Process equipment design and selection criteria

Summary of Unit operations
Unit operation
Purpose
1.
Fruit fermentation
To loosen fruit base from spikelets and to allow ripening processes to abate
2.
Bunch chopping
To facilitate manual removal of fruit
3.
Fruit sorting
To remove and sort fruit from spikelets
4.
Fruit boiling
To sterilize and stop enzymatic spoilage, coagulate protein and expose microscopic oil cells
5
Fruit digestion
To rupture oil-bearing cells to allow oil flow during extraction while separating fibre from nuts
6
Mash pressing
To release fluid palm oil using applied pressure on ruptured cellular contents
7
Oil purification
To boil mixture of oil and water to remove water-soluble gums and resins in the oil, dry decanted oil by further heating
8
Fibre-nut separation
To separate de-oiled fibre from palm nuts.
9
Second Pressing
To recover residual oil for use as soap stock
10
Nut drying
To sun dry nuts for later cracking

In designing equipment for small-scale oil extraction one of the key factors to consider is the quality required. ‘Quality’ is entirely subjective and depends on the demands of the ultimate consumer. For the edible oil refining industry the most important quality criteria for crude oil are:

low content of free fatty acids (which are costly to remove during oil refining);
low content of products of oxidation (which generate off-flavours);
readily removed colour.

The most critical stages in the processing sequence for a processor seeking to satisfy these criteria are: bunch sterilization as soon as possible after harvest; and effective clarification and drying of the crude oil after extraction.

By contrast, for the domestic consumer of crude palm oil, flavour is the primary quality factor. This is boosted by the fermentation that takes place within the fruit when the bunches are allowed to rest for three or more days after harvesting. Thus sterilization immediately after harvesting is not a crucial consideration. Herbs and spices for flavour are introduced during the oil-drying phase of operations to mask off-flavours. Therefore rigid process control during oil clarification need not be prescribed or incorporated in the design.

The free fatty acids and the trace tocopherols contained in the crude palm oil after natural fermentation also have a laxative effect, which is desirable for African consumers for whom synthetic substitutes are a luxury. The acidity imparts a ‘bite’ to the oil which some consumers prefer. Thus the quality requirements of one market, leading to certain processing imperatives, may conflict with those of another market.

The traditional manual methods are normally referred to as ‘low technology’ production. The mechanised units are likewise referred to as ‘intermediate technology’ production.
The village traditional method of extracting palm oil involves washing pounded fruit mash in warm water and hand squeezing to separate fibre and nuts from the oil/water mixture. A colander, basket or a vessel with fine perforated holes in the bottom is used to filter out fibre and nuts. The wet mixture is then put on the fire and brought to a vigorous boil. After about one or two hours, depending on the volume of material being boiled, the firewood is taken out and the boiled mixture allowed to cool. Herbs may be added to the mixture at this point just before reducing the heat. On cooling to around blood temperature, a calabash or shallow bowl is used to skim off the palm oil. Because of the large quantities of water used in washing the pulp this is called the ‘wet’ method.

A mechanical improvement, based on the traditional wet method process, is achieved by using a vertical digester with perforated bottom plate (to discharge the aqueous phase) and a side chute for discharging the solid phase components. The arrangement combines digestion, pressing and hot water dilution into one mechanical unit operation.

The ‘dry’ method uses a digester to pound the boiled fruit, which is a considerable labour-saving device. The oil in the digested or pounded pulp is separated in a press that may be manual or mechanical. Motorised mechanical presses are preferred, whether hydraulic or screw type.
Most medium- and large-scale processing operations adopt the ’dry’ method of oil extraction. This is because the fibre and nut shells may immediately used to fire the boiler to generate steam for sterilization and other operations, including electricity generation. If the huge volumes of fibre and shells are not used as boiler fuel, serious environmental pollution problems may result. Too much water in the fibre increases the amount and cost of steam required to dry the fibre. Hence the preference for the dry method in plants handling more than six tonnes FFB per hour.

Processing machinery manufacturers tend to make machines to fit individual processing operations. However, recent developments have been toward the manufacture of integrated machines, combining several process operations such as digestion, pressing and fibre/nut separation into one assembly. It is found that these machines fit into two key process groupings: batch and semi-continuous processes.

