nature of soil, the age of the individual palm and the
time at which the bunches are harvested.
Although as a rule the oil o fruit bunch percentages of Dura fruit and Tenera fruit are
such as to mark the difference between the two varieties, it is not feasible to define the
composition of the bunches of each variety once for all by a few characteristic figures.
For example, the variations in the fruit to bunch, nut to bunch and fibre to bunch ratios
within each type are too numerous and too great to lay down fixed proportions
applicable to one variety of oil palm alone.
In this issue we have pictured a number of block diagrams symbolizing the composition
of a given crop. The diagrams have been drawn up with the aid of comprehensive series
of figures collected in practice, so that they are rooted in reality.
The diagram of figure 1 shows that after sterilization and threshing 10,000 kgs of
bunches are broken into 3000kgs of empty bunches, 1050 kgs of evaporated moisture
and 5950 kgs of fruit and calyx fragments. The 3000 kgs of empty bunches consist of
750 kgs of moisture and oil free matter and 2250 kgs of moisture. The mass of
sterilized fruit comprises 2900 kgs of nuts and 3050 kgs of pericarp. Again, the 2900
kgs of nuts may be broken down into 580 kgs of kernels and 2320 kgs of shell, or more
specifically, 532 kgs of moisture-free kernels, 2020 kgs of moisture-free shells and 348
kgs of moisture. The pericarp contains 955 kgs of water, 525 kgs of solid-non-oil
(mainly fibre), and 1570 kgs of oil.
With the aid of these block diagrams and the oil and dry matter balance sheets, it is
possible to give a picture of nature an the magnitude of the various loose and of the
correlation between them,
The oil and dry matter balance sheets
In planning effective process control, it would be logical to proceed from the quantity
of oil contained in the crop that is introduced with the mill. On determination of the
amount of oil recovered and the various process loses (known lessons) it would be
possible to find both the oil extraction efficiency and the unknown losses.
As it is, drawing a representative sample of fruit bunches of analyzable size is not

feasible in practice. Indeed, even fresh fruit or sterilized fruit does not lead itself for the
purpose. In this reason, it is usual in the palm oil industry to base the mill control
system on the total oil losses sustained, it being assumed by definition that the quantity
77
of oil produced added to the total of known losses corresponds to the amount of oil that
was present in the fresh fruit bunches.
It is obvious, therefore that all the losses should be determined with the greatest
possible accuracy. The oil loss in pericarp fibre is one of the figures that stand in the
focus of attention.
Although in every day practice this loss is often regarded as a criterion by which to
judge the efficiency of the recovery process, the oil loses in condensate, in empty
bunches, on nuts and in sludge together form too notable a figure with respect to the
loss in fibre to warrant such a notion. The oil and dry matter balance sheets to be
discussed presently give a picture of the various losses involved, and they reflect
faithfully the values that have been found in practice.
During the milling process, the fruit mash may lose moisture through evaporation or
absorb moisture owing, for example, to dilution of the crude oil. In consequence, it is
very difficult, if not impossible; to indicate with a reasonable degree of accuracy how
the moisture originally contained in the fruit mash is divided between the pericarp fibre
and the sludge during the process. The moisture content of the pericarp fibre indicated
in the oil and dry matter balance sheets is in reasonable harmony with the figures found
in reality. The remaining moisture has been entered as a sludge component by way of
balancing figure, and, therefore, has no bearing whatsoever upon the values found in
practice. This does not affect the validity of the balance sheets, insomuch as it is only
the ‘oil’ and the ‘solid-non-oil’ we are interested in. The balance sheets are based upon
10,000 kgs of fresh fruit bunches. The quantities fed to and discharged from each
station are always in equilibrium. For instance, looking at the balance sheet of fig. 1 we
notice that 10,000 kgs of F.F.B. are fed into the sterilizing station, where 1050 kgs are
lost by evaporation, while 6 kgs of oil run off with the condensate. It follows that the
quantity of sterilized bunches conveyed to the threshers is 10,000 – 1050 – 6 = 8944
kgs.
In the threshers or bunch strippers these 8944 kgs are split up into 3055 kgs of empty
bunches – in which 55 kgs of it are carried off – and 8944 – 3055 = 5889 kgs of fruit
and calyx fragments. Arrived in the centrifuge, the mass of 5889 kgs of fruit is again
broken down into 4077 kgs of residue and 1812 kgs of crude oil. The depericarper, in
its turn, divides the 4077 kgs of residue into 2929 kgs of nuts, on which 29 kgs of oil
are lost, and 1148 kgs of fibre containing 470 kgs of solids-non-fat and 112 kgs of oil.
Finally, the 1812 kgs of crude oil recovered are conducted to the clarifying plant, where
they are split into 1362 kgs of pure oil and 450 kgs of sludge, which holds 55 kgs of
solid-non-fat and 6 kgs of oil.
Now, if we add the total amount of oil lost (208 kgs) to the amount of oil produced
(1362 kgs), we find that the F.F.B. must have contained 1570 kgs of oil initially. This is
equally indicated in the block diagram of Fig. 1. In a similar manner we can find the
total mount of solid-non-fat in the pericarp by adding together the 480 kgs s.n.o in the
78
fibre and the 55kgs in the sludge, etc and so on. With the help of these figures, which
are mutually related and in balance, it is possible to find others, which are of
importance in the mill control system and/or which may illustrate the efficiency of the
recovery process employed, as shown in Figure 1.
Comparison of Results
Extraction Centrifugals Versus Presses
When we compare the oil and dry matter balance sheets of Figures 1 and 2 the first
thing that arrests our attention is the higher oil losses in the sterilizing station in Figure
1. This is because the balance sheet in the question is based on a mill using vertical
sterilizers.
The reader will be familiar with the fact that both the oil losses in condensate and the
amount of oil absorbed by the stalks of the fruit bunches are higher with vertical
sterilizers than with horizontal ones. (The why and wherefore has been explained in
Vol. 1 Nr/ 3). The balance sheets under notice once more clearly demonstrate the point.
As a matter if course the difference in efficiency between an extraction station equipped

with centrifuges and one using presses has nothing to do with the losses suffered in the
sterilizing station.
For centrifugal extractors the losses in fibre and on nuts have been computed at 141 kgs
of oil per 10,000 kgs of fruit bunches and for presses at 56 kgs of oil per 10,000 kgs of
fruit bunches. These low figures cannot be attained, however, without the optimum nut
to pericarp ratio, to secure which the expedient of re-cycling the fibre may be resorted
to in the case under review.
The lesson to be derived from the comparison drawn is that provided endeavours are
made to approach the optimum nut to pericarp ratio as nearly as possible, hydraulic
presses may be expected to give a better oil extraction efficiency than centrifuges.
The additional oil recovered is not only the result of lower losses in fibre, but also of
lower losses on nuts, and it is especially this latter point which is liable to be
overlooked.
Fruit Low in Oil Versus Fruit High in Oil
A comparison of the oil and dry matter balance sheets of Figure 2 and 3 shows that
notwithstanding completely identical recovery methods and losses that are equal in a
comparative sense, the fruit with the highest oil content also give the highest extraction
efficiency. The higher percentages of fibre to fruit bunch (Fig.3), causes a higher
effective oil loss in fibre and in sludge. The lower percentage of nut to fruit bunch – at a
constant loss of oil on nut to nut – results in a lower effective loss.
79
The conclusion to be drawn from the comparison is that the extraction efficiency also
depends upon the original oil content of the fruit bunch. As a general rule, a high
extraction efficiency is concomitant with a high percentage oil to fruit bunch;
conversely, the oil extraction efficiency is reduced as the oil/fruit bunch percentage is
lower.
The Effect of the Fibre to Bunch Ratio
When we compare the oil and dry matter balance sheets of Figures 3 and 4, we find at
once that the rule laid down in the preceding paragraph is certainly not without
exception. Despite the higher oil to bunch percentage, i.e. 25% instead of 21% and oil
losses which, comparatively speaking, are in the same order of magnitude, the
extraction efficiency is 92.72% in case 4 as against 94.33% in case 3, in other words, it
is lower. The cause lies in the higher percentage of pericarp fibre to bunch, i.e. 13.25%
instead of 7%.
The reader might suspect us of having chosen rather extreme instances, occurring
rarely, if at all, in practice, when comparing the balance sheets here discussed. We can
assure him, though, that the block diagram of Figure 4 is a faithful representation of the
results obtained with certain T/P cross-varieties, which are now grown and processed
on a fairly large scale in Africa and which we have contrived to collect accurate data of.
Low N.F.P.Q. Versus High N.F.P.Q.
In the issues dealing with digesters and clarification we have dwelled at some length
upon the term Non-Fat-Pressing-Quotient and what it stands for.
The advantages of a high N.F.P.Q. are: lower oil losses in pericarp residue and/or a
higher throughput capacity per press unit. The drawback is that high N.F.P.Q.
inevitably entails higher oil losses in sludge.
A high N.F.P.Q. need not necessarily result from (over) intensive digestion alone; it
may be caused as well by (over) intensive agitation in the press cage. The particular
instance we have in mind in this connection is the screw press, for, owing to the high
wall pressures and the movement of the worm or screw, the amount of solids-non-oil
forced out through the perforations in the cage of this press exceeds that released from a
press of the hydraulic type.
A comparison of the oil and dry matter balance sheets of Figures 4 and 5 once more
reveals the fact that a high N.F.P.Q. implies reduced oil losses in fibre and increased
losses in sludge.
Now, if only the lower losses in fibre are taken account of (which, unfortunately,
happened all too frequent in practice, because the oil losses in sludge are more difficult
to ascertain), the opinion that a high N.F.P.Q. is synonymous with a high oil extraction
efficiency may easily come to prevail. One glance at the balance sheet is enough to
80
disabuse anyone of that idea. True, the loses in fibre are reduced considerably, but there
is an even more notable increase in the losses in sludge. The ultimate result is that per
10,000 kgs of fruit bunches of the composition indicated, 22 kgs of oil are loss
additionally. This is no adequate reason in itself to condemn the recovery method with
a high N.F.P.Q., for a proper judgement it is necessary also to weight the advantage of a
high throughput capacity per press unit (implying lower process costs) against the
drawback of increased oil losses.
For an illustrative calculation we proceed from the assumption:

