Join Date: Jul 2000
Thanked 125 Times in 115 Posts
Rep Power: 15
I run an onboard recirculation polishing system that I designed myself. Ive been involved in filtration engineering for the past 30 years.
I have a black iron 110 gallon fuel tank.
I started with a fully cleaned and mechanically scrubbed tank. The recirculation polishing system is fully independent of the main fuel delivery system: Independent dip tube that runs from the VERY bottom of the tank (to pick up free water and settled particulate ---> Walbro 12v. 3 gpm transfer pump (with integral suction screen) wired directly to the engine panel ---> 5uM industrial (10" long X 2.5" dia.) polypropylene spun bonded filter) ... filter housing has drain port to which I attached a clear (tygon) drain tube to visualize if water is in the bowl of the carbon steel filter housing (housing has 0-15 psi pressure gage to monitor filter condition) ---> back to tank. System will 'turnover' 100 gallons in approx 30 minutes (due to pressure drop losses). Constant turnover of tanakge will yield the resident particulate in the tank to submicronic levels (a 5uM filter will have a % of captuire sites in the submicronic range) ... I verify the particle resident ranges by nephalometry - an optical correlation to actual particle size. So in approx 8 hours of run time, I filter approx. 2000 gallons of oil through the recirc. system. I change the filter at a differential operating pressure of about 8-10 psid.
EVen if the 'worst' hapapens ... I pick up a load of bad fuel or I get a particle storm break loose from the tank walls, the RECOVERY of the system back to sub-micronic levels is quite FAST.
The advantage to a high 'turnover system' is that the oil is continually filtered at relative high (3 gpm) volumes to take advantage of the essentially % submicronic removal ability of a 'coarse' 5uM filter. When the filter gets 'dirty' the effect is even greater (using trapped dirt to filter more dirt). The reason for this is that if the submicronic particles are continually removed there are no 'nucleation sites' onto which particles will grow (agglomerate).
I do thoroughly mechanically scrub/clean out the tank every two years. I also carefully check any oil that I put injto the tank: I put the oil into a clear glass and hold it up between my eyeball and a strong white light, if there is any 'cloudiness' to the oil ... I close up and simply go somewhere else. I usually ONLY buy fuel from a high turnover source ... such as a fuel depot that services 'watermen'; if possible NEVER from a marina.
The Main fuel system used to have a dual block and bypass 30-->10---2ÁM Racor but Ive since removed all this 'complexity' with the exception of one single LARGE surface area 2uM Racor ... which simply doesnt get any challenge of particles due to the effectiveness of the recirculation system. The 2ÁM Racor is fitted with a block and bypass which allows me to switch to the bypass if I need to change the Racor (the 2ÁM engine mounted 'guard filter' temporarily doing ALL the main system filtration). The main system then delivers to a 3 gallon 'day tank' so in the worst case scenario I can switch solely to the (gravity feed) 'day tank' and run at WOT for about 4 hours until I sort out (later) any problems in the recirc or Racor. The main fuel delivery system is PRESSURE FEED with a 12v. diaphragm pump AT the tank and all the feed lines are double flared stainless tube.
The Walbro was ~$175.00
The carbon steel 10" X2.5" was about $80.00
5ÁM Osmonics polypropylene spun bonded (GX05-10) @ $8.00 ea.) .... dirt capacity is about 30 grams.
All the piping is stainless steel double flared. .... as copper is reactive with diesel fuel and compression fittings usually eventually leak (and suck air).
To prevent water 'uptake' of the fuel, I have installed a (regenerable Silica Gel with indicator dye) desiccant filter/trap on the tank vent line. Diesel fuel is quite dry when cracked/distilled but if left open to the atmosphere will pick up water by chemical equilibrium .... through the vent. The desiccant filter/trap prevents such water 'uptake'.
My main system is is pressure feed (12 v. diaphragm pump AT the tank) --> 2ÁM Racor (with block and bypass) ---> engine 2ÁM 'guard' filter ----> daytank ---->engine. Pressure feed is **vastly more efficient** for filtration plus as if a fitting begins to leak I dont 'suck air' (but I run a bombproof stainless piping system so that leaks are virtuallly impossible ... as I dont need to fill the bilge with oil or get huge fines from the CG).
Hope this helps.
Last edited by RichH; 07-04-2007 at 12:02 PM.