The extraction of palm oil from boiled palm fruit can be accomplished by handling successive batches of materials or continuously feeding material to the machines

The batch systems work directly on successive loads of boiled fruit to extract oil in one operation for clarification. The ‘wet’ method uses a vertical digester (Fig. 11) with a perforated bottom plate to pound a batch of fruit and then flush out the oil and other non-oil solids from the mashed pulp with hot water. The direct screw-press is designed to pound a batch of boiled fruit in the entry section of the machine while exerting pressure on the mashed pulp in another section to expel the palm oil in one operation.

The advantage of the wet system is that it is simple and completely leaches all oil and non-oily solid substances that can be carried in the fluid stream out of the digested mash to give clean and separated nuts and fibre. The aqueous effluent from the vertical digester goes directly to the clarification stage of processing. The amount of water needed to flush the pulp is normally the same as that required for diluting the viscous oil that comes from the mechanical press in preparation for clarification. An inexperienced operator may use too much hot water to leach out the oil and thus consume unnecessary wood fuel.

The ‘wet’ method yield of palm oil is severely reduced when the wash water is cold. In the course of digesting the fruit mash, in the presence of water, there is increased tendency to form an oil/water emulsion that is difficult to separate from the fibre mass. The emulsified oil loss in the fibre can be substantial if care is not taken to ensure full loading of the digester. Vertical flushing digesters, requiring loading and discharging of a specific amount of material, can thus only be used in a batch operation.

Continuous systems work sequentially, with one operation feeding directly into another, related to the arrangement and timing of machine operations. Careful engineering of unit operations is required to minimise discontinuities in the feeding of one stage into another. Otherwise some machines have to be stopped periodically for other stations to catch up. When there are discontinuities in the flow of materials between process stations the operations are known as semi-continuous. The dry extraction systems with separate digestion and pressing stations are usually semi-continuous.

Also when digestion and pressing stations are combined into an integrated unit and there is discontinuous feeding of boiled fruit to the digester inlet the operation is termed “semi-continuous”. Once operations have been integrated to attain full continuity the capital investment capacity of small-scale operators has been surpassed, because both machinery and working capital for raw material increases greatly with the increased level of mechanisation.
The dry systems do not need much water for processing, although they have the disadvantage of leaving substantial residual oil in the press cake. The oil content of the press cake can be quite considerable (2-3 percent), depending on the type of press used and the strength of manual operators.

The efficiency with which the various presses can extract oil ranges from 60 to 70 percent for spindle presses, 80-87 percent for hydraulic presses and 75-80 percent for the Caltech screw-presses. The first-pressing oil extraction rates also range from 12 to 15 percent for the spindle-presses, 14-16 percent for hydraulic presses and 17-19 percent for the motorised screw-presses. (Rouziere, 1995)

In many instances the first press cake is then sorted to remove the nuts, and the fibre is subsequently subjected to a second pressing to obtain more oil (an additional 3 to 4 percent on FFB). The second press oil is generally of lower quality, in terms of free fatty acid content and rancidity. Such low-grade oil is used in soap-making. Some village processors undertake the traditional hot water washing of the entire press cake immediately after pressing instead of sorting fibre and second pressing.

Local manufacturers have developed a wide range of machinery and equipment for processing palm oil and palm kernel to fit any budget. All the relevant unit operational machines can be produced to various degrees of finish and quality in the Sub-Region. It is the combination of the unit operation into an affordable process chain that distinguishes the manufacturers and their supplies.

From traditional technologies that rely solely on manual labour and simple cooking utensils, raising the level of mechanization depends largely on a balance between the quantity of bunches available for processing in a given locality and the money available for investment in machines.
The first consideration should be the availability of raw materials and how to compute the processing scale. Knowing the optimum scale of operations, it is then possible to consider the type of processing techniques. The higher the technology, the more skilful operators will be required to handle the machines. These technical considerations should lead to the equipment selection and examination of the capital investments needed to acquire the necessary machines.

Assume a Village Group decides to plant oil palm and establishes a program to plant a certain number of seedlings each year over a seven-year period. In the third year the first set of trees begin to bear fruit. The community wants to establish a processing mill and they call an expert. How is the estimation made of the size and type of processing unit required by the community?
Start by establishing the block of planted areas by year so the age of the trees may be determined. The oil palm tree begins to bear fruit from the third year and the yield per tree increases progressively with age until it peaks around 20 years. The yield begins to decline from year 25 through 40 when the economic life of the tree ebbs.

