Work on the fuel polishing system began in earnest upon our return from the Bahamas in mid 2004. While I had been thinking about and evaluating fuel polishing technology and approaches since 2003, our experience with dirty fuel and fuel tanks on our return trip from the Bahamas galvanized us into action. A rough crossing of the gulf stream stirred up a lot of debris, and our off-shore journey back to the Chesapeake Bay was cut short as we were running out of Racor filters!
I have heard different explanations for the origin of the sludge-like deposits that can occur in fuel tanks. Explanations include overwhelming infestations with micro-organisms (bacteria, fungi, yeast) and aging-related molecular transformations of the diesel fuel to various insoluble gum/resin/varnish compounds (also called asphaltenes). Aside from clogging filters, these processes can produce acids and other corrosive chemicals that can affect various engine components, as well as damage fuel tanks.
Addition of biocide to fuel is said to be preventative, but it is not a panacea. Micro-organisms can develop tolerance to the effects of biocides, and treating a heavily contaminated tank can promote precipitation of organic debris. I, like many others, believe that inspection and cleaning of fuel tanks should be part of routine maintenance. Biocides are toxic, and many of the chemicals are cancer-promoting. I worry that the presence of biocide residues in fuel tanks adds significantly to the health risks of carrying out this routine maintenance. Thus, I was anxious to implement a preventative program that did not rely on biocides.
POLISHING SYSTEM COMPONENTS
The fuel polishing system had several basic objectives: 1) Keep the fuel and the tanks clean; 2) Provide sufficient flow capacity to process a full load of fuel (1150 gallons) in a reasonable time; 3) Automate the priming of fuel lines and filters; 4) Allow easy transfer of fuel between tanks for trim adjustment.
The polishing system I installed consists of the following components: 1) A continuous duty 120 volt a.c. 5-diaphragm pump rated at 3.3 gallons per minute, purchased from ESI; 2) Racor 1000 FG Turbine Series filter, equipped with 2 micron filter elements; 3) De-Bug Magnetic Fuel Treatment Unit, model L1000 (rated for 4.4 gpm). Alex Marcus, ESI's owner and chief engineer is very knowledgable and was a great help with overall system design.
In addition, I plumbed in a small Walbro 12 volt fuel transfer pump equipped with a 4 psi cut-off switch. This pump is used only to pressurize fuel lines for priming purposes -- very useful when engine fuel filters get changed or if lines need to be bled of air.
The effectivness of magnetic fuel treatment in general, and De-Bug units in particular, is debated amongst boaters. De-Bug's website has a good review of the evidence supporting the use of their units. The theory is that exposing diesel fuel to multiple magnetic fields of alternating polarity kills micro-organsisms by disrupting their cell membranes.
I debated the use of a Racor versus a Separ filter for the system. My decision in favor of Racor was based on two points: 1) I wanted to use 2 micron filter elements, and I have been unable to obtain Separ elements in anything finer than 10 microns; 2) Racor filter elements have a hydrophobic coating that helps to remove moisture, whereas Separ elements lack this coating.
Duet's 4 fuel tanks each have 3 access pipes: Two pipes ('A' and 'B') terminate 1.5 inches above tank bottom. The third pipe ('C') is for fuel return and terminates at the top of the tank. The 'A' pipes on each tank were already plumbed to a fuel distribution manifold serving the main, wing, and generator engines. I decided to use the 'B' pipes plumbed to a separate manifold for fuel supply to the polishing system. This allows me to run the engine off one tank, while polishing a different tank. If we are underway, I return polished fuel using pipe 'C'. However, if we are stationary, and the fuel is not being agitated by boat motion, I return polished fuel using pipe 'A'. This is accomplished by flowing polished fuel 'backwards' through the fuel distribution manifold used by the engines (ball valves to the engines get closed first). The use of pipe 'A' creates a jet of flow which stirs up debris and water which have settled on the tank floor. Use of a return near the tank floor is recommended by ESI. A schematic of the system is found at this link. To get maximum fuel flow from the polishing system, it is important to install properly sized hoses, especially on the suction side of the pump. Data from Parker Hannifin indicated that hydraulic systems pumping 3 gpm should have at least 5/8" I.D. suction hoses. This recommendation has a built in safety factor since hydraulic oil is more viscous than diesel fuel. I used Parker FR221-12 hose with series-20 JIC37 female hydraulic fittings (see more on JIC37 below). The Parker series-20 line of hose fittings can be installed without a machine swage, just a sturdy bench vise and wrench. They avoid the inconvenience of taking all your individually cut lengths of hose to a shop that can machine-swage fittings. Also, they can be re-used (although for some high pressure hydraulic work, it is not recommended to re-use these types of fittings).
