THE MEMBRANE FILTER PRESS FOR
PALM OIL SLURRY CLARIFICATION
Mohamad Sulong
Palm Oil Research Institute of Malaysia
No.6, Persiaran Institusi, Bandar Baru Bangi, 43000 Kajang, Selangor
Ronnie C.W. Tan
Concept Engineering Sdn Bhd
Lot 7&9 Jalan PJS 11/22Taman Bandar Sunway, 46150 Petaling Jaya, Selangor
Click here for Publication list
ABSTRACT
Clarification of crude palm oil slurries by filtration followed by expression is reported in this paper. Pilot unit of membrane filter press with filtration area of 1.0 square meter was used in the study using fresh crude slurry ex-screw press as feed material. Feeding of slurries was done at 2 kg/sq.cm. pressure using air operated diaphragm pump and squeezing of cake was done by inflating the membrane elements at 2, 4 and 8 kg/sq.cm air pressure. These treatments produced cake with oil content of 18%, moisture content of 43% and solid content of 39% with suspended solid (as N.O.S.) recovery of more than 95%. Oil content in the cake was further reduced to approximately 3.5 kg/tonne FFB by initiating pre-washing before squeezing of the cake in the filtration cycle. This process is unique in it's ability to provide feed stream clarification and suspended solids dewatering. As demonstrated in this experiment, the entire operation is technically viable and the economics of a full scale plant is feasible.
INTRODUCTION
Most palm oil mills in Malaysia have utilized conventional process in clarifying crude palm oil slurry for more than fifty years. During 1996, PORIM and Concept Engineering Sdn Bhd initiated efforts to evaluate membrane filter press as a possible alternative to this conventional system as part of our strategies to minimize problems related to liquid effluent handling and treatments in palm oil mill.
In the conventional process, crude palm oil slurry is produced by mechanical pressing of the
cooked fruitlets. These slurry is then clarified in a vertical or horizontal clarifier. The oil layer
which is separated from sludge under gravitational force occupies the top portion of the tank.
The oil is then continuously withdrawn from clarifier by a skimmer. The sludge phase which
is compacted at the bottom part of the tank is composed of 6-12% oil which is then recovered
by a disc-stack centrifuge. In all instances, hot water is used to improve the settling behavior
of oil particles while the sludge separator operations require water to balance the phase
dynamics in the bowl. Under these circumstances, tremendous quantity of water is used and
discharged into waste treatment plant. It is estimated that 430 tonnes per day of waste water
is discharged by a typical 40 TFFBPD palm oil mill [
Another disadvantage of this conventional process is slow settling rate of oil particles which require rather large settling area for effective oil separation. Standard design of oil clarifier allows oil to settle within 4 to 7 hours under standard 30% hot water dilution. Under these circumstances, the oil is easily susceptible to hydrolysis, oxidation and other chemical reactions which deteriorate final oil quality.
In quite recent developments, decanting centrifuges are used for solid recovery prior to oil
clarification and as simultaneous solid recovery and oil clarification [
In keeping up with these drawbacks of oil clarification systems, PORIM and CONCEPT Engineering Sdn Bhd recognized that possibly a more efficient clarification system existed. Because of this, membrane filter press system was studied with specific areas of improvement over the conventional gravity clarification and centrifuge clarification.
An early application of filter press for palm oil process was noted by Van Heurn [
What is filter press
Filter presses are liquid/solid separation devices that utilize feed pressure and, in some cases,
squeeze pressure to reduce the liquid content in process or waste slurries or to reduce solid
content in a product (clarification). Pietila et.al. described briefly the theoretical approach to
the filtration equation using Darcy's law which was used by many researchers in evaluating
filtration equipment for a given duty [
Each face of plate is covered with a filter medium to form a series of perforated chambers into which process material is introduced under pressure. The filter medium retains suspended solid particulate. As a filter cake builds in the chamber, liquid flows through the drainage surface of the filter plate, and is collected from the filter through drainage outlets. When the filtration cycle is completed the closing force is released and the filter plates are separated, permitting the filter cake to be discharged from the chambers.
Membrane filter press filtration
Membrane filter press plates utilize an impermeable, flexible member attached to the filter
press plate core. This terminology should not be confused with submicron "membranes" used
in ultrafiltration or reverse osmosis. In filter presses, a membrane, also called a diaphragm, is
designed to inflate by compressed air or pressurized water to squeeze residual moisture out
of the formed filter cake. Polypropylene membrane plates have been in used for more than ten
years with significant success. These plates are currently used in many severe applications,
some operating two cycles per hour, 24 hours per day for two years, tens of thousand of
cycles without failure [
Conventional filter press filter only 75-80% of the batch in the first 50% of the cycle time. If the plate stack is opened before the full cycle is completed, the results is a wet, sloopy cake. A membrane filter press, however, can be stopped at the most efficient point in its cycle. By applying pressure to the membrane built into the filter plate, the cake is squeezed to achieve final solids content in a significantly reduced cycle time.
