Filtration 101

[CraigAClark]

Here is an incident that occured in the last 6 months. We sold a machine to a vessel well over a year ago. It incorporated a small combo tower ( filter/ separator) with filtration @ 7500 cu ft based on an 80F inlet temp., 3.5 scfm block (Bauer). The captain e-mailed notifing us of the incedent and blaming us for it. That is their version of what happened. The hose leading to the manifold (portable) ruptured due to the pressure resulting from the seive (13X) creating an internal fire. Fortunately no one was injured.

After some questions, we found out that the operator was new and the instruction manual provided with the machine was lost. So, you have a dangerous situation to begin with. If you calculate the processing capacity of the filter at the ambient operating conditions, they should have been working on a 75% capacity, we figured 25hours max. The filter also had a litmus indicator in it. Given that the installed life of the filter was only 6 months, the filter was over a year old and was run for 40 hours. Well they changed it out before operating the machine. You can imagine how much water dripped out into the tower when the old, over used filter was removed. The tower should have been disassembled and cleaned before use. As you can imagine, since you understand and know about the " undiscussed" potential of 13X when subjected to a slug of water under pressure, what actually happened. They returned the old filter with the machine and the litmus strip was beige and it was saturated with oil. Someone had closed the ACD overide valve. No one knows how long the machine was operated in this condition, nor how many drain cycles were missed. The replacement (new) filter was in pieces, sieve was discolored and the carbon became the fuel. The filter tower itself was scortched on the inside.

Now, I can imagine the version that "untrained" operator will tell his buds about the incident. That's reality, especially when he seeks a new position. The incedent was obviously our fault.

Craig

 

[iain/hsm]

Gentlemen.

You can blame the Europeans for the low temperature conditions, as we don’t get any sunshine over here, just rain and cold. But may we offer some detail as to water vapour, water condensation, pressure dew point and pressure vessel design.

Regarding the suitability of 7075 alloy IMHO Jordair are simply misleading you with this statement. It is not the 7075 material that is the problem. It is Jordair’s own lack of suitable design and there own miss application of pressure vessel engineering.

Firstly there is (or should have been) a difference in the design at Jordair between the filter shell (the part that carries the filter chemical) and the separator shell (the part that separates the condensed water/oil particulate). It is critical to note that a filter shell is designed for pressure cycles of 500 to say 5000 psig (or whatever MWP) as opposed to a filter separator designed for pressure cycles between (zero) 0 and 5000 psig, and included in that design is the compensation required to ensure full pressure cycles of zero to max working pressure three to four times an hour (during auto drain operation over the life of the product). This is known fact to filter design engineers. That the pressure cycle in separators is greater than that of filter shells and requires further consideration as to:

1. The wall thickness.
2. The architecture of the vessel, vessel design stress calculations.
3. The radius of all the internal thread undercuts.

All these issues are well known to vessel designers and should have been to Jordair’s.
Further in the Jordair statement they state that the separator consists of a shell with a plug and a base. This tells you that they were using a standard filter shell design and not a separator design and IMHO this is there own design problem and not that of the material. Putting out a blanket scare and blaming it on material is frankly wrong. Understandable but wrong and misleading non the less.

 

[pescador775]

OK, Iain, to compensate for the low ductility of 7075, the mfgrs are using thicker metal. I get it, thanks. I forgot about that outfit in England (fish symbol) that has been making filters/coalescers out of aircraft aluminum for so long. Is there any way of finding out what alloy they were using in 1973? Do you know?

 

[iain/hsm]

OK, Iain, to compensate for the low ductility of 7075, the mfgrs are using thicker metal. I get it, thanks. I forgot about that outfit in England (fish symbol) that has been making filters/coalescers out of aircraft aluminum for so long. Is there any way of finding out what alloy they were using in 1973? Do you know?

Yes and No,
Yes, was the marketing manager for said “fish” outfit when they used to manufacture in house. Now left, and now they no longer manufacture. But I have continued to supply both in the design and manufacture with HSM Engineering Technology.

In 73 the material was HE30TF alloy for both filters and separators. Incidentally there was and still is three methods of manufacture, in those days the shell was either gun drilled from solid (look inside the barrel if you can see the drill marks and drill end angle at the end it was gun drilled from solid, Then later as with other filters made for Compair, Bristol Pneumatic, American Bristol and others we made a larger diameter longer length filter that was designed using a Luxfer heavy wall extrusion (diving type extruded shell) again look inside down the barrel if the inside end is a smooth radius then it was manufactured from a diving cylinder shell (before the head closing or neck closure procedure). In those days the neck closure on Luxfer cylinders in UK was known a “Swan Neck”. We were always different to the U.S. both in material and in neck design.

One engineering problem with the extruded shells is they are to all intent and purpose “banana” shaped so when machining the inner thread you end up with a greater degree of slack or tolerance, Fit up an American Bristol shell without the 0-ring and back up ring and look at the degree of “wobble” in the threads. The thing rocks around quite alarmingly. Agreed trepanning the shell removes this problem.

The other problem using extrusions is the softness of the material and the pressure life cycle, Quite the reverse to what you would expect, pressure vessels from extrusion are of a thicker wall than from tube due to material “grain”, tensile hardness and more important elongation and deflection under stress.

The third method is using high tolerance aircraft quality tube, you can tell these by the threaded cap and plug at each end normally 7075 but specially treated.

This allows longer lengths and also a larger diameter, with a higher or greater tolerance, less wobble on the thread. The higher tensile material in tube form is found with the majority of European compressor suppliers. Bauer is an example.