JIC37 (more properly called SAE J514, 37 degree flare) refers to fittings that have sealing metal surfaces that are flared at a 37 degree angle. They are widely used in hydraulic systems and provide a vibration-resistant fluid-tight seal at high pressures without need for thread sealants. These types of fittings are in use throughout Duet's engine room, and they function quite well. An alternative type of fitting recommended by ESI is the Straight Thread O-ring Boss. These fittings use the same threads as JIC37, but a sealing o-ring takes the place of the flared metal surfaces. The inlet-outlet ports of the Racor filter and De-Bug units are designed to accept these types of fittings. Other than those two components, I chose not to use O-ring fittings elsewhere because I found sourcing more difficult and costly.
3/4" full-flow ball valves were used throughout the system, to avoid losing the advantage of low flow resistance offered by the large internal diameter hosing. Full-flow ball valves have less flow resistance than regular ball valves of the same size. The manifolds used in the system were also for 3/4" plumbing, as were all the hose fittings.
A note about hydraulic hose fitting size designation is in order. Begin by realizing that the nominal size of most plumbing fittings (not just hydraulic) is not the same as what you would measure with a caliper. For example, a male pipe fitting with nominal size of 3/4" would be closer to 1.06" if you measured the outer diameter with a caliper (the female fitting gets measured on the inner diameter and is a snick smaller). The hydraulic industry has another wrinkle in that they use a dash number which expresses nominal size in sixteenths of an inch. So, in the case of the female JIC37 fittings I used, there were three size designations: Dash number 12, nominal fitting size 3/4", and actual male thread measurement of 1 1/16-12. The 12 following the actual thread measurement refers to threads per inch, not dash number. The following table, excerpted from the Dayco Hydraulic Hose and Coupling Catalog, illustrates this nomenclature as it applies to variously sized JIC37 fittings.
The dash numbering system is also used to express the actual measured internal diameter of hydraulic hose in sixteenths of an inch. For reasons I do not fully understand, just a few hydraulic hoses do not conform to this standard. The Parker FR221 hose which I used is one of those rare exceptions. With a dash number of 12, the I.D. should have measured 3/4", but in fact it is 5/8".
If you are confused, as I was, by the complex nomenclature surrounding hydraulic hose and fittings, the 'Dayco Hydraulic Hose and Coupling Catalog' is a great resource. It also describes best practices for installation of hydraulic hose, which is quite interesting. You may download a copy from the Parker Hannifin Hose Division website. The pumps, Racor and De-Bug units, and manifolds were mounted on Starboard which was then attached to a wall in the engine room. I wanted to be able to thru-bolt the various components to the Starboard, but I also wanted an easy way to detach these components in the future without pulling the Starboard off the wall. The solution was stainless T-nuts which allow one to permanently install a female threaded fitting inside a pre-drilled hole. With the T-nuts installed, all the various polishing components just get bolted to the Starboard using ordinary machine screws. They are effectively thru-bolted since the T-nuts have a flange which spreads the load out behind the Starboard panel, much like using a standard nut and washer. Pictures of the mounted system are here and here. As part of this project, I color coded all the valves according to the fuel tank they serve, to improve my ability to easily determine fuel routing. The manifold which controls return of polished fuel is mounted separately, above the return manifold used by the engines. Plumbing connections between the valves and manifold use standard pipe threads, not hydraulic fittings. To insure a good seal, I used Loctite 569 thread sealant.
ESI strongly recommends use of Pri-D fuel treatment, in conjunction with a regular schedule of fuel polishing. Pri-D is manufactured by Power Research Inc. The purpose of this additive is to prevent the aging-related molecular transformations of diesel fuel referred to earlier. It is not a biocide. I have found various suppliers on the internet.
Aside from regular use of the polishing system, and periodic tank inspection/maintenance, it helps to be scrupulous about the seal on your diesel thru-deck fill caps. It is easy for dirt, as well as wear and tear, to degrade the seal on the o-ring. Whenever I open a fuel fill cap, I carefully clean the threads and sealing surfaces. A light coating of silicone plumber's grease on the o-ring helps to insure a water-tight seal. If I have absolutely any doubts about the appearance of the o-ring, it gets replaced immediately. These steps should help to insure that sea and rain water do not gain access to the tanks.
You can check on the effectivness of your water-excluding methods by using a fuel-impervious water indicator paste on the end of a long stick. Poke this down to the bottom-most point of the tank, and see if there is color change. One such paste I have used is "Kolor Kut" (Kolor Kut Products Company, Houston, Texas).
Lewis Marine: Parker hose and hose fittings. Walbro fuel transfer pump.
ESI: 120 volt continuous duty 5-diaphragm fuel pump. De-Bug unit. Straight thread o-ring boss adapters for Racor and De-Bug units.
McMaster-Carr (mcmastercarr.com): Excellent on-line industrial hardware supplier. Various plumbing components including ball valves and adapters to use between hydraulic fittings and standard pipe fittings.
Fuel manifolds were obtained from Magnaloy. The units are aluminum header bar manifolds.