Purifiers and Clarifiers differ only in that clarifiers are not set up to remove water. Their design are similar to the point that most purifiers found on board can be converted to use as a clarifier with simple alteration of the gravity disc
If an oily water mix is placed into a tank then separation of the two parts will
begin with the lighter element rising to the top. The rate the separations occurs
is governed by several factors including the difference in specific gravities and
the force of gravity acting upon it.
For mixes placed into a settling tank there is little that can be done about the gravity but the difference in the specific gravities can be increased by heating. This because water density changes at a much reduced rate when compared to oil. The limiting factor to this is that the water cannot be heated above 100'C for obvious reasons.
A wide shallow tank will increase the rate of clarification over a tall thin tank
Principles of operation
When a volume of light oil is placed into a tank contain a weir and a quantity of water the fluids will tend to arrange themselves as shown above. The height of the water in the weir rises to a point governed by the volume ( and thereby relative height) and specific gravity of the light oil.
Knowing this it is possible to form a rudimentary purification system
As a oil/water mix is fed into the tank separation begins with heavy particulates falling to the base of the tank along with water which joins the other water excess overflowing the heavy phase weir. Hopefully clear oil passes over the light phase weir. The problem arises that to ensure their is sufficient time to allow for full separation of the oily mix the flow would have to be very small relative to the size of the tank.
Principle of separation in centrifuge containing angled plate stack
Fluid moving between two plates has a velocity greatest at mid point and minimum
approaching the plates.
a particle entering into the plates will tend to be pushed upwards by the fluid flow. All the time centrifugal force tends to retard the horizontal component of the movement causing the particle to approach the underside of the top disc. As it approaches the fluid flow velocity reduces. The centrifugal force eventually overcomes the force acting on the particle due to fluid movement and the particle starts to move towards the outer rim. The centrifugal force acting on a particle is proportional to its mass therefore a small particle will tend to move further under the influence of fluid flow. Indeed a particle small enough will be carried through the plates and out with the discharge. In this way it can be seen that reducing the flow rate to a purifier will tend to increase the quality of the output.
The basic centrifuge differs than that described above most obviously by sitting
on its side. In reality it takes the form of a round bowl a cross section of which
will show something like that seen above. Gravity is replaced by centripetal force
as the bowl is spun at high revolutions thereby creating very high g-forces.
A disc stack is incorporated to encourage a laminar flow improving the separation effect. Dirty oil is introduced via a centreline oil feed dip tube. The oil is led to distribution holes which are reflected in the disc stack but not the dam
The following factors are of importance when understanding the function of the purifier
Choosing Gravity Disc
The graph shown above is one typical of one found in a purifier instruction book for selecting appropriate gravity disc size. Shown on the diagram is an example of an oil of sg 0.93 at 0'C. The sg at 15'C for use with this graph is found by projecting along a horizontal line to 15'C. This step would be omitted if the sg at 15'C was already known. A line is then drawn parallel to the pre-drawn sloping lines. Where the drawn sloping line cuts the appropriate oil supply temperature isothermal then This becomes the selection point for the disc. This is found simply by ascertaining which size band the point lies in.
Self cleaning centrifuge
The majority of purifiers found on board are of the self cleaning type in that they are able to open the bowl to discharge any accumulated sludge. Apart from the sliding bowl the main difference is the centripetal pump over the simple design. In this a fixed centrifugal style impeller is mounted in the light phase outlet drawing the oil and discharging it at pressure sufficient to deliver it to the receiving tank. A discharge valve is fitted which is adjusted to give a constant back pressure in the bowl. The adjustment of this back pressure tends to move the position of the interface but more importantly increases the oil in the light phase delivery chamber increasing the immersion depth of the lip of the pump. This reduces possibility of air being entrained and removes foaming.. In the event of bowl failure back pressure will fall, this may be detected by a pressure switch initiating a shut down
For the bowl shown above a typical sequence of events would be
Typical alarms and shut downs
The following gives a general list of alarms only some of which may be fitted.
Sharples constant sludging
Here's one to send a shiver down the spine of anybody of my age.
This consisted essentially of a standard non desludging bowl into which were drilled small holes on the circumference fitted with nozzles. Seal water was pumped continuously from a small catchment tank mounted adjacent to the purifier into the bowl where it passed though to be ejected through the jets. It then drained to the catchment tank. Dirty oil would float to the surface where it would overflow though a surface mounted skimmer to the sludge tank. Theoretically the bowl could run for considerable periods without cleaning. The reality was one to two weeks, bowl cleaning included patiently trying to clear the small bore nozzles. I remember loading bunkers which were brick red, it contained lots of sand. The purifier was permanently overflowing to an extent we had to use the second purifier to run on the sludge tank. Bowl cleans were every day with the other engineers playing the 'it best if one person concentrates on them' card. The worst aspect was the Chief Engineer who used to lie in wait for me when I was called out during the night. On sneaking back to for some sleep he would drag me into to his cabin ( which was next to mine) for a thank you drink- this inevitably lasted until 8 am
The most obvious trend is that towards online sludging. In this during normal operation a small quantity of extra seal water is added and the bowl opened for an extremely short period of time thus removing the need to interrupt the process.