Membrane filter press operations allow a low-pressure pre-squeeze before cake washing is begun. This pre-squeeze eliminates channeling (cracks and less well-formed areas of the cake) and produces a uniform, homogenous filter cake and optimum wash efficiency.
By applying a high pressure squeeze to the filter cake (up to 100 psi with compressed air or 225 psi with pressurized water) one can produce the final cake with solids content from 50% to 90%, greater than can be achieved with conventional filter press. This is important for both maximum yield, where a high value liquid is your product, or high solid-content, where reduced moisture in the final cake is critical. The membrane filter press provides a dryer cake than can be obtained using high feed pumping pressure in a conventional filter press.
EXPERIMENT
Preliminary test work with a membrane filter press
for crude palm oil clarification
Since filter plate presses had not previously been used in the milling process, it was necessary to carry out preliminary test work on crude palm oil slurry to determine some of the operational parameters, such as filtration time, cake thickness, filter media type, filtration pressure etc. On-site investigations were carried out at Krau Palm Oil Mill using pilot equipment supplied by Concept Engineering Sdn. Bhd. This pilot unit comprised a three chamber 470mm X 470mm skid-mounted membrane filter press together with ancillary equipment, such as feed pump, compressor etc. Facility was available to change filtration pressure and squeezing pressure as required while cake thickness was maintained at 25mm and temperature of operation was maintained at 95oC. This filter press was used to process fresh crude slurry, thus simulating conditions on a full sized plant.
Apart from determining the filtration characteristics of the crude palm oil slurry, the tests were
important to provide information for correctly sizing the production filter press, together with
the ancillary equipment for the required duty at palm oil mill. Of all parameters to be
evaluated, the followings were considered to be most important: Filtration pressure between
2 -4 kg/cm
Although these are listed as individual parameter, they are interrelated, i.e., a change in one will have an effect on at least one other. Of prime importance was the need to determine the shortest possible filtration time in order to maximize the throughput of the filter press and minimize capital expenditure. While it is desirable to achieve the maximum cake thickness to aid cake discharge this has to be balanced against the longer filtration time.
RESULTS AND DISCUSSIONS
Typical performance curve of the filtration process using crude palm oil slurry were shown in Figure 1 for cumulative filtrate volume vs filtration time. Our test results indicated that oil losses in the cake after squeezing were ranged from 15% to 35% depending on the squeezing pressure and squeezing time. The final moisture content of the cake was correlated very well with the squeezing pressure and time. Separation of oil from water in the absence of solids was achieved within five minutes at 85oC. This indicated that a simple device such as liquid hydrocyclone can be used to clarify the oil from water within short retention time. Our calculations have indicated that solid recoveries were within 95-98% and it depends very much on the cloth type and the nature of slurry during the treatment. The absolute oil losses in the cake were calculated to be 3.5 kg PTFFB when cake washing was initiated before applying squeezing pressure. This figure is slightly lower than those normally encountered during conventional oil clarification process using gravity clarifier or decanting centrifuge. As far as oil quality is concerned, the peroxide value of the clarified oil was within one meq/kg when the mill is processing reasonably good fruit bunches.
With the above data we were able to design a system for typical commercial plant in palm oil mill.
PROJECTED ECONOMICS OF A FULL SCALE SYSTEM
Conceptual Design
The design bases for 20 TFFBPH are as follows:
Diluted crude slurry is allowed to settle partially in a pre-clarifier (15 tonnes) for 40 minutes. The oil is withdrawn using a skimmer and then purified using self ejecting centrifuge
It is anticipated that the oil content in the sludge is between 15%-18% under these circumstances
Approximately 6 tonnes of sludge is produced after oil recovery
Two unit of filter presses are required to ensure the process is close to continuous operation with each press having 25 membrane plates (1200 mm X 1200 mm). The cake thickness is 40mm and each filter chamber has 2.3 square meters filtration area. The required number of plates for a given duty is estimated from the pilot data performance.
An excess of 15% to the required filtration area is added to the calculated figures as a safety factor
Feeding into presses is done in two steps. Firstly the sludge from the middle part of the
clarifier is fed into the press using air operated diaphragm pump. This is done to eliminate
the sudden blockage of the filter media by the solids. The second phase feeding is initiated
before the first feeding is terminated to avoid sudden decrease in feeding pressure. The
heavy sludge is then fed during second phase using air operated diaphragm pump or using
a screw pump. Total filtering time is estimated to be 40 minutes, after which pre washing
is initiated for 10 minutes using hot water. Cake squeezing is initia ted in three stages, i.e.
using 2 kg/cm
Liquid filtrate is collected in a filtrate tank (8 tonnes) whereby oil and water separation is taking place under gravity. The oil layer is mixed with the clarified oil of the pre-clarifier. The water phase is partially discharge (20%) into effluent treatment plant and the balance of water is used for dilution purposes.
