The engineering group here has been discussing kidney loop pumps and tank turn-over. What is the general rule of thumb regarding tank turn over, or pumping the tank contents through the kidney loop. For example, if I take the simple approach, given a 300 gallon tank and a 30 GPM kidney pump, I would turn the tank in 10 minutes. But that assumes no mixing of oil that has been passed through the loop with oil that has not yet been through the loop. I realize that the actual results depend a lot on tank design, but as a general rule of thumb, what would I expect for a tank turn over rate?

I've had excellent results with Filtering Hydraulic Sysytem with OffLine Filter Circuits. That is the main type filtration I design into almost all hydraulic circuits to supplement other required filters such as those required for Proportional or Servo valves.

It has been my experience that filtering the oil at a 3 Micron level at a complete tank change interval of 4-8 hours works very well,

I also use very large Filter Packs so oil flow velocity is just a trickle through the elements and they are electrically separated from the systme run controls. That way they run 24/7 and do not require a system shut down to be serviced.

On a 300 Gallon tank that was operating in a DIRTY atmosphere it took about 30 days to clean the oil to where you could shine a light in the tank and see the bottom like there was fressh water in it.

On that circuit the pumps went from changeout every year, at least, to over three years at last check with no loss in producton time.

Same company then put a similar system on another DIRTY 6,000 Gallon Tank with similar results.

Both of these systems were in atmospheres that were super contaminated with some nasty grit. Also both had much higher than average oil deoth change since they powered some Single Acting Rams. Another thing that was done is the Tanks were sealed and the air being transferred back into the Tank during a normal cycle was also filtered to the same level. All possible other places that could allow air to enter were sealed all around the Tank.

The old saying "Pay me Now or Pay Me Later" can certainly be applied to filtration of a Hydraulic Circuit.

Thanks for the feedback Bud.

We use a "clean tank" philosophy where absolutely no fluid gets into the tank without going through a filter. High pressure filters are on the pump outlets and anything returning to the tank enters through a filter. The case drain goes through a dirt box with a high surface area screen so that there is no pressure drop but protecting the tank so in the event a pump comes a part we don't fill the tank with metal shavings. An offline filter/cooling loop has Hand/Off/Auto modes. Auto mode turns on automatically anytime a pump motor is started and it runs for about 15 minutes after the pump shuts off, but we have recently increased this to 1 hour on some systems. Tank filling uses the kidney loop. A three way valve on the kidney pump suction is switched to the barrel suction hose so that new oil passes through the off line filters before it enters the tank. We also perform monthly lab analysis of our oil.

So far our clean tank philosophy has paid off - I count 78 hydraulic pumps here (not counting the itty bitty integrated OEM packs on some machine tools) but we only replace 4-6 pumps a year. Except last week we replaced 5 pumps, sometimes IT happens in large quantities.

Anyways, even though we increased the time for the kidney pump to run one hour after the main pump shut off, we have been discussing whether to run the kidney loops forever, never turning them off even when the machines are idle except for maintenance and whether that would add any benefit and/or risk. The plant runs 24/7/360.

Alaric;

You can see my Training book take on the subject of Filtering hydraulic circuits at this link:
http://www.hydraulicspneumatics.com/200/eBooks/
Go to Ch. 7 Pg. 6 to see my recommendations for the ultimate filtration setup. All the filter locations may not be required on all circuits but need to be checked out on a one on one basis.

Cheapest insurance anyone can buy when it comes to hydraulic circuits.

Often a hard sell to the end user since many have seen the oil that comes out of the crankcase of their vehicles engines and believe oil that much contaminated will work for high pressure hydraulics also. Seems the Almighty Dollar of the First Cost of the circuit is what a buying decision is based on.

However, once they see the results they change their buying spec's., if they have them, to reflect their new found knowledge.

But, that is what you would expect with a regimen that is sold on lowest bid ater being designed by the untrained or those who are somewhere between starting and graduating from the College of Hard Knocks.

I wanted to find a rule of thumb to estimate how long it takes to actually pass all of the oil through an off-line filter when oil that has been filtered mixes with the unfiltered oil upon return to the tank. It occured to me this afternoon that if I assume that as soon as filtered oil is returned to the tank it mixes evenly into the unfiltered oil and doesn't stratify and that if the tank is always evenly mixed then 1/2 of the oil the kidney pump pumps has already been through the filter at least once or more - so that gives me a worst case to estimate from.

So

V*.5^N = u
Where V is the initial volme of the tank.
u is the amount of unfiltered oil.
N is the number of times I have to pump the tank volume.

.5^N = u/V

Lets deal with u/V now, I wanted a rule of thumb for having confidence I have filtered 99% of the oil in the tank at least once. Expressing unfiltered oil as a percentage of tank volume I get .01V, and .01V/V gets rid of the tank volume from the equation and leaves me with just .01.

.5^N = .01
ln(.5^N) = ln(.01)
N*ln(.5) = ln(.01)
N = ln(.01)/ln(.5)
N = 6.6.

I'll choose 7 for my rule of thumb,
So if I have a 300 gallon tank and 30 gpm pump, then tank volume / pump flow rate = 10 minutes.
Using my rule of thumb then I can say that in 70 minutes 99% all of the oil in the tank has been through the filter at least once.

Good tank and baffle design will improve the time dramatically, but does 7 * tank volume/pump flow seem to be reasonable for a rule of thumb?

This message has been edited. Last edited by: Alaric,

These web sites are exellent reading on oil cleanliness in hydraulic systems.

http://www.oilanalysis.com/article_detail.asp?articleid=194

http://www.noria.com/

quote:

So if I have a 300 gallon tank and 30 gpm pump, then tank volume / pump flow rate = 10 minutes.

The value, 1 min, you just calculated is the filtering time constant. 5 Time constants should do it ideally because because of imperfect mixing 7 is a good choice.

quote:

Using my rule of thumb then I can say that in 70 minutes 99% all of the oil in the tank has been through the filter at least once.

Yes.

quote:

Good tank and baffle design will improve the time dramatically, but does 7 * tank volume/pump flow seem to be reasonable for a rule of thumb?

It does to me.


I don't understand your .5^N calculations
it should be more like

Contamination=InitialContamination*exp(-flow*t/Volume);

It is the solution to a simple first order differential equation.

edit, the volume should be the whole volume of oil in the tank and oil in the circuit.
If you meant tha half the oil is outside the tank and unfiltered then the volume should be 600 Gal. If you are just trying to calculate the filtered vs unfiltered oil then the equation would be.

UnfilteredOil(t)=(UnfilteredOil0/(Filtered0+unfiilteredOil0))*exp(-flow*t/(FilteredOil0+UnfilteredOil0);
The variables with the 0 at the end are the initial values.

This message has been edited. Last edited by: Peter Nachtwey,