Control and Operating water
Water must be supplied at a fixed pressure to ensure that the quantity supplied to the purifier is constant for the set parameters. The water normally comes from the vessels hot water system or is independently heated to reduce thermal shocking and to prevent cooling of the hot oil
Considerable torque would be required to direct drive the bowl upto speed using an
appropriately sized electric motor. In addition very high loading would occur on
the gear train, to prevent scuffing due to oil film breakdown would require large
mating areas thereby large gear trains which would again increase the starting load.
A centrifugal clutch arrangement is fitted which has between 2 and 6 ferrodo lined brake pads. These are designed to slip during the start up period and also to a much lesser extent during the speed up period after de-sludge. Purifier manufacturers will usually quote a maximum and MINIMUM start up time. As the pads wear it may be necessary to remove and restore the mating surface to keep the start up time correct. As a last measure the number of pads should be altered
The electric motor may be of special design allowing for a long period of slight overload during the start up period.
The gear train is generally a single stage worm and wheel arrangement with the wheel being made of a softer material. Lubrication is normally splash only, the viscosity of the oil is essential to prevent wear as the form of lubrication is mainly boundary therefore the wear is governed by the viscosity and additives contained within the oil.
When wear occurs it will be scuffing and relative movement between the mating faces polishes out any pitting. As wear worsens galling occurs destroying the running surface. This damage is reflected in both elements therefore both should be changed.
As well as overload other causes of premature failure are poor design ( step forward
westfalia), poor material choice, poor lube oil choice, too long a de-sludge period
relative to supplied oil quality, out of balance bowl, failing bearing set in particular
the vertical shaft upper resilient bearing arrangement
The use of planned maintenance is essential particularly with respect to bearing changes. It is strongly recommended to monitor condition using vibration analysis
Should be carried out at regular intervals not exceeding manufacturers recommendations. Every care should be taken not to score the surfaces of the bowl especially the sliding surfaces for de-sludging types. The disc stack is generally numbered and should be built up as per this system as the stack is a balanced unit.
This was a techniques employed some time ago to improve purification of lube oil and to remove acids. It involved continuously adding a small quantity of water at oil temperature to the oil inlet which would pass through and overflow. This is much out of favour as it tends to remove the essential oil additives in particular detergents. An alternative is to inject steam which improves the removal of colloidal carbon by causing it to coagulate
Shown is a typical circuit for a lube oil system although it can equally be applied to a fuel system. Control is achieved by the three way valve which either diverts oil to re circulate or sends it to the purifier.
Oil flow rate is controlled by the oil control valve situated before the positive
displacement delivery pump which is driven off the purifier horizontal shaft via
a weak link arrangement
Back pressure from the purifier is controlled at outlet via the back pressure control valve
Shown is typical damage to the sealing face of a sliding bowl. This has been caused
by either poor assembly or by hard material being trapped aft the bowl closes. Unfortunately
it is more likely to be the former.
Failure is detected by loss of sealing water as seen down the sludge shute. In addition there will be carryover to the heavy phase and loss of discharge pressure as the seal is exhausted.
Operation of a Centrifugal oil purifiers
The following description relies on the reader having an understanding of the function and internal design of a modern self cleaning centrifugal purifier.
The method describes the operation of a manually controlled system. It is accepted that the vast majority of units are automated, it is intended allow better understanding of the automated process by doing this. The " Automation " section will clarify this.
The unit described is a sliding bowl type, does not have on line sludging capability with light phase requiring heating. Operating water is required for both opening, and closing the bowl. It is accepted that in common designs only opening water is supplied, once supplied this water leaks out until it reaches a level determined by a weir arrangement. The water remaining beyond this weir acts to close the bowl
Operating, displacement, seal and sludging water are delivered in finite quantities governed by the type and size of bowl, and the supply water pressure. Although not mentioned it should be taken for granted that water introduced is of a set quantity
Starting and stopping
Before starting the purifier the correct sized gravity ring must be in place for the light phase specific gravity at required delivery temperature.
Most units have a brake arrangement fitted acting on the bowl to slow during stopping. This should be check to be disengaged.
The use of this brake is determined by manufacturers recommendations. The brake should be seen as an emergency device to slow the bowl in the event of some problem, typically an out of balance of the bowl caused by sludge not being evenly removed from the bowl.
Where no preference is given it is the authors recommendation that the brake be used. The reason for this is to allow the bowl to pass as quickly as possible through any critical vibration harmonics.
Correct operating and sludging water supply pressure should be checked
The light phase (oil) supply, discharge valves may be opened, the flow control valve and back pressure valve may be set at approximate initial settings. The three way valve is set to recirc (dump)
Remember that when the oil is introduced to the bowl the bowl is cool and so the oil will be cooled and will tend to pass over with the heavy phase. This action may be reduced by having the back pressure discharge valve slightly more open than required closing as the bowl warms.