Table 3 describes the potential yields of palm fruit bunches (in metric tonnes) from the planted hectares per year. Estimates in Table 3 are used to calculate the expected annual yield for each annual block. For example, 8 700 seedlings planted in 1998 began to yield fruit in 2000 at the rate of 3 tonnes per hectare to give 198 tonnes for the year. By Year 7 all planted areas will be in production, at different yield rates. The estimated annual yield per planting block is calculated and then the column for the year is added to give the potential raw materials available for processing. For example, in Year 7, when all planted blocks are yielding fruit, the total is 8 919 metric tonnes (see the row designated ‘TOTAL’). How the annual yield is distributed over the entire year needs to be determined in order to know which period demands the attention of processors.

The oil palm tree yield is distributed over the entire year. Most of Central and West Africa experience two rainfall seasons. The oil palm bears fruit in response to the rainfall pattern and hence there are two peak harvesting periods in these regions. Southern hemisphere tropical monsoon regions such as Malawi, Zambia and South East Asia experience only one long rainy season and therefore tend to have a single peak-harvesting season.

For Central and West Africa the annual monthly distribution pattern for produce is expected to show the following variations:
Month
Percent yield
Seasonal contribution

March
9
April
12
May
16
50 %
June
13
July
8
August
7
September
8
34 %
October
11
November
7
December
5
January
3
16
February
1

In the peak harvesting month it is estimated that 12 to 16 percent of the annual yield is generally available for processing. The plant that is installed must be capable of processing the peak month output, which is generally estimated as 15 percent of the annual output. Conservatively, it is estimated that the plant will work two shifts during the peak season.

Table 3: Estimated annual yield per hectare (from year of planting)

Year
1
2
3
4
5
6
7
8
9
10
11
12
15
20

Estimated yield
(Tonnes)
--
--
3.0
4.25
5.5
6.0
7.25
8.2
8.6
9.5
10.5
11.0
12.5
13.5

Table 4: Estimated FFB yields after planting and related plant capacity
Year/yield in metric tonnes

Hectares
1
98
2
3
4
5
6
7
8
9
10
11
12
15
20
66
--
-
198
281
363
396
479
541
568
627
693
726
825
891
190
-
-
570
808
1 045
1 140
1 378
1 558
1 634
1 805
1 995
2 375
2 565
800
-
--
2 400
3 400
4 400
4 800
5 800
6 560
6 880
7 600
8 800
10 000
400
--
--
1 200
1 700
2 200
2 400
2 900
3 280
3 440
4 400
5 200
400
--
--
1 200
1 700
2 200
2 400
2 900
3 280
3 440
5 000
Total
198
851
3 571
6 041
8 919
10 619
12 526
14 121
15 558
17 041
19 840
23 656
Peak Month
29.7
128
536
906
1 338
1 593
1 879
2 118
2 334
2 556
2 976
3 548

Plant
Capacity/hr Plant

0.09
0.4
1.7
2.8
4.2
5.0
6.0
6.6
7.5
8.0
9.5
11.0

Source: Poku, K. Feasibility study on Malawi palm oil mill establishment

In Year 3 there is the potential of processing 198 tonnes of fresh fruit bunches. Assuming that the total quantity were to be processed in one location over a 20-day period using 8 hours in the day, we would need a processing unit that handles 186 kg per hour, or 93 kilos/hr if the choice was made to operate 16-hours per day. Table 4 shows capacity based on a 16-hour working day. For this capacity a wet type digester or the dry spindle-press operation would be recommended. By Year 5 the community would require a fully mechanised mill using motorised digesters and presses.

Before the sixth year the community would have to decide whether they want to stay in the small-scale milling category or move up to a medium-scale operation using a continuous system of machines. If the option is to stay small-scale then the community will need to place orders for additional small-scale processing modules. The new set of processing machines can be placed to run alongside the existing facility or located in another village to minimise bunch transportation costs.

The best plant size option for rural Africa is still unknown. Large-scale operations normally require high-skilled labour and management expertise. Most villages do not have such a pool of skilled labour. The villages also lack the social infrastructure such as good accommodation, schools and hospitals that would attract high-skilled labour. Thus, in order to establish a large-scale processing operation, labour needs to be imported from other parts of the country. To maintain these ‘alien’ workers and managers a provision must be made in the capital investment for housing, schools and clinics near the processing estate. Some of the schooling and medical services must be extended to the whole community or there will be resentment towards the ‘alien’ workers.

Large-scale operations also require rapid transportation of harvested bunches to the processing site, hence the need for investment in roads and civil works. The establishment of large-scale operations creates an overhead burden that is beyond the capacity of a village community.
Many of the large-scale operations established in the early 1970s have declined along with the national economies of African nations. The cost structure of these establishments has rendered the output products non-competitive on the international market.