a. that, for example, a continuous screw press has a throughput capacity equivalent to
12 tons of bunches per hour;
b. that a day’s wages of one mill hand are equivalent to the value of 5 kgs of palm oil.
Per shift of 8 hours it is possible to handle 8 x 12 = 96 tons of bunches. On the basis of

the differences between the oil and dry matter balance sheets of Figures 4 and 5, we
find that in a plant having a high N.F.P.Q. (some 30% in the case under discussion), the
oil loss sustained per shift will exceed by:
96 x 22
10 = 211.2 kgs that in a plant having a low N.F.P.Q. (16% in this case).
These 211.2 kgs are equivalent to some 42 day’s wages. Consequently, in economic
grounds and despite the higher losses the recovery method with a high N.F.P.Q. (in our
example based on a screw press) should be given preference over the other if at a
throughput capacity of 12 tons of bunches per hour, 42 or more mill hands could be
dispensed with.
Inasmuch as 12 tons of bunches/hour are easily handled in 4 normal hydraulic presses

usually requiring 4, and at the outside 6 attendants, we need not give any further
comment concerning the outcome of this calculation.
What the comparison teaches us is that in general a high N.F.P.Q. tends to affect the
process results adversely. It will never do only to heed part of the oil losses (more
especially the losses in fibre) and to ignore the rest, if one wishes to achieve effective
mill control.
81
Oil & Dry Matter Balance Sheet of a Palm Oil Mill Using
Vertical sterilizers, digesters without perforation and static clarification
3000 empty bunches 750 dry matter 580 kernels 532
1050 water 2250 water 2320 shells 2020
5950 fruit & calyx 955 water 348 water

fragments 2900 nuts
525 solid-non-fat
3050 pericarp
1570 oil
- loss of moisture by evaporation 1050 kgs

F.F.B.
10,000 kgs Sterilization oil in condensate 6 kgs 6 kgs
- Sterilized fruit bunches 8944 kgs
Threshing empty bunches 3055 kgs

oil loss in empty bunches 55 kgs 55 kgs
- fruit 5889 kgs
Centr extractor
Crude oil 1812 kgs
- Pericarp residue 4077 kgs
oil 112kgs
Depericarper fibre 1148 kgs water 566 kgs 112 kgs
solids-non-fat 470 kgs
Nuts 2929 kgs oil 29 kgs

Nuts 2900 kgs 29kgs
Clarification Sludge 450 kgs oil 6 kgs

Solids-non-fat 55 kgs 6 kgs
Water 389 kgs
1362 kgs oil produced Oil loss 208 kgs
82
Oil extraction efficiency total ………………….. = 1362 : 1570 = 86, 75 %
Oil extraction efficiency centrifugal extr ………. = 1368 : 1509 = 90, 66 %
Oil extraction efficiency clarification ………….. = 1362 : 1368 = 99, 63 %
Oil loss in fibre to dry matter …………………... = 112 : 582 = 19, 24 %
Oil loss in sludge to dry matter ………………… = 6 : 61 = 9, 83 %
Oil loss in condensate to fruit bunch …………… = 6 : 10000 = 0, 06 %
Oil loss in empty bunch to fruit bunch …………. = 55 : 10000 = 0, 55 %
Oil loss on nut to nut …………………………… = 29 : 2929 = 0, 99 %
Non-fat-pressing-quotient ……………………… = 55 : 525 = 10, 48 %
Figure 1
83
Horizontal sterilizers, digesters with perforation, hydraulic presses and static

clarification

fragments 2900 nuts
525 solid-non-fat
3050 pericarp
1570 oil

F.F.B.

- fruit 5909 kgs
Hydr. Press recycling

Crude oil 2097 kgs
oil 44 kgs
Depericarper fibre 900 kgs water 446 kgs 44 kgs

Nuts 2900 kgs 12 kgs

Water 545 kgs
84
Figure 2
85

clarification

fragments 2330 nuts
700 solid-non-fat
4070 pericarp
2100 oil

F.F.B.

- fruit 6359 kgs
Hydr. Press
Crude oil 2830 kgs
oil 59 kgs
Depericarp fibre 1189 kgs water 535 kgs 59 kgs

Nuts 2330 kgs 10 kgs

Water 735 kgs
86
Figure 3
87

clarification

fragments 1210 nuts
1325 solid-non-fat
5490 pericarp
2500 oil
F.F.B.

- fruit 6657 kgs
Hydr. Press
Crude oil 3317 kgs
oil 112 kgs

solids-non-fat1113 kgs

Nuts 1210 kgs 5 kgs

Water 765 kgs
88
Figure 4
89
Horizontal sterilizers, digesters and/or presses with high NFPQ and sludge centrifugals

fragments 1210 nuts
1325 solid-non-fat
5490 pericarp
2500 oil

F.F.B.

- fruit 6657 kgs
Press
Crude oil 3652 kgs
oil 83 kgs

Nuts 1210 kgs 5 kgs

Water 885 kgs
90
Figure 5
91
Depericarpers
General
In order to give the reader a clear insight into the various methods which in the course
of years have found application in separating nuts from fibre, we think it expedient to
recall to mind what has been written in Vol. 1. Nr. 5 dealing with digesters, i.e.:
In the early years, the recovery of palm oil gave great difficulties. Not only was there a
considerable loss of oil in the fibre, but the oily residue from the presses caused no end
of trouble in the further stages of the process. It could not be treated effectively in the
depericarper, with the result that the loss of nuts was greater than strictly necessary.
Moreover, an appreciable amount of oily fibre remained attached to the nuts, which
fibre formed a cake on the wall of the nut crackers so that they had to be put out of
service and cleaned very frequently. The fibre present in the cracked mixture fouled the
claybath in the kernel and shell separators, causing excessive loss of kernels. Further,
fibre particles were carried off with the kernels and lastly they spoiled the appearance
of the final product.
Naturally the attempts to cope with the difficulties encountered in those days were not
made in the pressing station, but whenever they were likely to produce some effect, i.e.
also in the depericarpers. The major influence of digestion upon the pressing effect had
not yet been discovered. As a rule, the pericarp was insufficiently digested, and
sometimes the mash contained fruits the flesh of which had hardly been bruised and
continued to envelop the nut completely.
It is obvious that under these circumstances it was very hard to obtain clean nuts and
that the nuts were made to travel a long distance in the depericarper and were subjected
to a certain amount of friction.
By the fricative action, i.e. by causing the nuts to rub against each other and against the
depericarper wall, it was endeavoured to loosen the fibre from the nuts.
As the pericarp fiber had a high moisture content as well as a high oil content, it was
only too prone to cling to the nuts and to the depericarper wall and thus to cause a great
deal of awkwardness. It is again quite obvious that better results were obtained when
the pressed residue was subjected to drying. As a consequence of improved
sterilization, new developments in digester design, more accurate checks on digestion, a
better pressing effect etc., the make-up of the press cake had changed. Already in the
digester the nuts are detached from the pericarp, which owing to improved sterilization
methods is softer (better ‘cooked’). As well, the oil content and – no less important –
the moisture content of the press cake have been drastically reduced in the course of
years so that on the one hand the recovery equipment in existence is better able to fulfil
its task and/or has been simplified, while on the other hand it has become possible to
evolve new machines and add them to those available.
92
Depericarpering Methods
The various types of depericarpers can be divided into three distinct main groups:
1. Hydraulic depericarpers.
2. Pneumatic depericarpers.
3. Mechanical depericarpers.
Hydraulic Depericarpers
Essentially, hydraulic depericarpers are washing machines in which the pressed residue
is washed out with a generous amount of water. In the process the fibre will float and
run off with the water whereas the nuts settle on the bottom. In its simplest form, such a
depericarper consists of a concrete on brickwork gutter with water running through.
The pressed residue is dumped into the gutter and is loosened and spread with the aid of
rakes handled by labourers standing in the water. The fibre is carried off with the water
current and caught against a strainer placed down stream. The nuts sink and are
scooped out at regular intervals.
A mechanical type of hydraulic depericarper is composed of a through in which a shaft

fitted with arms loosens the cakes and washes them out. The fibre-laden water is
pumped away, for example to the boiler house, where the fibre is separated out by
means of strainers and the water led back to the washer. The nuts are removed in an
even flow by, for instance, paddle wheels.
This type of depericarper has the following advantages:

1. Suited for wet and oily residue.
2. Great simplicity of design.
Its drawbacks are:

1. The manually operated type is dependent on the presence of running water;
2. The mechanically operated type gives high labour costs;
3. The mechanized type has a comparatively high power consumption;
4. Fibre and nuts are delivered wet; moreover, oil which otherwise might have been
utilized for fuelling purposes is lost with the water passing through.
Hydraulic depericarpers have been very popular in the palm oil recovery industry and
are hardly applied today. They are only found in those rare mills which still use the
obsolete and practically abandoned method of recovering palm oil by washing the fruits
in the digesters with large quantities of hot water.
Pneumatic Depericarpers
Although there is a considerable variety of design, all pneumatic depericarpers are

basically the same. They consist of a long rotary drum with baffles mounted on the
inside of the periphery to take the pressed residue upward and drop it again. A current
93
of air, either hot or cold, is passed through the drum and this is just strong enough to
carry the loose fibre off against the force of gravitation, the nuts remaining under the
influence of this force.
With these depericarpers the press cake is broken up and loosened in advance in so-
called cake breaker conveyors. The hot cake is here allowed to evaporate and to dry,
steam and jacketing being provided to assist the process.
The adjustable paddles on the shaft of the conveyors are set in such a way that the run
of the press cake to the depericarper is as long as possible and the fibre is completely
detached from the nuts on arriving at the inlet.
With the Stork depericarper, also called drier-depericarper (refer to fig 1), the press
cake is dried further in the top drum through an inlet which is conical so as to avoid
excessive leakage of air.
The fibre and the nuts both move in the same directions, towards the end of the drum.
The position of the baffles at the inside of the drum periphery controls the retention
time of the residue in the machine.
Arrived at the end of the drum, the nuts will drop by their own weight and enter the
polishing drum, while the fibre is carried off with the air stream to a cyclone situated at
some convenient point in the boiler house and serving the functions of separating the
fibre from the air. In the polishing drum the nuts are rubbed against each other and/or
the wall for a considerable time in order to loosen any fibre particles adhering to the
shell. A weak unheated current of air is conducted through the drum and entrains these
particles, joining the main current at point A. A slide controls the proportion between
the two air volumes and, with it, the temperature of the drying air.
The drying process in the depericarper has the advantage that even press cake that is too
wet and/or too oily for the cake breaker conveyor to give it a thorough preliminary
treatment, can be handled successfully and without a hitch. There is a subsidiary good
point in that drying increases the heating value of the fibre.
The drawbacks are among other things the comparatively high steam consumption
(which is only partly offset by the higher moisture value of the fibre), and the fact that
the final moisture content of the fibre is subject to variations and is largely determined
by the load on the depericarpers.
The variations in the final moisture content will cause a change in the natural slope, i.e.
the drier the fibre the smaller the angle of the slope. This implies that with boiler
furnaces incorporating step grates the incline of the grates should be variable in order to
meet any changes in slope of dried fibre. On the other hand such a variation in incline is
only possible within well-defined limits.
94
We have here one important reason why preference is sometimes given to ‘cold’
depericarpers, in which the air required to separate the fibre from the nuts is not heated
beforehand.
To these machines which as a rule consist of a single long drum set at a slight incline,
the pressed residue is admitted at the high end. The nuts leave the drum at the low –
open – end, and the air entering at that point emerges at the high end, carrying the fibre
with it (refer to fig 3). In other words, the nuts and the fibre move in opposite
directions, thus permitting separation over a longer distance. The theory that this
system causes greater selectivity, i.e. that the line of division between fibre that is
removed and the smallest nut that just falls short of being carried off is more accurate,
has never been proved in practice.
The placing of the fan outside the dust stream (i.e. either as delivery fan infront of the
machine or as suction fan after the cyclone) instead of within that stream (between
depericarper and cyclone) has often been a subject of discussion and every possibility
of location has been tried out in practice.
Placing the fan outside the dust stream has the advantage that modern fans of high
efficiency can be employed, while wear on housing and impeller is practically
excluded.
The drawback of arranging a fan infront of the depericarper is that the system comes
under over-pressure. It being difficult to keep the connection between the stationary air
duct and the rotary depericarper drum reasonably tight, a veritable dust-nuisance may
be the result.
The problem attendant upon placing the fan after the cyclone is how to make the duct
discharging the fibre from the cyclone sufficiently leak-proof. Failing this, disturbances
are bound to occur throughout the system. Furthermore, the difference between air
suction and delivery is that with the former any leakage will as a rule, favourably affect
the air to fibre ration and increase the rate of speed of the air towards the fan, both of
which factors are conducive to the pneumatic conveyance.
On the other hand, with the fan in the delivery position any leakage in the system, apart
from entailing the risk of dust-nuisance, may cause a change for the worse in the air to
fibre ration and may reduce the rate if speed of the air in the duct. In other words the
same two factors now have an inhibitory effect on the pneumatic transport. The system
most frequently adopted is to arrange the fan between depericarper and cyclone. In that
case underpressure will prevail in the depericarper which it is so difficult to keep
adequately dust-tight, whereas there will be over-pressure in the delivery duct and the
cyclone so that the latter needs no fibre sluice and the awkward problem if keeping this
leakproof is dispensed with.
The latter arrangement, however, is not without taking its toll, for obviously when the
fan is placed within the dust stream it will be exposed to the abrasive action of the fibre
95
and the sand particles it carries along. The – often considerable – wear and tear this
caused makes it necessary to install a fibre fan of such a design that the impeller blades
and the curved wall of the fan housing, which will suffer the most from the erosive
action, are easily renewable. For that purpose certain concessions are inevitable, which
results in a lower efficiency and, consequently, a higher power consumption than with
the two other systems.
Pneumatic depericarpers, which now find general application in large and medium-
sized palm oil mills, are available for almost every output. In their present form they are
sure to give trouble-free service under almost any circumstance, and their separating
effect is very good, not to say excellent. They have one avowed shortcoming in that
they are rather large and take a great deal of horse-power in comparison with other
units in the mill (14 to 18% of the total power consumption).
Mechanical Depericarpers
As regards mechanical depericarpers there exists no great diversity of design. As a

matter of fact there are only two main types, the first and most frequently employed
being the screening drum in which it is attempted to screen out the fibre and thus to
separate it from the nuts. If the press cake has received adequate pre-drying in a cake
breaker conveyor so that the fibre is reasonably dry and detached from the nuts on
entering the drum, a fairly good result may be expected.
As with all screening drums the capacity depends largely on the size of the screen
meshes. That is why they are better suited for palm fruit having a large nut and a
comparatively thin layer of pericarp than for fruits having a small nut and a thick layer
of pericarp. The capacities are low with respect to the size of the screens so that the
screen apertures are generally chosen as large s possible, which sometimes entails
considerable losses of small nuts in the fibre fraction.
The mechanical drum depericarper marketed by Stork has an octagonal section and its
sides are fitted with expanded metal. With regard to the direction of rotation this
material is so arranged that the screen gauze acts on the nuts like a rasp, i.e. the sharp
edges of the screen meshes rub, as it were, the fibre off the nuts. This is pictured in fig
4 and 5.
The second type can be clearly recognized in all its forms and varieties as a
modification of the cotton ginning machines, the fibre is pressed or drawn through a
narrow aperture, which is of such a size as to preclude damage to the nuts. Generally,
these machines have a low output, while the separating effect is indifferent.
The depericarper designed and patented by Werkspoor-V.M.F is a most favourable

exception to the rule. This machine, which is rather small considering its capacity
(approx. 4 metres long, 2 metres wide and 3 metres high – refer to fig 6) has a rotary
cage whose wall, is formed by rollers in pairs revolving in opposite directions. The
rollers are of such a diameter that they have no grip on the nuts but remove the fibre to
96
the outside of the cage (refer to fig 7). Owing also to a special surface treatment
imparted to the rollers, the separation of nuts and fibre obtained is all but flawless. The
capacity is high, corresponding as it does to 15-18 tons of bunches per hour. The power
consumption is low (max. 15 H.P.) and does not exceed 4 to 6 % of the total energy
requirements of the mill.
However, as we all know, absolute perfection is rarely attained in the sphere of human
invention. So too the machine in question has its weak spots. For one thing it is the
problem of conveying the fibre from the depericarper to the boiler house, for another,
the large number of pivoting points are apt to the arouse fears as to the possibility of
effective upkeep. That is why in spite of its undeniable qualities this depericarper has
never really stood in great demand.
Replies to readers
The two numbers of our reviews dealing with clarification have induces several of our
readers to write in for our explanation of certain points connected with the matter. As
most of the questions asked were either identical or of similar purport, we are pleased
to devote some space in the present issue to a discussion of what has aroused the
interest of many.
Questions
a) One reader wrote that it is not illogical to expect the effect of the separation of
water and dirt in a purifier (centrifuge) to depend among other things upon the
load on the purifier. For a smaller supply of palm oil to the purifier will prolong
the retention time of the oil in the purifier bowl and thus increase the possibility
of more water and dirt being removed.
In practice, however, no support can be found for this theory. Indeed, whether the
machine runs at ¼ load or ½ load or full load, a fairly constant final moisture content in
the order of 0.25% is measured.
The question laid before us was whether we were able to account for this phenomenon.
B. A question posed by several readers can be summarized as follows: Palm oil when
dried intensively will be hygroscopic and has a tendency to absorb moisture from the
atmosphere. In this view of circumstances, is it really justified to dry the oil with the aid
of rather expensive vacuum operated equipment?
c. The same question was asked by others in slightly different words: Considering the
hygroscopic properties of the oil, is it at all possible to dehydrate it to a low moisture
content of, say, 0.05% at atmospheric pressure and in the presence of humid air?
97
Answer:
German technical journal: ‘Fette, Seifen, Anstrichmittel’ the well-known research

