HP Air Distribution Design 1

[tc7]

We are laying out plans for a high pressure air distribution system in two of our testing facilities. The flow requirements are modest, ~40 SCFM (we will use 2 each 20HP recips); operating pressures will be 2000 to 2500 psi. Advice is asked on a couple of issues:
1. Is ASME B31.1 the correct specification to apply for design/inspection? or ASME B31.3?
2. We will be installing dryers (probably regen types) to provide -20 degree pressure dew point; What method is suggested to monitor the air moisture quality? I have seen indicator light type monitors, but I am interested in a more quantitative indication.
3. With air dryness as indicated above, should the distribution pipe line still provide drip lines? What is a safe way to drain HP drip lines?

Thanks for all advice anyone may offer.
Tom

 

[Montemayor]

Tom:

Since you are in the planning stage, this is a good time to kick around opinions, experiences, and recommendations on what will be your final installation. Allow me to make the following comments on the air dryers:

1. Drying 2,500 psig air is no problem. I’ve done it many times. I recommend you employ an adsorption dryer with Activated Alumina adsorbent. This dryer can effectively and consistently dry your product air down to -100 oF (approx. 1 ppm vol.). Note that with adsorption type dryers you don’t design nor operate with a specific dew point product. You normally produce a product with a dew point that is well below your minimum requirements. This is because it is very awkward and not cost effective to try to control the dryer’s capabilities below what it is capable of. Normally this is an attribute of this type of dryer because a super-dry product just makes downstream operation that much better and smoother. In your case you want to employ the dryers at the highest available pressure (2,500 psig) because this is where you will do the minimum of water removal.
2. As with all dryers, you must positively remove all the condensed water in the HP air feed prior to entering the dryer. This means that you must exercise efficient and timely draining of your feed air line. A conventional ball-float trap or automatic drainer is used to carry out this function together with an efficient vapor-liquid separator. In your case you’re probably talking about a separator that is 8”-10” diameter and 3’ high. Always allow for manual drain valves also – as backup and for testing.
3. Employ an adsorbent filter on the dryer’s product line. Sometimes, due to aging and pressure shocks, the adsorbent can release adsorbent “dust”. You don’t need this in your application, so filter it out. This will normally be a small pile of dust that fits in your palm over the duration of a year or two. You normally should inspect the adsorbent every year or two anyway. The adsorbent is usually replaced with fresh material after a certain empirical period of approximately 3-5 years – depending on the service and frequency of use.
4. Pick a NEMA type cycle for your adsorption dryer that is suited to your operation. This will make the regeneration of the unit a simple and direct operation. For example, if you are operating on 8-hour shifts, then you could design your dryer so that one bed is drying for 8 hours, while the other is regenerating during those same 8-hours. Two beds are normally used alternatively. For your stated capacity, you are probably talking about beds that are 4’-6” diameter x 4 feet tall each. Electrical heating will probably be the regenerative energy source. I would recommend you use a small blower to force atmospheric air heated to 500 oF through the dryer being regenerated. You can also use product dry air to regenerate – if you accept the operating cost of using such superior quality air for this purpose. If you do, you eliminate a blower and the adsorbent bed gets smaller and more efficient.
5. You can operate the dryer manually or have it function automatically – whichever is your need or desire.
6. For practical purposes, water in your 2,500 psig will be non-existent downstream of the dryer. However, as I stated, you will always have a need to drain out the condensed water that collects in any high pressure, saturated air stream feeding an air dryer. The vapor-liquid separator complete with draining facilities that I mention above takes care of that – ahead (upstream) of the dryer.
7. If you can use the super-dry air product of an adsorption dryer, then your dew point monitoring is reduced down to testing from time to time – maybe once a week. This is another terrific advantage of a dryer that produces such a high quality air. Of course, if you don’t believe me or if you want to make double sure that you never exceed the -20 oF dew point, you can always install dew point meters in the product line. If your capital monies are bare, you can use an old-fashioned dew point cup tester with dry ice + acetone as your coolant. The dew point cup method, of course, is a manual operation.
8. Another great attribute of adsorption dryers –especially those with activated alumina – is that they selectively adsorb any invading oil particles that make it that far. In other words, they do not allow any oil introduced upstream (such as by your reciprocating compressors) to continue on to your process downstream and contaminate it. Your product air should be so pure and dry that it literally dehydrates your skin when you blow it on yourself. If you are using it to test or experiment in a laboratory setting, then I strongly advise that this is the preferred way to produce your air requirements.

This is what I can offer you for now. If you have any specific questions or doubts, let me know. I hope this helps you out.



Art Montemayor
Spring, TX

 

[pmover]

to supplement art's posting... - excellent process advise.

suggest reading the introduction section of the asme b31.1, which outlines application of appropriate piping code. most likely, asme b31.3 is applicable to your application.

good luck!
-pmover

 

[tc7]

Hey Art-
Thanks for the expert treatise on drying! As luck would have it, I am looking at and am very attracted to a unit which is using activated alumina.

I'd like to learn more about your suggestion of the "dew point cup method" to test dryness - can you refer me to any material on this?