The bowl may be started. Drive to the bowl is via a centrifugal clutch arrangement reducing the starting current on the motor. The start up time is determined by the slippage of this clutch which is in turn determined by the number of pads. The oil is now being circulated though the heater. The heating medium may be introduced to bring the oil to purifying temperature
It should be noted that for some designs the motor is non-standard being able to carry high starting currents over a longer period. When requesting replacements this should be noted.
Too few pads causes an overly long start up time. In addition, the drag of the liquid as it leaves the bowl during a sludge cycle causes a reduction in speed. For automated systems, this speed must be regained before the oil is reintroduced otherwise carryover can occur. With some designs correct sludging is determined by an expected rise in drive current caused by the motor trying to bring the bowl back to speed. Too few pads may cause problems in both these cases.
Too many pads leads to excessive force on the drive gear leading to premature failure especially of the wheel and the electric motor.
When the bowl has reached it operating speed the bowl may be closed by introducing closing operating water.
this may be checked by viewing the external speed indicator or by observation of a reduction in amps as the bowl reaches its operating speed. The best method is a combination of the two. Observe the amp reduction and visually check the speed indicator to confirm that a fault has not occurred in the clutch.
Once correct oil temperature is observed the purification process may begin. Seal water is introduced to the bowl. The three way valve is operated to supply oil to the bowl. Correct flow and back pressure should be set once full flow is achieved. The amount of heating medium to the heater should be increased as required
The sludge port should be viewed to ensure no leakage from bowl- remember to close port before sludging. Check heavy phase (water) shute to ensure no oil overflow. Check operating water shute to ensure valves have isolated
It is unlikely that the bowl will immediately come " on-line ", this is generally due to the bowl being cold, cooling the oil, increasing its specific gravity and causing it to pass over the heavy phase (water) shute.
The amount of carryover may be reduced by opening the back pressure valve. The best solution is to removed the cool oil that is in the bowl by sludging. This may be repeated couple of times before the bowl has reached operating temperature.
The correct function of the desludging mechanism should be checked.
Once the unit has been proved on line and operating correctly the alarms and shutdowns should be tested. Where the units overflow to a sludge tank the correct operation of the alarm should be checked.
The testing of the alarms and shut downs is paramount. Once completed, especially for main engine lube oil purifiers, a note should be made in the engineroom log book
It is common to find dedicated sludge tanks for the purifiers. The level in which is kept artificially high and just off the high level alarm. In this way the tank acts as a back up alarm for the purifier.
The vast majority of units will be fully automated for UMS enginerooms. Sludging will involve pressing a single button. A controller will then cycle through the operating, displacement and seal water valves as well as operate the three way valve. The heater will have an independent controller although a zero output signal may be generated by the purifier controller during the sludge period.
Detection of correct desludging may be by drive motor current, as discussed, flap arrangement which is struck by the discharge from the open bowl or by measuring the discharge pressure which falls to or below zero pressure during the sludge cycle.
The use of Homogenisers both in initial installation and retrofit found favour with
several ship owners due to perceived benefits. Service experience has revealed a
variance on the results perhaps attributed to the appropriateness of the installation
The Homogensier is placed in series immediately before the purifiers
A typical homogeniser consists of a drum containing a number of disc stacks. The disc stacks are spatially contained in a rigid frame but may rotate on their axis. The rigid frame is driven causing the disc stacks to rotate around the casing.
Fluid is flowed into the the Homogeniser where it is acted upon by the disc stack and casing mounted tyre. Mechanical impact and Shear forces act on the fluid breaking down impurities and larger components. Water not removed by the purification process is finely integrated into the fuel
The theoretical effects of this is to improve combustion and thereby fuel efficiency.
In service investigation has failed to demonstrate real improvements in SFC
Reduction in Purification efficiency
Purifier manufacturers such as Alfa Laval strongly advise against the use of Homogenisers as they have an adverse effect on the separator efficiency. Particularly water is very difficult to separate once the fuel is passed through a homogeniser
Reduction in Sludge
The Homogeniser is able to process sludge which would otherwise be lost via the purification process. This may lead to an increase in overall fuel consumption but a reduction on the requirements to discharge waste and ease the load on the incinerator.
Reductions up to 70 to 80% are claimed
Water Injection and reduction in NOx
research was carried out whereby water was deliberately metered into the fuel entering the homogeniser. The theory behind this was that a water nucleus would form in the fuel droplet when injected into the combustion chamber. The water would rapidly expand and flash off blowing apart the fuel droplet and thereby improving atomisation.
In service investigation has failed to demonstrate real improvements in SFC
A reduction in Nox is achievable approximating to 1% for each 1% injection of water. This is associated to a lowering of combustion temperature
Soot and particulate emissions
Noticeable reductions in in-service soot emissions have been found
Extracts taken from "Green Efforts for Existing Ships- Fuel Pre-treatment summary report" published by Marintek Sintef group