worker Prof. Dr. Ing. M. Loncin in June 1955 published an article: ‘Die Loslichkeit des
wassers in Fetten and Oelen und der Dampfdruck des gelosten Wassers’. (The
Solubility of water in oils and fats to aldehydic and Ketomic oxidation and enzymatic
decomposition by the action of lipases as causes responsible for rancidity in oils, the
author has found a thick factor contributing to decay, i.e. autocatalytic hydrolysis which
occurs in all glycerides oils.
This hydrolysis is a mono-molecular reactions of the glycerides with water, which

reaction occurs in the homogeneous phase and is catalysed by the fatty acids present.
As has been said in Vol.1 of this Review, Mr. Loncin has been the first to demonstrate
that the velocity of the reaction alluded is directly proportional to the free fatty acid
content and that this velocity increases very swiftly with a rise in temperature. For
example, in palm oil having an FFA content of 3.3% and containing a certain amount of
excess moisture, the acidity will rise to 4.95% after storage for 199 days at 37oC. At
60oC, however, the rise will be as follows:
Storage period
In days: 0 25 50 75 100 119
FFA 3.3% 5.5% 7.2% 11.6% 17.3% 20.0%
(These values have been derives from the graph published by Mr. Loncin).
The increase in the FFA content remains the same, irrespective as to whether the oil is
merely saturated with water pr contains excess water. If there is excess water there will
be a second phase in which micro-organisms potentially responsible for enzymatic
decomposition may develop in the presence of impurities such as mucilages,
phosphatides, etc.
In the article referred to Mr. Loncin describes how he and his co-workers have
proceeded in determining the solubility of water in oils and the vapour pressure of the
dissolved water. In his conclusions the author holds that the solubility of water in oils
depends upon the temperature and the FFA content, which was illustrated by the
following table on palm oil of 60oc.
FFA in % 0.2 2.5 6.5 1.5

Water dissolved, in % 0.23 0.27 0.32 0.45
According to the author, the solubility tends to diminish with the chain length and
appears to be the same for glyrides and fatty acids.
98
Later experiments (such as those conducted by Dr. K.F. Heinisch of the AVROS
experimental station at Medan – Indonesia) correlation between water solubility
increases with a rise in temperature no conclusive evidence has been found.
With the tests carried out by Mr. Heinisch the water solubility was found to increase
with the temperature up to 70oC, beyond which point it decreased again, considerably.
By centrifugeing only the non-dissolved water can be separated out. However, it is

precisely the dissolved water that affects the quality of the oil. Palm oil from which the
non-dissolved water had been removed with the help of a purifier has been found to
break down with the same velocity as palm oil containing a generous quantity of non-
dissolved excess water.
This implies that the only way to improve the quality of the oil is to remove the
dissolved moisture by evaporation, which evaporation depends upon the vapour
pressure of the dissolved water.
From the tests conducted by Mr. Loncin it appears that below the limit of solubility the
vapour pressure is proportional to the moisture content, whereas beyond that limit the
vapour pressure remains constant and corresponds to the vapour pressure of free water
at the same temperature.
This is illustrated by the graph on the next page, applicable to palm oil of 60oC with an
FFA content of 2.45%.
On the basis of the graph it is possible to find out at what moisture content the oil will
start behaving like a hygroscopic substance.
When the vapour pressure of the water dissolved in the oil exceeds the partial pressure
of the water vapour in the air above the oil, the oil will deliver moisture to the air.
Conversely, when the partial vapour pressure in the air exceeds the partial vapours
pressure in the oil, the water will absorb moisture from the air and behave like a
hygroscopic substance.
If we assure that the air to be in such a condition as is frequently found in the tropics,
i.e.: t – 30oC; relative humidity = 90%, the partial pressure of the water vapour in that
air will be 28.4mms mercury column. If the air is heated to 60oC, the relative humidity
will drop to approx. 20% but the partial vapour pressure remains 28.4 mm Hg. This air
is now in equilibrium with palm oil having a moisture content of 28.4 / 150 x 0.27 =
0.51%, and the oil will deliver or absorb moisture according as the actual moisture
content is higher or lower than 0.51%
The rate of absorption or deliver will be low in view of the small differences in in
vapour pressure arising under the circumstances generally prevailing in the oil recovery
process.
99
Summing up the foregoing brief exposition we arrive at the following answers to the
queries raised:
a. By means of centrifugals onlt the non-dissolved water can be removed. The

moisture percentage remaining after the centrifugeing process depends upon the
FFA content. At any rate it will be too high to secure the oil from taints, sothat
further dehydration is certainly recommended.
b. Broadly speaking, under the circumstances prevailing in oil mills palm oil will
not develop hygroscopic properties before the moisture content has dropped to
0.05%. As a general rule, dehydration to such a low moisture content will prove
justified economically. There seems to be little sense in reducing the percentage
beyond 0.05.
c. Dehydration to a final moisture content of 0.05% at atmospheric pressure and in
the presence of moist air is possible theorectically. In practice, however the
moisture content will probably be slightly higher. In that case the oil will
definitely not hygroscopic.
Question:
Is palm oil soluble in water?

This question was asked by one of our readers on the strength of the following
considerations:
A higher NFPQ necessitates a stronger dilution of the crude oil with hot water.
A higher NFPQ implies increased losses of oil in sludge.
If palm oil were soluble in hot water, this would be plausible explanation for the
increased losses consequent upon stronger dilution.
Answer:
In this connection the only data at our disposal are the never published results of an
experiment conducted in 1957 by Dr. K.F. Heinisch of the AVROS experimental
station at Medan, for the purpose of determining the solubility of palm oil in water.
The solubility appeared to vary from 0.01 to 0.08%, but no direct correlation was found
between temperature – FFA content and solubility, as is shown in the table below:
FFA content of the oil

Temperature in oC 0.08% 2.61% 5.33% 11.20%
70 0.01% 0.05% 0.03% 0.02%
80 0.02% 0.04% 0.04%
90 0.08% 0.05% 0.04% 0.05%
Notwithstanding this failure to lay down any such correlation, it is safe to assume that
water has a certain – if slight – dissolving effect on palm oil. Consequently, part of the
100
increased oil losses in sludge resulting from stronger dilution of the crude oil may in all
likelihood be attributed to the phenomenon concerned.
With extreme dilution of crude oil (as is usual in screw presses), the solubility of palm
oil in water may cause considerable (invisible) losses.
Kernel Recovery
Introduction
In continuation of the preceding issue, in which special attention was given to the
perforation of the nut grading and cracked mixture screens, we shall devote the present
number of our magazine to the other equipment required in the recovery of palm
kernels.
1. Preparative treatment of the nuts
For a successful nut cracking process it is necessary to give the nuts such a preparative
treatment as will loosen the kernel in the surrounding shell. This is an effective method
to prevent the kernels from vein cracked along with the shell, while moreover it
minimizes the risk of kernel particles adhering to the shell after cracking. If the cracked
mixture contains such adherent particles, they will for the most part be discharged
along with the shell fraction and will thus form an important source of losses in the
kernel recovery process.
The preparative treatment of the nuts commences at an early stage, i.e. during the
sterilization of the fruit bunches, as has been pointed out in the issue covering this
subject. Sometimes sterilization alone will be sufficient and no further treatment is
called for. This is particularly the case when the duration of the sterilizing cycle
exceeds 90 minutes or when ‘triple peak’ sterilization is used. With the latter method
sterilization is accomplished in three stages, the sterilizer being blown down three times
during the cycle. The increased desiccation through flash evaporation resulting from
this procedure not only ensures better bunch stripping but also prepares the nuts more
thoroughly for the cracking process to follow.
Again, the sojourn of the nuts in the digester tends to improve the cracking effect.
It is generally assumed that the kernel will shrink and thus be loosened from the shell as
a result of desiccation. This theory which to all appearance is borne out by many
observations made in every day practice has led to the construction of various types of
nut driers. The type most commonly used is the silo drier, a picture of which is shown
in this issue. Here the nuts are stored for a period varying from 12 to 16 hours during
which they are dried by a current of hot air (appr. 55°C.).
101
Next to this autoclaves are sometimes employed in which the nuts are exposed first for
20 minutes and then once more for 10 minutes to saturated steam having a pressure of 3
kgs/sq.cm. Through flash evaporation of moisture from the kernel— the autoclave
being twice blown down abruptly —the purpose of effective desiccation is achieved.
A picture of the autoclaves in question is also given in the present number. In fairness
we must add that other methods are applied with success beside those conditioning the
nuts with the aid of heat.
First there is the method already referred to, by which the fruit bunches are sterilized in
such a manner that not only the stripping effect but also the cracking effect required is
made a criterion for the total sterilizing time. When this system is employed the nuts
emerging from the depericarper are led direct to the kernel recovery station.
In some mills the discovery has been made that when the steam operated air heaters of
the nut silo driers are put out of action the results obtained are just as good. The fans are
left in operation however, because cooling is deemed necessary for a proper treatment
of the nuts.
In other mills the opinion prevails that the best way to ensure good conditioning of the
nuts is to store them for a considerable length of time (i.e. a couple of days) in silos that
102
are not ventilated mechanically. The reader will understand that all these systems
should not be judged purely by their own merits but should be viewed in connection
with the preceding sterilizing process and with the type of nut to be handled. It is an
established fact for example that small nuts are more difficult to crack than large ones,
while small Tenera nuts in their turn give more trouble in cracking than those of the
Dura variety.
On the other hand it is precisely the small nuts which can be brought into a much better
condition for cracking by an additional treatment (e.g. drying). Below we give a few
figures illustrating and supporting this statement. They represent the outcome of
experiments conducted on an estate where it was not usual to subject the nuts to any
additional process after the sterilization.
The figures indicate average values derived from a series of observations and they
relate to nuts smaller and larger than 13 mms which were dried at a temperature of
approx. 70°C.
Nut fraction smaller than 13 mms.