Thanks again for the great response.

 

[Montemayor]

Tom:

Dew point cup:
An apparatus consisting of a small, externally polished, stainless steel (or Chromed) cup placed in a glass container into which is passed the sample gas. The glass container is vented to the atmosphere and the small, sample gas flow rate is continuously pointed at the polished external cup surface. The temperature of the polished surface is lowered by immersing dry ice (solid carbon dioxide) pieces in an acetone solution contained in the cup while stirring it with a thermometer. The temperature at which a fog first appears on the cup’s exterior surface is the dew point of the sample gas at atmospheric pressure.

The dew cup technique entails manually cooling the polished cup, usually made of chromium-plated copper. The cooling of the cup is done by filling it with acetone (or methanol), and small chips of dry ice are dropped into the liquid until the dew point is reached; the dew point is determined when condensation has formed on the outside of the dew cup – as visually seen by the operator when looking through the glass enclosure. The temperature (dew point) is measured by a thermometer placed in the acetone. It should be noted that the dew cup method is a one-time measurement and is dependent upon the skill and training of the operator.

This method is low-cost, but taxing in inconvenience. As stated, it requires a trained, skilled operator and a timely supply of Dry Ice chips on hand. The Dry Ice sublimes away at -109 oF, so it is very difficult to store and have it handy. You have to arrange to have a simple handful available at the time of the test and this can be an inconvenience – unless you can produce it on the spot. Some operators do this by expanding liquid CO2 into a small metal container and producing Dry Ice snow. This is actually nothing more than a CO2 fire extinguisher operation. But I cannot recommend that a CO2 fire extinguisher be used for this because it would entail taking away a critical fire-fighting piece of equipment and possibly jeopardize your fire-fighting capabilities if the extinguisher were not duly and faithfully refilled immediately. There used to be a design of a dew point cup that used a small CO2 cylinder to expand the liquid directly onto the polished surface to differentially reduce the temperature while looking at it. I don't remember its name.

But you will find Dew Point Cups, chilled mirror types, and many other dew point analyzers listed in Google’s search engine on the Web. There’s a method and a price for every application. I have operated year-round production of high purity Oxygen and Nitrogen from Air Separation plants that employed adsorption dryers with nothing more elaborate than a Dew Point Cup manual analyzer. I never had an incident where water moisture passed through my dryers or undetected by my operators using the manual Dew Point method. And I operated my plants with a minimum dew point on the feed air of -100 oF. I used a combination of activated alumina + Mol Sieves in my adsorption dryers. I preferred this test method because it produced a knowledgeable and skilled operator who knew and appreciated the importance of the dew point reading and understood exactly what it represents in terms of water moisture. However, I also appreciate that with the passage of time great improvements have been made in the instrumentation and its automation while the cost of labor has gone up. The operator of today is further separated from the workings and understandings of how the instrument works and what it does. But that is supposed to be progress. I still claim the ability to duplicate the accurate read-out of a modern dew point tester with the use of only a simple dew point cup.

I hope this helps you out.


Art Montemayor
Spring, TX

 

[jay165]

The choice of 31.1 or 31.3 will depend on the type and location of the facility where the unit will be located. Jurisdictional regulations may require one or the other or something else. I agree that 31.3 is the most appropriate standard to use.

 

[tc7]

Jay and Pmover-
After reading through B31.1 and B31.3, it seems to make little difference for my application which to apply insofar as design stresses, materials or inspections. We want to use copper-nickel pipe and sil-brazed socket fittings (I need nothing over 1 1/2" pipe size) throughout the facilities. It is not clear to me what level of inspection can be perfomed on this type of construction; aside from final visual and pressure test, what else is appropriate ?? To what extent must an inspector observe in-process joint fabrications, if any??

Thanks,
Tom

 

[jay165]

The only time I've used CuNi is in ANSI 150 and 300 # seawater service. For fillet welds we used dye-penetrant in addition to visual and hydrotest. Frankly, I wouldn't use CuNi and brazed socket fittings for your high pressure application. Maybe someone else can suggest some additional NDE techniques.

 

[Montemayor]

Tom:

You don’t state how long your small capacity (nominal ½” pipe) distribution header will be, but I can alert you that you will have support problems with such a high pressure, small diameter line. I can almost guess that all (or most of it) will be located indoors, probably anchored or supported up against available walls for easy identification and availability.

I’ve installed a lot of high pressure (2,000 – 5,000 psig) piping in the past; most, however was for high purity industrial gases. We standardized on using ¾” brass pipe, schedule 80 as I recall. The elbows and Tees were also brass and it was all screwed. However, we always silver soldered the screwed fittings. Our supplier used to be RegO. Our justifications were:

1. We didn’t tolerate any leaks at all. Some installations were in hospitals and other public places. Since the gases were high purity and free of any contaminants – especially solids – it was a naturally obvious conclusion that we would never have a reason to internally inspect or open the piping up. Therefore, 100% screwed and soldered connections were a common sense decision. This also furnished good, solid mechanical integrity.

