Van's Air Force
Don't miss anything! Register now for full access to the definitive RV support community.
Trapped air in oil cooler
- Thread starter Bavafa
- Start date
Even after an initial flush/bleed, it is common for Heat Exchangers and filter containers to accumulate entrained air. The resulting condition is called "air bound" and for heat exchangers, not all of the available surface area is utilized and the heat xfer goes down proportionately. In industrial applications, it is very common for HExs to implement a trickle vent at the top of one of their headers. As for our applications, try to have at least a little slope with flow through the cooler bottom->up. You're gonna get anecdotal "evidence" here to the contrary. That's because our components tend to have (sometimes significant) margin. Best to keep margin as margin. Up front consideration of the aforementioned can save you some frustrating rework later. My $0.02
AlexPeterson
Well Known Member
Take the example where the cooler is mounted horizontally, with air going vertically through it: I suspect that there would be very little air trapped.
If oil were very, very (glacially!) slowly pumped through an empty cooler at essentially atmospheric pressure, one can see that there would be air trapped above the ports. However, oil is pumped quite vigorously through the cooler, and it also at around 5 atmospheres of pressure. The oil will take the path of least resistance, which means it would prefer to flow where there is only air, since the viscosity of air is extremely low compared to oil, thereby pushing the majority of the air through.
If oil were very, very (glacially!) slowly pumped through an empty cooler at essentially atmospheric pressure, one can see that there would be air trapped above the ports. However, oil is pumped quite vigorously through the cooler, and it also at around 5 atmospheres of pressure. The oil will take the path of least resistance, which means it would prefer to flow where there is only air, since the viscosity of air is extremely low compared to oil, thereby pushing the majority of the air through.
Desert Rat
Well Known Member
This same question came up about a year ago. At that time, pacific Aero had just told me that no oil would be trapped in their cooler regardless of mounted orientation. I asked their tech engineering department about it specifically because I had seen the showplanes 13 row oil cooler mount that orients the cool upside down with no apparent ill effects.
That entire discussion is in this thread.
vansairforce.net
That entire discussion is in this thread.
Trapped air in oil cooler
I was wondering*with the oil cooler that is mounted with its port in horizontal*direction and on the sides, would it be possible to get air trapped in it and not fill completely*due to trapped air.**
vansairforce.net
Last edited:
Desert Rat
Well Known Member
wait...is this Groundhog Day? I just noticed that two of you guys who are in this conversation were also in that one.
Viscosity has nothing to do with determining flow path. Delta P rules and gases are compressible. If there's no associated device pressure drop budget, u is really only used to verify turbulent flow in the HEx tubes via the Reynolds number. The header is quite the opposite story (reference the attached link).
In the end, trust the OEM. Their design could compensate by not including header area above the outlet (depending on orientation or course), surface area margin (with appropriate increased turbulation) , etc. My experience with most of the aviation oil cooler OEM include non-written comments like, " we prefer a vertical orientation", "bottom->up flow is desired", etc. No OEM wants to have a written installation restriction that the others don't. I'll stick by my comments here and in the attached thread.
Groundhog day on a Friday. Is that a good thing or not?
PilotjohnS
Well Known Member
At 90 psi, i would imagine the trapped air is pretty small volume( if I did the math right, 1/8 the volume at rest). It seems the slug of air would get pushed out, or dissolved into the oil. My experience working on missile hydraulic systems is that the hydraulic fluid, and oil, can entrap lot of air. JMHO
No Sir. The headers are designed for very low velocity/to be manifolds. There are no turbulators applied in the fin pack passes; thus, no way to adjust back pressure through losses. The header design is really the only way to achieve such. Air can get trapped there and can accumulate in these high points. If on the inlet side, you’ll get dead tubes and a proportionally lower heat xfer
The cooler designer may have added margin in these areas where area above the inlet/may not have been considered in the design calcs. This adds weight and cost.
Ultimately, if the cooler OEM has installation requirements or even recommendations, best to adhere to them.
I do not have personal experience based on testing to debate whether air can remain trapped in an oil cooler with both of the fitting ports oriented towards the bottom, but I do know that there are certificated aircraft that have a special fitting with a riser tube installed on the outlet port of the cooler to assure that the cooler has to fully fill with oil before any starts flowing out the exit side.