Drying period in hours 0 4 8 16
% broken kernels 1.7 1.4 3.4 17.6
% kernel on shell 3.8 1.9 1.3 0.2
% uncracked nuts 3.2 2.6 1.5 1.2
Nut fraction larger than 13 mms.

Drying period in hours 4 8
% broken kernels 2.3 2.9
% kernel on shell 5.1 2.4
% uncracked nuts 3.8 4.1
These tables warrant the conclusion that for small nuts the drying process gives a
marked improvement of the cracking results, whereas the effect is less obvious where
larger sized nuts are concerned. Furthermore, it would appear that unduly prolonged
drying tends to increase the percentage of broken kernels.
2. Screens for nuts and cracked mixture
In the preceding number we have dwelt at some length on the task allotted to the nut
graders and the screens for cracked mixture. Particular stress has been laid on the
method of determining the theoretical optimum size for the screen plate perforations.
Of grading screens provided with such perforations we may reasonably expect that the
fraction dropping through does not contain any undesirable matter, but we cannot be at
all certain that in the fraction passing over the edge there is no material which in fact
belongs to the other fraction and should have fallen through the perforations. When the
screen is overloaded this is even quite likely to happen for there is a constant interplay
103
of capacity and screening efficiency. To clarify this point we shall now proceed to
define the term ‘screening efficiency’.
If we designate the material supplied to the screen as ‘feed’, the material falling through
as ‘product’ and the material remaining behind on the screen as ‘overflow’, the
screening efficiency can be said to imply the inclusion of the maximum amount of
desirable material from the feed in the product and the exclusion of the maximum
amount of undesirable material from the product. The efficiency is expressed as a
percentage and is calculated in the following manner.
dp = proportion by weight of desired material in product.

dv = proportion by weight of desired material in feed.
do = proportion by weight of desired material in overflow.
P = total weight of product.
V = total weight of feed.
O = total weight of overflow.
Recovery = (dp.P) / dv.V
Rejection = 1 - [( 1 - dp) . P] / [(1 - dv).V]
Screening efficiency = recovery x rejection =
dp . P ( 1 - dp) .P
1-
dv.V ( I - dv) .V|
A single example will serve to explain this. A nut grader is fed with material consisting
of 50% small nuts and 50% large ones. The perforation used is such that no large nuts
can drop through. The screen capacity however, is insufficient so that part of the small
nuts is passed to the overflow.
104
If the screen concerned is provided with larger perforations the amount of product will
increase. Beside an increased quantity of desired small nuts this product will then
105
contain a given quantity of unwanted large nuts, so that the picture may look as
follows:
This example demonstrates beyond any doubt that by providing a screen which is too
small with perforations of a size deviating from the one called for theoretically, we may
not only obtain a better distribution of material over the two fractions (product and
overflow), but also improve the sscreening efficiency.
In general, capacity and screening efficiency are closely related factors. If a low
efficiency is acceptable the capacity can he high and, conversely if a high screening
effect is required this will involve a low (or lower) capacity.
Further the output of a rotary grading screen is conditional on a great many other points
of which we mention: the size of the perforations (the larger this size, the larger the
proportion by weight of the material falling through per unit of time); the ratio of the
free screening area to the total area; the speed of the screen and the nature of the
material handled.
Naturally, if the feed contains a large proportion of material that is just a little smaller
in size than the perforations, this is bound to affect the capacity or the screening
efficiency to a considerable extent. It is this circumstance which may cause trouble in
the grading of Tenera nuts. Tenera nuts namely, differ from Dura nuts in a few
important respects i.e.:
1. Dura nuts have a thicker shell, so that the difference in size between kernels and
half-cracked nuts is far greater than with Tenera nuns.
106
2. Dura nuts occur in a much greater range of sizes than do those of Tenera type,
which means that the frequency histogram for the latter is higlter aml narrower
than that for the former. One might say that the belt or area within which the
separation should he accomplished is considerably narrower than with Dura nuts.
This is why grading screens which have never given rise to complaints when used for
Dura nuts, may prove unsatisfactory as soon as they are required to handle Tenera.
An adequate treatment of Tenera nuts mostly necessitates a larger screening area than is
applied to grade Dura.
In the forgoing we have alluded to the fact that the mother of rotations of the grader
affects the screen capacity in large measure. As a rule, the capacity of a screening drum
will increase along with tlte rate of speed, up to a certain point. For nut graders and
cracked mixture screens the optimum is found at one third of the critical number of
rotations: n = 1/3 nk.
As will be known the critical speed is nk = 1800

d
in which formula the diameter d is expressed in metres. For instance, given the
diameter of a nut grading screen is l.20 metre; required the optimum number of
rotations of the screen:
nk = 1800 = 38.7 r.p.m.

1.2
n = 0.33 x 38.7 = 12.9 r.p.m
From the point of view of engineering the grading screens are very simple contrivances
which hardly need any description.
Small end inexpensive recovery units are mostly equipped with screens of the open
type. For larger units the screens are generally enclosed in a casing of steel plate which
considerably reduces the dust nuisance for which open installations are notorious.
3. Nut crackers
In palm oil mills two types of nut crackers are employed i.e. roller crackers and
centrifugal crackers.
3.01 Roller crackers
In these machines the nuts are cracked between two fluted rollers revolving in opposite
directions. As the clearance between the rollers is invariable whereas the nuts are of
different sizes they cannot all be cracked effectively.
107
To meet this drawback the rollers are sometimes made conical. Also there are crackers
incorporating cylindrical rollers, the centre lines of which are not parallel, so that the
clearance between the rollers differs from point to point. In order to get the nuts at the
proper point between the rollers according to their size they should be graded.
A grading screen however, is not accurate enough for this purpose. Therefore separate
grading rollers are often mounted above the cracker, the clearance between then being
slightly wider than that between the cracker rollers. Like the latter, the grading rollers
are sometimes cylindrical with divergent centre lines so as to obtain a gradually
increasing space between them, while in other cases they are conical. The nuts are fed
lengthwise along the grading rollers which revolve in opposite directions and away
from each other. They fall between and through these rollers in order of size and thus
come between the cracker rollers precisely at the point where they can be cracked most
advantageously.
The output of the crackers is limited by that of the grading rollers and is not very large.
It varies from 250 to 500 kgs of nuts per hour. The cracking effect is extremely good,
being in the neighbourhood of 98%, so that it is scarcely worth the trouble to separate
out any uncracked nuts. Therefore this procedure is mostly dispensed with.
The crackers have a drawback though, in that they are subject to considerable wear and
tear. Another weak point is the fluting of the cracker drums which does not always grip
the nuts properly.
Both on account of the small output and the heavy wear it is subject to this type of
cracker is seldom used nowadays. We have only mentioned it for completeness’ sake.
3.02 Centrifugal nut crackers
The nut crackers commonly employed in present day palm oil mills are of the
centrifugal type. They are available in two models, one with a horizontal and one with a
vertical shaft. The former model is used more frequently than the latter.
In principle the crackers consist of a high speed rotor provided with it number of radial
channels or slots across the face. The mtts to be cracked are fed to the centre of the
rotor at a uniform rate and enter the radial slots. Thence they are hurled with great force
against a cracker ring surrounding the rotor, so that the shells are broken. According to
the diameter at the rotor the number of revolutions of the shaft may vary from 1000 to
2500 per minute. The peripheral velocity of the rotor is 25 to 30 metres/sec.
Not all the nuts are cracked the first time they are fed into the cracker. It should be
possible though, to achieve a cracking effect of 90%, while adequate preliminary
grading and conditioning of the nuts may raise this percentage to 95.
By increasing the speed of the cracker rotor, the cracking effect can be improved even
further, but this entails a greater risk of broken kernels. A close observation should be
kept on the performance of a nut cracker with the help ol two factors, i.e. the cracking
108
effect and the percentage or broken kernels. To assess the cracking effect, approx. 10
kgs of cracked mixture should be separated out by hand (incidentally, this test requires
a temporary interruption of the feed of half-cracked nuts ex cracked mixture screen).
Now the uncracked nuts and half-crackecl nuts from the sample are weighed and the
percentage by weight with respect to the total amount of cracked nnixture in the sample
determined.
For example: untracked nuts 1.8%

half-cracked nuts 3.4%
5.2%
Cracking efficiency 94.8%
100.-%
To calculate the percentage of broken kernels all kernels and fragments of broken
kernels in the same sample should be separated out visually and manually. This is quite
a time taking procedure, the more so because in order to obtain an accurate picture it is
necessary to sort out even the smallest fragments of kernel.
Thereupon all the kernel particles are put together and weighed. The percentage by
weight found with regard to the total amount of kernels in the sample then gives tlte
percentage of broken kernels. On top of this, well-run mills also assesss the kernel
extraction efficiency. For this purpose, the kernel losses in the pericarp residue and the
shell are determined at regular intervals.
As the ratio or fibre and shell to bunches is known by weight, it is possible to calculate
the percentage of kernels that is lost with the fibre and shell fragments. The kernel
extraction efficiency is then found as follows:
Kernels produced x 100 = K.E.E.