The increase in diameter from 3.750” OD material used for the U.S. market to 104mm material used in Europe allows Europe a bigger diameter filter cartridge typically increasing it to use a 70mm cartridge this allows around 30% more chemical to be used and this in advertising increases filter life expectancy.
Filter spigot dimensions is a minor detail but the gas flow is very important, The best choice is allow the gas up straight into the filter medium as in Bauer, and Lawrence Factor, The other cheaper method is to use the filter shell as an extra cooling condenser or separator and have the gas travel from the top down into the filter chemical, this is as per Bristol, and Compair and you can tell with the oily residue inside the shell. One method is not best suited for Nitrox production.

 

[iain/hsm]

OK, Iain, to compensate for the low ductility of 7075, the mfgrs are using thicker metal.

Almost forgot the first part of the question. No, in fact it is quite the reverse. In the early days the working pressure to burst pressure using HE30 was 2.5:1 that is about a 12,000 psig burst pressure. But they had a heavy wall due to the softness of the material elongation and permanent expansion problems with soft HE30 material.

Our 7075 material filters are designed with a 4:1 safety pressure rating i.e four
(4) times the working pressure, so our minimum burst pressure is over 25,000 psig. This allows a 6000psig working pressure vessel with a greater degree of safety. 4:1 also complies with the U.S ASME specification.

However burst pressure is only half the engineering. Cycle pressure loading is just as important maybe more, We have always supplied separators designed to be loaded from zero to full working pressure 100,000 times and have the test results showing that. However by increasing the minimum pressure from zero to say 500 psig in the shell the pressure cycle is indefinite. It’s just a better design of auto dump that makes all the difference. Instead of blasting away all the condensate and the all the gas each and every time say (3 times per hour) a gentle bleed of just the condensate not the gas would be much better.
Like they say the devil is in the detail.

 

[pescador775]

Instead of blasting away all the condensate and the all the gas each and every time say (3 times per hour) a gentle bleed of just the condensate not the gas would be much better.
Like they say the devil is in the detail.

From the horse's mouth, heh. I never understood the big blow down except as lazy design.

The higher tensile material in tube form is found with the majority of European compressor suppliers. Bauer is an example.

I believe Bauer is using 6061 T6

What is the new designation for HE30-- 6351 or 6082?

 

[iain/hsm]

From the horse's mouth, heh. I never understood the big blow down except as lazy design.
I believe Bauer is using 6061 T6
What is the new designation for HE30-- 6351 or 6082?

Thanks Pescador.
As I'm usually called the horse's ass I think this is a step up!. Lol.

The Bauer Separators manufactured by Bauer Munich with the M76 thread and 95mm OD and 270mm long (wall length) had a rated 350 barg working pressure and 455 barg hydro. The filter shells would also have (from my inspection) been made from hollow tube stock 102mm OD with a 14.5mm wall. The German material designate was ALMG SI1 F31 this is 6082 but precipitation hardened to T6 condition with a tensile strength of 270Mpa. (By comparison 7075 is over 500 Mpa).
The 10” 20” 27” and Securus filter shells from Germany were made also in 6082.

This material coupled with the cyclic loads and the M76 x 2 thread had design problems from the beginning with the form in the 95mm OD separator shell had a stress area around a 30 degree 0-ring chamfer to a 90 degree x 2mm landing to shell, this was a weakness in the design. They were subject to eventual replacement due to a number of explosive failures.

Interesting to note that a diving scuba cylinder gets filled say once a day yet the compressor separator gets filled 3 to 4 times an hour, each hour every hour the compressor is running. In a dive shop running say 1000 hours a year thats 3000 to 4000 fills, or compression cycles, As some filters on compressors get to be say 10 years of age that’s 40,000 fills, Use an old mil spec surplus 20 year old filter and we are up to say 80,000 fills. At 85K fills you have reached the maximum design life of the shell. Add to this corrosion, stress, strain and any time you exceed 50C at some point the shell is about ready to retire. I think the term is “To go out with a bang”!!

 

[pescador775]

Yes to the old Bauer filters. There were a couple of incidents in Australia. The investigative report stated that due to almost constant operation on dive boats which occurred over many years, the 80,000 cycle rating had been exceeded resulting in said explosions. I don't recall that there were any injuries, just broken furniture.

 

[veriqster]

Guys,I think both of you are right. Basic physics: Higher temperature - more water in the air & higher energy of the vapor form water which BOTH contribute to the same result. Now, if you want to bash each other's heads I'll step aside.
Just my 9th grade physics 2 cents, no pun intended.

P.S. Sorry, I see this got constructive at some point, glad to see so many details.
Thanks a LOT.

 

[pescador775]

Iaian. Using a 4 X 1 safety factor with such a rigid metal as aircraft aluminum seems appropriate. Three years ago, an internet friend was selling compressor systems which were based on Walter Kidde pumps, complete assemblies fabricated in his garage. In order to be price competitive on EBay, he sought to build filters from scratch. The design he showed me was of a combined filter/coalescer made from 7075 pipe. Since the market included Scuba and paintball the press capability had to be higher than usual ( for USA). I tried to dissuade him from this but he was enthusiastic for this approach and so we finally settled on using Barlow's formula to calculate a minimum 4 X 1 factor. We looked at the threads and O rings for worst case purposes and overshot a bit on the strength of materials. I remember we calculated a yield press around 20,000 psi and mind you the paintball guys jam tanks up to 300 bar, above 4000 psi. The filters ultimately worked well and there have been no reported problems. I used to worry about the assumptions we used but based on what you said we did not err on the side of caution but hit it about right.

The two filters which I built for my own system use 2024 T6. No problems, knock on wood.