Today decentralised small-scale processing operations are preferred in most parts of Africa.

Once the required plant size has been determined, the next item to consider is the amount of money required to buy the necessary machinery. The more money available, the more units can be bought, to minimise the drudgery of processors.

The wide array of machinery options makes it possible for a processor to start operations with a manual spindle-press used to pound the palm fruit. Another may start with a single motorised vertical wet process digester. Further up the investment scale are those who can afford the combination horizontal digester and screw-press or combination horizontal digester and hydraulic press along with the associated sterilizers, threshers, and oil clarifiers. Another combination that is yet to be tried is the combination of a horizontal motorised screw-press in combination with a second stage vertical flushing digester for maximum palm oil extraction and fibre/nut separation.

Type of unit
Key machines
Rated capacity
(k g FFB/hr)

Extraction efficiency
(%)

Capital investment
(US$)

Single batch unit
Dry

Spindle

100-200
55
150-200
Hydraulic
200-300
67-74
5 000-7 000
Screw
250-400
77.4
1 500-6 000

Wet
Vertical digester
500-800
80-90
1 500-2 500
Dry
Motorised horizontal digester
(only)
500-1000
55
2 500-3 000
Dual separate units
Dry
Digester + Spindle presses
200-300
60-70
3 000-5 000
Digester + hydraulic press
400-800
67-78
7 000-10 000
Semi-continuous combined units
Motorised digester +
500-850
70-87
10 000
Dry
hydraulic + spindle-press
-15 000
Digester + screw-press
500-850
76-90
12 000-15 000

The extraction efficiency refers to the percentage of oil that the machine can extract in relation to the total oil in the boiled fruit. The type of fruit mix (Dura/Tenera) presented for processing greatly influences the extraction efficiency of all units.
Many of the installations that use single spindle and manual hydraulic press units require manual pounding with wooden mortars and pestles, foot stomping, etc. Thus the throughput capacity of such a mill is determined by the manual pounding rate. The presses are usually not mechanised and hence the processing capacity of the press is also limited by the size of the press cage and the operator’s energy level for turning the press screw or pumping the hydraulic fluid mechanism.

Another limiting condition is the affordability of capital equipment. Where the capital equipment cost exceeds a certain value villagers will shy away from taking loans to purchase the combination of operations. The designer must bear in mind that until the rural/urban migration of village youth is reversed the villages will be mainly populated by the elderly. These elders are naturally reluctant to take up long-term loans and the local banks are reluctant to lend to a predominantly aged community group. In Ghana, for instance, capital equipment costs should be around US$10 000 to be affordable to village-based individuals or groups.

Because of the need to keep initial capital investment to a bare minimum it is imperative that unnecessary mechanised unit operations are eliminated. Work that can be done manually - without overly taxing profitability - should be, thereby taking advantage of surplus labour and creating a stream of wages and salaries in the local community. Operations that are usually associated with drudgery by processors, such as fruit digestion and oil extraction, can be mechanised. Other less strenuous tasks, such as fruit separation and fibre/nut separation, can be contracted out to elderly women and unemployed youth.

“Small-scale” does not necessarily mean a significant decrease in efficiency. It does, however, mean a reduction in working capital and operating costs. The small mills can be placed at the heart of local communities, minimising reliance on vehicular transport that is normally unavailable in rural communities, given the poor condition of road networks and other infrastructure. This increased accessibility serves to dramatically reduce fruit spoilage and consequent post-harvest losses.

Culturally, men cultivate or produce while women process and sell. Traditionally, women decide the form in which the produce is to be traded and hence determine the degree of processing they are willing to undertake. These decisions form the basis of traditional technologies upon which innovations are to be derived.

The operating philosophy for equipment innovation should therefore be an attempt to develop machinery to alleviate the drudgery of female processors while providing additional avenues for the employment of those displaced by the improved technologies, keeping some operations labour-intensive. It is therefore important to mechanise the key drudgery-alleviation equipment that can be easily handled by women.

Prime mover power is also a major consideration. Most villages do not have electricity and hence the diesel engine is the main source of power. Thus, for cost reasons there cannot be a multiplicity of these engines to drive the required unit operations. Where there is the need to drive several machines the answer could be to use diesel power to generate electricity. The cost and maintenance of this power source would eliminate most small-scale processors and communities. The power source in such instances acts as a limitation to the number of unit operations that can be mechanised and powered. Systems of pulleys and gears to drive operational machines should be actively considered when designing for village based groups.