Kernels produced + known losses
This calculation does not guarantee complete accuracy, however, because no allowance
is made for the amount of kernel dust that is discharged along with the shell fragments
in the cracked mixture screens or with the wash water. Those unknown losses may he
comparatively great, so that the test described above is important as an additional
check.
As the nuts are weaker at the thick head (germination holes) than at the tapering tail
(beard) there is some sense in finding out whether they actually strike the cracker ring
head first. The poor cracking ressults obtained with small Tenera nuts having a long
beard have sometimes given rise to the oppinion that at least a part of these nuts hit the
ring with their resilient tails and thus impede proper cracking. To give the reader an
insight into this mattter we shall now record the following experiment:
Of a centrifugal nut cracker with a horizontal shaft, rotor diameter 460 mms, internal
diameter of the cracker ring 710 mms, speed 1050 r.p.m., the cracker ring was covered
109
at the inside with moulding clay in a layer of approx. 10 mms, in order to catch the nuts
and to locate their point of impact. Each time only a limited number of nuts were fed
into the cracker.
110
The experiment produced the following results:
1. The large nuts were all completely cracked in spite of the thick layer of clay on
ring.
2. The mediun-sized nuts were only partly cracked.
3. The small nuts were not cracked at all and were trapped in the clay.
4. They stood in a slanting position thick end first.
5. They were not evenly distributed over the periphery of the cracker ring: by far the
greater part of the points of impact was confined to a limied area.
From these observations the following conclusions can be drawn:

a. The speed of the rotor should be adapted to the size and weight of the nuts.
b. A mixture of nuts of different sizes can never produce a satisfactory cracking
effect. Careful preliminary grading is therefore required.
c. The speed of the nuts probably exceeds that of the air surrounding them, so that
they are brought into the correct position for cracking by their aero-dynamical
shape.
d. Apparently the clearance between rotor and cracker ring (710 - 460): 2 = 125
mms in the case under review) is sufficient to allow the correct positioning of the
mus.
In regard to the latter point we would observe that if the end of the radial slot in the
cracker rotor has been rounded through wear, the nuts may be given some sort of
111
rotation on leaving the slot. Whether the clearance of 125 mms will then be enough to
bring the nuts into correct position is an open question.
Consequently, whenever a poor cracking effect is obtained this may be attributable to