Filtration cycle for the second press is initiated soon after termination of feeding in the first press
Filter cake is collected in the receiv ing tank and it is conveyed out of the press unit using a conveyor
Capital cost estimation
The capital cost of a full scale system is derived in the following manner. First, all major auxiliary equipment and membrane filter press systems are sized based on pilot data. The auxiliary equipment such as pre-clarifier, filtrate collection system were costed through vendor quotes and catalogs. The filtration system costs are divided into hardware and membrane plate costs.
On- site engineering design and installation expenses are calculated as percentage of the equipment costs. Different multipliers are used for the auxiliary equipment, engineering and installation than that of membrane filter press. The percentage increases for ins ta l lation and design used ar e as follows: Administration & Supervision (2%), Contingency (10%) and inflation (10%).
Installed equipment cost and engineering cost are totaled and the main categories with the respective multipliers used in sequence to obtain the total installed cost were as follows: Auxiliary equipment (installation - 40% of equipment cost, Detailed engineering design - 12% of installed cost) and Filtration system & Purifier (Installation - 15% of equipment cost & Detailed engineering design - 5% of installed cost)
Building cost to house auxiliary equipment and filter press is not included in the estimate
The projected total installed capital cost is estimated at RM 2.6 million for 40 TFFBPH palm oil mill
Operating costs
The operating cost consists of several major elements. labor is assumed at 8 hrs/day working 2 shifts, 8 hrs/shift. Current labor expense is RM 15/day. Electrical power is calculated based on similar consumption of the conventional system. i.e. RM 2.05/tFFB. This corresponded to 10% of the total electrical power required by conventional mill. The only disposable item used in the process is the filter cloth. It is costed at RM 70/sq meter. Repair and maintenance costs are consisted of material and labor. maintenance material was calculated as 1.5% of the hard goods cost (total equipment cost less membrane replacement cost). The maintenance labor cost was assumed to be equivalent to the material cost. The final element for this cost category is insurance and taxes. This cost is assumed to be one-half of the material cost.
Depreciation expense has two element: membrane plate replacements cost and equipment amortazation. The membrane plates are depreciated over three-years life at their replacement cost. The other facilities are depreciated more than 10 years on straight line basis.
The process is assumed to be operating 16 hours per day, 305 days per year
The estimated operating cost for 40 TFFBPH throughput is RM 1.52/tFFB. The major cost factors are membrane plate depreciation, facilities depreciation and maintenance materials.
Operational cost saving of RM 600,000.00 is estimated if the filtration system is used in oil clarification process with the estimated payback period of 2.5 years.
Although the filter cake is considered as waste in this process, it is actually a good substrate for composting process to produce organic fertilizer or grow media.
CONCLUSION
From our study it can be concluded that the membrane filter press is a viable system for crude palm oil clarification. The economics of the full scale system as projected is cost-competitive. The entire systems offer an effective way of reducing BOD load into effluent pond while minimizing the volume of liquid effluent discharge from the mill.
REFERENCES
Olie, J.J. and Tjeng, T.D. (1972) Traitement et evacuation des eaux residuaires d'une huilerie de palme (Treatment and disposal of palm oil mill effluent). Oleagineux, 27, 215-218
Chan, K.W, P'ng, T.C. and Mohd A., M.R, Palm Oil Mill Effluent Utilization and its future research directions in the oil palm industry
Fanguin, J. (1985), Essais d'une decanteuse horizontale a vis convoyeuse Sharples en clarification d'huile de palme, Oleagineux, Nol 40, 4, 211-214
Jorgensen, H.K (1982)., The U.P. Decanter-drier system for reduction of palm oil mill effluent, PORIM Regional Workshop on palm oil mill technology and effluent treatment, 17th -18th Aug, Kuala Lumpur
Lim, C.L. (1987), The application of decanters for palm oil clarification, 1987 International Oil Palm/Palm Oil Conference, 29 June - 1st July, Kuala Lumpur
Sulong M. (1987), Application of scroll centrifuge to recover solids and oil from the underflow of crude palm oil settling tank, Proceedings of the regional symposium on waste utilization in Asia and the Pacific, ENSEARCH, Kuala Lumpur, 69-77
Van Heurn, F.C. The Preparation of Palm Oil, Investigations on oil palms, Ruygrok & Co., Botavia, 1922
Diprose, R.J.(1978), Effluent Research at Unipamol, Proc.3 MOPGC Symp. Treat. Disp. Palm Oil Mill Effl. 1977-78, 15-20
Pietila, K.A. and Joubert, P.J.( 1981), Examination of process parameters affecting sludge dewatering with a diaphragm filter press, Jounal WPCF, Dec, 1-9
Holyfield, G.W. and Brown, D.W. (1995), Application of a filter press in the recovery of potash at Boulby Mine, Filtration & Separation , May, 455-459