Manufacturers don’t usually add cost and complexity unless they found a specific reason that it was required.
Fluid dynamics is one of my areas of expertise.
The situation you describe can happen with a non-pressurized system. Where both the air and the oil will exist at the same pressure, and the oil will flow past the trapped air. Because it lacks pressure to expand into all the available space.
Once you pressurize the system, the air gets entrained in the oil as the oil expands to fill the voids. It's the main reason that oil coolers are on the pressure side of the pump and not the suction side.
Not my first heat exchanger rodeo. Believe what you want to believe.
Oil doesn't expand, gases contract.
Entrained gas actually likes to separate from the bulk fluid at low velocity areas like a well designed header. There tends to be plenty of entrained air in our related systems because the reservoir residence time is fairly low. The ullage area available to facilitate the air release isn't great either but I've seen worse.
System NPSHa and pump NPSHr are the most typical driver for HEx placement in a system. Losses can easily make the required pump suction pressure exceed that supplied. This limits pump type to positive displacement types or means added inducers for centrifugal types. There's also the safety concern of imploding a HEx on the suction side of a PD pump in an upset condition. More on that follows.
The HEx overall U value is primarily governed by three parameters. The internal film coefficient, the external film coefficient, and the parent material k value. The HEx walls are made very thin to help the last variable. Thickening them to apply for a suction application reduces the overall q so, surface area must be increased which adds costs and weight. This would require a bypass. Hot oil is better than no oil. Again, more weight and complexity.
Back to the OP's question, it comes down to the OEM's design considerations which we don't know. As people tend to make price point a main driver, I'm sure some cooler manufacturers take advantage of this and cut cost; any resulting application requirements tend to be hidden deeper in the sales lit. Not realized until later many times. Again follow the cooler manufacturer's requirements/recommendations for installation. Doing sh!t again is expensive and annoying.
The OEM's design considerations are maximum heat rejection, minimum size/weight, and cost per unit. I'm one of the people who works with the OEM application engineers to size and source things like this. There would be zero ullage area in a heat exchanger like this, it's not a reservoir like an oil tank. Ullage would defeat the transfer of heat from oil to air.
please read again. Ullage (and residence time) was referring to the design limitations of the oil sump -> why there is air entrained in the oil -> why it can accumulate in high points/low velocity points in the system -> why cooler orientation can be restricted based upon the OEM’s design considerstions. You’re previous stated reasons for HEx placement in a fluid process are far off as already noted. You’ve truly brought zero benefit to this discussion.
Last edited:
It seems common sense to me that air will get trapped. I'm just an industrial construction worker, but I've spent enough time around process systems, heat exchangers, and engineers to recognize Freemasm is correct. The academic arguments to the contrary are just that, academic.
The famous chevy LS v8's have a spot in the heads with a vent/ crossover line to prevent air from being trapped there. If not vented properly when filling the coolant system, that air will allow the head to overheat and warp. BMW's convoluted cooling systems are notorious for air pockets and problems. Just a couple examples of air being trapped in pressurized fluid systems. I've dealt with these problems in both cases.
The famous chevy LS v8's have a spot in the heads with a vent/ crossover line to prevent air from being trapped there. If not vented properly when filling the coolant system, that air will allow the head to overheat and warp. BMW's convoluted cooling systems are notorious for air pockets and problems. Just a couple examples of air being trapped in pressurized fluid systems. I've dealt with these problems in both cases.
I will tell the fine professors at ERAU who use their Cray supercomputer to analyze my data that they are idiots, as determined by some posters on the Van's forum.
They can contact the Van's Forum to get the latest info on computational fluid dynamics research, and stop spending so much time with their expensive supercomputer and the associated professors who don't know anything. I guess the manifolds we made from plexiglass for visualization of the fluid flow through the system are also junk.
Last edited:
those are low pressure systems, operating at 15 psi or less. Not up to 110 psi, as in the case of Lycoming oil systems.
Like I mentioned in my previous post, this subject is way outside of my knowledge base, and in those cases I always defer to others that work within those disciplines, but one thing I have learned is that even if you have the most advanced computing capabilities and designers, after you have a result from all of the high tech. design and analysis you do actual physical testing to prove that the result is valid.
Has that been done?
If not, it would be pretty easy to do if someone cared to know.