wear on the slots in the cracker rotor.
e. The many points of impact in a small area most probably result from the fact that
the nuts which are fed to the centre of the rotor and thence fall down always start
on their spirally curved path through the rotor and the cracker housing at almost
the same point. This may cause the re-bounding shell fragments of a crached nut
to impede the cracking of the nut immediately succeeding it. The effect of the
nuts invariably traversing the same route in crackers with a horizontal shaft is
demonstrated by the photo below showing two cracker rings. In one spot a rather
deep pit is clearly to be seen, while hardly any traces of wear are apparent on the
rest of the ring.
Although in spite of the fault mentioned crackers with a horizontal shaft usuaully
produce a very good cracking effect (90 – 95%), another type of cracker incorporating a
vertical shaft has been developed especially for the treatment of Tenera nuts.
The reader will understand that as with this type of cracker the nuts can commence their
spirally curved route through rotor and cracker housing at any place the points of
impact of the nuts on the cracker ring will be found throughout the periphery. This
112
reduces the risk of the nuts to be cracked meeting re-bounding fragments of shell on
their way to the cracker ring, which may improve the cracking result.
Often the gap between rotor and cracket ring is made extra wide which gives rather a
large apparatus.
The idea underlying this construction is to enable the nut to be brought into the
optimum aerodynamical position (i.e. thick end foremost) and to make it strike the
cracker ring at the most favourable angle.
This can be explained with the help of the schematic diagram of fig. 1. The nut to be
cracked entering the cracker rotor at x, traverses the spirally curved route xyz to leave
the rotor at point z.
The radial outgoing speed is Vr = w R2 – r2,
if we leave the friction of the nut with the rotor out of account. The peripheral speed Vo
= wR. As r is small with regard to R. Vr is practically equivalent to Vo and the angle
between Vo and Va is something less than 45o. The angle at which the nut travels
toward the cracker ring at the absolute rate Va = Vr2 + Vo2 is more obtuse as the
diameter of the ring is smaller. In other words, the risk that the nut will glance off the
cracker ring is smaller as the cracker ring diameter is larger.
Furthermore it should be observed that at the same absolute rate of speed Va the
resolvend perpendicular to the cracker ring is larger as the cracker ring is larger. Thus
in fig. 1 K2 is larger than K1.
As the square power of this speed determines the cracking force and the square power
of the absolute speed Va is a measure for the load taken up by the cracker, in theory the
conclusion would be warranted that the power consumption of a centrifugal nut cracker
having a larger cracker ring is sligthly lower than that of a cracker with a smaller ring.
This theoretical advantage however is of little importance when matched against the
other plus points claimed for crackers with a vertical shaft.
Of both tvpes of nut crackers a picture is shown in these pages.
4. Kernel and shell separators
To divide the cracked mixture into kernels and shell many methods have been tried in
the course of years. There are ‘dry’ and ‘wet’ separation methods. The dry method is
employed in winnowing systems which separate the kernels from the shell by means of
an air stream, the idea being that the small and smooth kernels offer less resistance than
the shell, so that the latter will be discharged with the air.
Other systems of dry separation are based upon the specific shape of shell and kernel,
the kernels being made to roll down an inclined surface and the shell fragments with
113
their sharp edges remaining behind. With the wet separation methods, which are
applied on a larger scale, advantage is taken of the difference in specific gravity
between kernels and shell.
They can be classed into methods employing a liquid of which (e.g. by the addition of
clay or salt to water) the specific gravity is raised to a point between the gravities of
kernels and shell and method employing clear water.
4.01 Winnowing systems
The simplest winnowing device used rather frequently and designed to effect partial
separation is placed between the cracked mixture screens and the kernel and shell
separator proper.
In the years when only thick-shelled Dura fruit was handled it had for its purpose to
blow small fragments and shell dust out of the cracked mixture. If this dust were
allowed to enter the clay bath, it would raise the viscosity of the separating liquid and
thus impede proper separation. The device consists of a vertical tube or duct with a
rectangular section in which the cracked mixture meets an air current directed upward.
The material discharged with the current is separated out by means of a cyclone. One of
the walls of the vertical duct is movable, so that the section can be made larger or
smaller to control the rate of speed of the air passing through.
The task of this winnowing machine has become more important after the introduction
of Tenera fruit, the cracked mixture of this variety containing a much larger proportion
of light shell fragments, abortive fruit and small nuts with a long beard which cannot be
cracked.
114
The material in question has a specific gravity which differs little or nothing from that
of the kernels, so that it cannot be separated out in sufficient measure in the subsequent
wet separators.
The winnowing system however, is not selective, i.e. it does not permit accurate
separation. When the speed of the air stream is increased more undesirable material can
be separated out, but this causes a larger proportion of broken kernels and small Tenera
kernels to be blown out along with it. Consequently where such winnowing systems are
employed a careful observation should be kept on the loss of broken and small kernels
discharged with the ~lust. This loss should not exceed the cost involved in manual
separation of unwanted material from the final product.
As the air speed required to discharge large fragments of shell is so high that a
considerable proportion of the broken and small kernels is liable to be dragged along,
other — often quite ingenious — contrivances designed to separate kernels and shell by
means of air have failed to become popular. As a rule the selectivity of these appliances
is rather poor: too many impurities are found in the kernel fraction and too many
kernels in the shell fraction.
4.02 Mechanical separators
With most of the mechanical kernel and shell separators in existence the cracked
mixture is spread out in a thin layer on a moving belt or carpet. The carpet is provided
with closely set thin steel needles in much the same way as the belts applied to card
wool and cotton fibres. The carpet is arranged at an incline and moves in upward
direction, while the mixture of kernels and shell has a tendency to travel downward by
the force of gravity. The principle underlying the system is that the round and smooth
kernels can roll down against the movement of the carpet, whereas the sharp-edged
shell fragments are retained by the needles and carried upward to be discharged
afterwards. An ideal separation cannot be effected with these machines, because on one
hand broken and flat kernels are removed by the carpet, while on the other hand small
uncracked nuts will roll down along with the round kernels.
An advantage of this type of separator is that the kernels are not wetted and therefore
do not require intensive drying. It is an error, though, to believe as is sometimes done
that the kernels thus recovered need no drying at all, for the moisture content of the
seemingly dry kernels is too high after cracking to prevent the development of mould
and an accelerated increase in acidity.
The separators described have another attractive point in that they do not require any
clay or salt, so that the kernels need not be washed afterwards
They are seldom used nowadays, although them interest in this system as a means of
secondary treatment to remove the considerable amount of impurities found among
dried Tenera kernels seems to have revived.
115
4.03 Clay bath separators
There are many types of clay bath separators. The differences between them are found
in the design rather than in the working principle and they are mostly named after the
manufacturers or after the estates where they were employed for the first time. The
method of separation is based on the difference in specific gravity between kernels and
shell, the gravity of the kernels being approx. 1.07 and that of the pure shell fragments
1.3 to 1.4 The cracked mixture of Tenera nuts, however, contains (light coloured) shell
of a much lower gravity and fragments of shell and/or small nuts to which the long
beard — one of the characteristics of Tenera nuts — is still attached. The specific
gravity of this material comes very near to that of the kernels and somentimes is even
lower, so that it is sure to cause trouble when separated according to gravity.
The bath in these separators generally has a specific gravity of 18°Be. Very often a clay
suspension is used, but salt solutions, dilute molasses and/or waste material from soap
factories may answer the purpose equally well. The clay required to prepare the
suspension should be of a fine structure and may not contain any appreciable amount of
sand or organic matter. For every 100 tons of bunches about 2 cu.m. of clay are needed.
As suitable clay is often hard to come by, it is sometimes attempted to save clay by
allowing the wash water to settle, or—since water is needed anyway to prepare a fresh
clay suspension—by keeping the amount of wash water just as large as will be
necessary for making the new suspension. In using clay bath separators it is necessary
to keep the clay well suspended and to prevent it from settling on the bottom. That is
why the machines are so constructed as to maintain a more or less forced circulation in
the clay bath.
116
The simplest type of clay bath separator designed for manual operation consists of a
rectangular iron reservoir in which a three-cornerecl perforated tank is suspended on
hinges. By means of a handle this tank can be made to turn on the hinges which we
attached to one of the upper edges of the external reservoir. Thus the tank is tipped up
to a point where it will discharge its contents.
The external reservoir is filled with a clay suspension of the requisite gravity. The
cracked mixture is dumped into the lowered internal tank and the clay water thoroughly
stirred. Next the kernels are allowed to rise to the surface and scooped off. The
fragments of shell which will go to the bottom are removed by tipping up the perforated
tank. The clay water will now run back into the main reservoir, while the shell passes
over the edge. A picture of this simple seperator is shown here.
Another clay bath separator, of the type manufactured by Stork, is shown in the
sectional drawing of Fig. 2. The machine is composed of a trough in which is mounted
a central shaft carrying two paddle wheels and two washing drums. This system not
only separates the kernels from the shell but also removes the adherent clay from the
two fractions after separation.
The mode of working is as follows.
The central section of the trough is divided into two compartments A) and B) by a
baffle set at an incline. At the top of the baffle is an overflow through which the clay
water can pass from A) to B). As the shaft rotates a paddle wheel constantly scoops
clay water out of the compartment B) and on moving forward pours it into the
compartment A).
The surplus of clay water thus formed in A) will flow bank into B). The mixture to be
separated is admitted to compartment A), where the shell will sink and the kernels
come to the surface. Now the kernels are passed through the overflow and to the kernel
washing drum by the constant flow of clay water. In the first section of the drum the
117
excess of clay water is removed and allowed to run back into compartment B). The
second drum section serves the function of washing and rinsing the kernels with clean
water.
The sinking shell fragments pass along the inclined baffle.
Arrived at the deepest point of the section, they are licked up by the other paddle wheel
to be discharged into the drum connected with this wheel. Here again the surplus of
clay water is removed in the first section, while the shell fragments are rinsed with clear
water in the next.
This automatic kernel and shell separator requires no supervision beyond maintaining
the requisite quantity of clay water by replenishing the bath at regular times and
keeping a check on the gravity of the clay suspension. The check can be carried out in a
simple manner with the aid of an areometer. If as a result of excessive dilution the
gravity has dropped to the point where the kernels are in danger of going to the bottom
and being dischargecl along with the shell, fresh clay water having a density exceeding
18 Be should be added.
To prepare the high gravity suspension simple clay mixers are available consisting of a
trough with a stirring device of robust design.
118
The principle described above admits of many variations. For example, there are a great
many possibilities ol removing the kernels and shell from the bath, mechanically as
well as automatically.
Whatever the design, the clay bath separators of this type invariably keep the clay water
in forced circulation, both to avoid sedimentation of clay particles and to float the
kernels recovered towards the discharge.
4.04 Kernel and shell separators using clear water
The oldest machine known in this category has much in common with the washing
equipment employed in coal-mining to separate coal from stone. It consists of a tank
filled with water in which is suspended another tank having a perforated and sloping
bottom. Across the machine is a shaft imparting a rapid up and down movement to the
internal tank by means of eccentrics and connecting rods.
The cracked mixture, which is admitted to the internal tank in an even flow, first goes
to the tank bottom. The tank however, makes a dancing movement and each downward
stroke causes water to be forced through the bottom perforations. The vertical jets of
water thus produced have the effect of washing the lighter kernels out of the cracked
mixture and floating them to the surface.
The heavier fragments of shell remain on the bottom but, this being inclined, they will
move towards the lowest point. A couple of adjustable slides are mounted in the
internal tank in order to keep kernels and shell apart after separation by the water jets,
and to permit a separate discharge of both fractions.
The seperator in question has a comparatively low output while the separating effect is
indifferent. Even with Dura crop it is impossible to avoid a second treatment of both
kernels and shell to sort out undesirable matter.
During the last decade hydro-cyclone separators have stood in ever greater demand.
These installations which were originally designed for washing coal are now applied on
a large scale both in palm oil mills and in many other branches of industry. The
advantage of hydro-cyclones over clay bath separators is that nothing except clean
water is required to bring about the separation, while the separating effect is as good.
Consequently, they are most welcome on estates where clay of the requisite
composition is not available or must be hauled over long distances.
Moreover the use of hydro-cyclone separators renders the clay mixers superfluous, so
that a much higher standard of hygiene can be observed in the mill. Against this we
must set a few drawbacks in that the separator requires a higher capital outlay, while
the power consumption is far in excess of that of a clay bath separator.
The working principle is shown in Fig. 3.
119
The cracked mixture is admitted into the main bath A which is completely filled with
water. By means of pump C 1 water and cracked mixture are now forced to the cyclone
D. The greater part of the kernels and of the water leaves this cyclone through the
vortex opening. The water is allowed to flow back into the bath A through the washing
drum F. The shell and the balance of the kernels are passed from the cyclone to the bath
B through the apex opening. Next the pump C 2 forces this mixture (which consists
largely of shell fragments) to the cyclone E where the final separation takes place, most
of the shell fragments being discharged to the washing drum G through the apex
opening and the kernels and remaining shell being held in the vortex and fed to the bath
A whence they are re-cycled. The hyclro-cyclones are so dimensioned that most of the
water is forced to leave through the vortex opening. The water and cracked mixture are
fed under pressure at a tangent to the periphery of the cyclone. Thus a centrifugal action
is set up, those fragments heavier than water being driven to the conical cyclone wall.
Apart from the rotatory movement or whirl imparted to the water in the cyclone, non-
tangential currents are set up opposite to the direction of the centrifugal force. The
reason is that most of the water entering the cyclone at the periphery is discharged
again through the vortex opening in the centre. Consequently, two opposing forces are
exerted on each particle in the cracked mixture: first the centrifugal force which
attempts to throw the particle to the wall and second the force of the water current
directed to the vortex opening which attempts to drag the particle along. In Fig. 4 the
non-tangential currents are represented diagrammatically.
If the particle is light and/or if it has a high flow resistance it will leave the cyclone
with the vortex.
On the other hand, if it has a high specific gravity and/or a low flow resistance the
centrifugal force will be the more powerful, so that the particle is driven to the cyclone
wall and will be discharged through the apex opening along with the smaller part of the
water.
With hydro-cyclones, therefore, it is not only the gravity which decides the separation
as with clay bath separators, but also the flow resistance. In other words, the size and
shape of the material handled play a significant part, which would warrant the
conclusion that a hydro-cyclone separator must be less selective than one of the clay
bath type.
In practice, though, no evidence has been found to support this theory as far as cracked
mixture of Dura nuts is concerned.
The case is different with cracked mixture of Tenera nuts, for here as with clay bath
separators no perfect separation is feasible. Abortive fruit, light shell fragments and
small uncracked nuts (as far as still found in the cracked mixture after winnowing) are
apt to be discharged through the vortex opening along with the kernels as a result of the
slight difference in gravity and because they have a higher flow resistance than the
smooth kernels.
120
5. Kernel driers
The last process to which the kernels are subjected is drying. This is indispensable
because:
1. moist kernels are more liable to become mouldy.
2. when kernels are stored in a moist state the acidity will increase more rapidly than
when they are stored dry.
It may be taken for granted that the question whether the kernels will develop mould
depends on their moisture content. The critical point seems to be in the neighbourhood
of 8%, for below this percentage no mould is found
The development of mould is also promoted by storage of the bagged kernels in rooms
having an excessive relative humidity or by the bags being piled too high. It is
necessary therefore, always to store the kernels in a dry and airy condition.
It has been observed that kernels containing more than 7% moisture, which were stored
under normal conditions, dried down to 7% moisture after a certain period of time,
whereas kernels with less than 7% moisture were found to absorb moisture from the
121
atmosphere until the balance of 7% was reached. This proves that there is no need to
dry the kernels to it moisture content below 7%. When drying kernels high
temperatures must be avoided, for overheating is sure to cause discoloration. This
subject has been dealt with in the chapter on sterilization. Discolouration of the kerned
is undesirable, because it greatly affects the colour of the kernel oil recovered.
Now discoloration is certainly not caused by drying alone. As a matter of fact the
sterilization process is even more responsible for this. When drying is carried out at
excessive temperatures however, a strong additional discolouration may ensue.
For the reader's guidance we repeat below the summary stating the quality of the kernel
oil recovered in relation to the quality of the kernels, as given in vol. 1 nr 1 (March
1960):
122
1. Kernels from fruit that is sterilized for a very short period at normal temperature
will yield oil of a slightly yellow colour, although the kernels themselves remain
quite white.
2. The oil recovered from normal kernels is of a light yellow colour and is easy to
bleach.
3. The oil recovered from kernels which have not been adequately dried is of a
grayish-yellow colour and is hard to bleach.
4. The oil recovered from light brown kernels is again light yellow and can generally
be bleached without trouble.
5. The oil recovered from kernels that have been stored wet and/or are slightly rotten
is dark red and hard to bleach.
6. The oil recovered from dark brown kernels is dark brown and cannot be bleached.
An important factor to be observed in drying kernels is the rate or drying. If this rate is
too high the surface of the kernels will become more or less glassy and hard, so that the
evaporation of moisture from the inside of the kernels is impeded.
As a result the moisture content is likely to remain too high.
With regard to the drying temperature it can be said that at the beginning of the process
when the kernels should still be considered as wet bodies containing a large proportion
of surface moisture.
Air of a high temperature may be used without trouble for at that stage the kernels
cannot assume a temperature in excess o[ the wet-bulb temperature of the drying air.
However as the surface moisture will evaporate rapidly even when lower temperatures
are employed and as the drying ol kernels is mainly a matter of diffusion drying (the
kernels assuming a temperature corresponding to the dry-bulb temperature of the drying
air) it is generally thought advisable not to use drying temperatures in excess of 50 to
55°C.
A circumstance to be taken into account in deciding on the drying temperature to be

adopted is that, at high temperatures the kernels will exude oil, especially if they are
under some pressure. The fatty apppearance they thus acquire need not necessarily
stand in the way of a ready sale. Indeed, for reasons we are at a loss to explain this fatty
appearance is even sought after by many buyers.
Recapitulating the conditions to be fulfilled to ensure good drying results are a drying
air temperature of 50 - 55°C, and a drying period of from 14 to 16 hours which values
are based upon an initial moisture content of 20 to 22% and a final moisture content of
7%. The following types at driers are in frequent use on palm oil estates.
123
Tray driers
With tray driers or hurdle driers the kernels are dumped into shallow trays having a
perorated bottom. The trays are then taken automatically through two drying shafts,
during which journey the kernels are dried by heated air which is passed through
countercurently.
The trays of one such drier which we know of had a length of 2500mms, a width of
2500 mms and a depth of 250 mms. The kernels were admitted into the trays in layers
of approximately 200 mms. There were 24 trays divided between the two shafts. The
drying time was 15 hours and the output 1400 kgs of dried kernels per hour.
The trays being comparatively shallow, it is not easy for the kernel oil to exude.
Another advantage is that the kerned traverse a well defined route through the shafts so
that a unform result is obtained. Considerable drawbacks are the amount of supervision
called for, the work involved in loading and unloading the trays and the expensive
maintenance resulting from the large number of moving parts.
Silo driers
These machines are of a very simple design, consisting of a large silo in which the
kernels are stored. Through the silo heated air is either blown or sucked by a fan. In the
drier are mounted roof-shaped ducts which serve a threefold purpose.
Firstly, they take care of the most uniform distribution of the air possible throughout the
section of the silo. Secondly, they reduce the pressure exerted on the kernels. Owing to
the careful dimensioning and arrangements of the ducts the load on the kernels is no
higher at the bottom of the silo than at some 60 cms from the top, irrespective of the
height of the machine, which may be onsiderable.
I.astly the roof shaped ducts serve the function of ensuring that the kernels are given
exactly the same rate of flow (= drying period), regardless of their position in the
section of the dryer (i.e. at the side or in the centre).
124
The wet kernels which on emerging from the clay bath or the hydro-cyclone separator
are generally rid of the surplus of adhering water by means of a steam jet, are admitted
at the top of the silo in a continuous flow. The dried kernels are discharged
continuously at the bottom through a shaking grid. The stroke of this grid and thus the
rate of discharge is adjustable and adapted to the rate of feed, so that the drier is kept
permanently filled.
At the maximum throughput the drying time is 14 to 16 hours.
If this maximum capacity is not wanted, the drying period is prolonged and the air
temperature reduced in a measure corresponding to the fall in capacity.
125
Tlte machines in question tlo not require any supervision beyond adjusting the rate at
discharge of the shaking grid and setting the temperature of the drying air.
The photograph depicts a Stork kernel silo drier composed of three successive zones
through which the drying air is blown.
The air temperature is adjustable for each zone individually.
126
127
PRODUCTION PROGRAMME for the palm oil industry
• Sterilization
Horizontal sterilizers with quick-action • Kernel recovery
door shutting device, complete with Nut silo driers
sterilizer cages and bogies in capacities Nut autoclaves
up to 18 tons/charge Nut graders and cracked mixture
Vertical sterilizers in capacities up to 6 screens
tons/charge Centrilugal nut crackers with
horizontal or vertical shaft
• Threshing Clay-bath kernel/shell separators
Rotary threshers in capacities up to 24 Hydro-cyclone installations for the
tons of bunches/hour separation of kernels and shell
Beater arm bunch strippers Kernel silo driers
• Extraction • Transport
Steam heated digesters or various Electric capstans
types and capacities Electric hoists
Extraction centrifuges Elevators for bunches - fruit - nuts and
Hydraulic presses for 1.5 and 3 tons of shell or kernels
bunches/ hour Screw conveyors - vibrating conveyors
Automatic presses, capacity 4.5 - 6.5 - flight conveyors etc.
tons of bunches/ hour
• Power and heat supply
• Clarification Fire tube and water tube boilers with
Vibrating screens for crude oil and specially designed furnaces for
sludge burning fibre, shell and or empty
Pumps for crude oil and purified oil bunches
and sludge Horizontal and vertical steam engines
Continuous first clarifiers for crude oil in various designs
Discontinuous clarification tanks Diesel motors
Sludge centrifugals Generator sets
Purifiers Complete switch boards and
Vacuum oil driers of various designs distribution boxes
Autontatic oil weighers Complete factory lighting system
• Drying and separation of press • Water supply

cake Pumping stations
Steam heated cake-break conveyors Water towers
Mechanical depericarpers Water purification units
Pneumatic drier - depericarpers
Fans and ducting for the pnematic • Oil storage
transport of fibre Storage tanks
Fibre cyclones Special oil delivery pumps
128

StorkReview PDF

Uploaded by

Copyright:

Available Formats

StorkReview PDF

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

StorkReview PDF

Uploaded by

Copyright:

Available Formats

STORK

Development of Free Fatty Acids before Sterilization

Pressure and Influence of Fatty Splitting Enzymes

Oil after Peeling %FFA

A = peeled fresh fruit

Temperature oC FFA% of oil recovered

Speed of the Fat Splitting Process

FFA% in oil recovered from bruised and pounded pericarp

Many researchers have applied themselves to determining the percentage of free

• An extension of collecting intervals will result in larger periods of temporary

Spontaneous Autocatalytic Hydrolysis

Treatment Initial % %FFA

K = 23/to (log a2 - log a1)

Oil + 0.25% moisture Same oil + 5% of moisture

There is no need to be doubtful; of the possibility of an increase in acidity brought

1. Inactivation of the oil-splotting enzyme

2. Coagulation of the albuminous substances, which will cause trouble in the

3. Hydrolysis of the mucilaginous matter.

4. Loosening the Fruit in the Bunch

We now come to the question of bunch stripping itself.

1. In the bunch stripper or thresher, varying quantities of bunch trash containing

2. Inadequately stripped bunches. As longs as it is a matter of one or two stray

(i) Faulty Action of the Stripper or Thresher

(ii) Abnormal bunches (knot-head bunches)

Only, if a bunch is reaped in a completely immature state will it become an all-out

(iii) Inadequate sterilization

5. Preparation of the Fruit for Further Treatment in the Extraction Plant.

Sterilization and the quality of palm oil

Furthermore, good bleachability necessitates a proper air-release prior to sterilization,

The Size of Bunches

Age of Palm tree Time of Steam press inadequately

Degree of Ripeness of the Crop

Definition of under-ripe, over-ripe and ripe.

Chief points controlling harvesting standards

(d) An increase of percentage of detached fruit is likely to impair the bleaching

Class 00 = no detached fruit = a%

This is controlled by a tendency of the kernel to discolour as a result of over-sterilizing.

Effect of Sterilizing Temperature on the Oil Recovery Quotient

Reduction of Moisture Content in Fruit through Sterilization

Optimum moisture content is 12% ± 2%.

According to Dalton’s Law,

3 ATMOS pressure = Partial air-pressure (1 ATMOS) + Partial Steam Pressure 2 ATMOS

Hence temperature of steam instead of 132.9oC →119.6oC.

Quantity of Steam Required For Sterilization Process

On a basis of figures obtained in practice, it maybe assured that in order to sterilize 1

Corrosion and Erosion

Sets of Temperature of % of oil

Composition of the crude from: Digester Press

Explanation Of Various Observations Taken

(I) Rupturing Of Oil Bearing Cells

Viscosity of the Crude Oil

The crude oil then consists of:

1. Non-fat: fibre and cells remnants plus soluble components

Boiling Of the Digester Contents

The requirement of proper digestion

6. During digestion the mash may not be allowed to boil.

1. Constructional Details of Digesters

The Digester Drive