2. I believe that you will find a problem in obtaining solid, sound mechanical integrity with the small pipe size of ½” – in spite of the high schedule number or wall thickness. It will behave like a string of spaghetti and be very unstable. A credible case of a worker or other persons grabbing the pipe or using it as a support in the future will cause you grave problems. I would guess that you might have to support a ½” schedule 80 pipe every 5 to 6 feet. That’s something I recall from our decision to go to ¾” size.

3. We also liked the brass pipe because it remained clean and didn’t require outside painting or maintenance. We color coded our pipe by using color tape stripes every 3 feet.

4. If you are in a laboratory or pilot plant environment, I suggest you make it very difficult or impossible for people to use the piping for support, hanging other pipes from it, shelving, etc.. It isn’t very difficult to break or mechanically damage a small pipe and fittings while it is under 2,500 psig.

5. I ran about 150 feet of ¾” brass, schedule 80 pipe from our main air separation plant out to the loading docks where we filled SCUBA air tanks in a pigtail manifold for local divers in Jamaica. It was very important for me to ensure that the purity of the air being filled into these breathing apparatus was very high and totally free of any solids or liquids. This kind of business was really only a service to local people because we made little or no money from such a sideline. However, the specter of liability was always on my mind and brass piping ensured the purity and quality of the product air. I employed brass fixtures up 10 feet against building walls to support the piping on its run to the loading docks.

I believe you will find that ½” piping will easily take the 2,500 psig pressure. I maintain that supporting this small pipe will be more of a problem and hazard. If you are interested in the copper-nickel pipe and sil-brazed socket fittings because you want to keep your air clean and not leaking, then the brass pipe might be cheaper. I know the brass pipe works from experience.

I hope this information and experience is of some service.


Art Montemayor
Spring, TX

 

[tc7]

Art-
I must admit that I had not even started to worry about supportting the pipe lines as yet and probably would have settled for a ~10'-12' spacing, but you brought to mind exactly some conditions that I saw when I first came to this facility - long-long lengths of high voltage conduit unsupported (60-75 ft!!!) and horribly contorted from the abuses as you mentioned. I will keep your suggestions in mind.

Thanks again.
Tom

 

[StoneCold]

tc7
You got some great advice from Montemayor as usual. The only thought I would add is you might look at using 3/4" stainless steel tubing with Swagelok fittings. It is much easier to make modifications or additional drops for your facility. The up front cost is pretty high but labor costs are very low. Just a thought.

Regards

StoneCold

 

[tc7]

Stone-
You are now one of several that have suggested this to me. I have been trying to interpret, apply and adhere to B31.1 or B31.3 criteria, although it is presenting some hurdles. The idea of Swaglok fittings seems to me far less suitable to Code criteria. I have since discussed the project with a person who claims to be an "authorized" inspector and neither he nor people in his peer group have seen anything like what I am doing at the pressures I will be working at (2000-2500 psi). Basically I really think I'm on my own. As long as I work with certified brazers and keep my materials records in order, they (the Inspectors) don't seem too interested in what I am doing!!!
Tom

 

[Hookem]

Montemayor: Do you have any method for evaluating competing vendor's claims for filter-separators? Chesney@uei-houston.com

 

[Montemayor]

Hookem:

Tom, I don't have any method whereby I can evaluate vendor's claims on filters or separators on gaseous fluids such as high-purity Oxygen, Nitrogen or even instrument air. What I've done in the past is to specify to competing vendors that they must furnish their warranty of performance complete with a submittal of detailed description(s) of the method and instrumentation used to confirm their compliance with specified quality delivery for approval by purchaser. In other words, I normally put the monkey on their back to prove that they have complied with the required, stated specifications. This probably has a tendency to inflate the quoted price, but it liberates the user from having to furnish and apply techniques and/or expertise that may not be readily available - and cost money to develop and acquire.

 

[jcfoley]

tc7,

I have been out of town quite a while so a lot this thread piled up while I was gone. Montemayor has given you some excellent advise for the process side. I have been working in the compressor industry for a decade or so and have one idea that you could consider. You mentioned using a couple of recips to supply your air. Are you taking this from ambient with the recips? If so, at what point in the process are you drying it? I have designed several high pressure (5,000 psig) air systems in the past and used AA adsorptive dryers between stages in order to dry the air at the most cost effective level. HP dryers will eat up your budget very fast. The ideal method is to pull the air off the second stage (135 psig or so) of the recip, then run it through a coalesing filter to remove oil and water. Then you are able to use a commerical off the shelf dryer like Hankison or Zebec to dry the air. Then, the 3rd+ stages of the recip can boost the dry air to the required pressure. The only issue then is the final clean up for your required purity levels (which you did not specify).

Also, I would only do this project with Swagelok or Parker A-lok fittings and tubing. Unless you have exteme purity requirements (less than 10ppm total impurities) these connections will serve you well. If you insist on brazing this, take a look at Parker or CPV for a seal-lok type fitting.

Chris Foley
Midland, TX

 

[tc7]

OK-
I am surprised that this thread has resurfaced so much recently! perhaps some updates are called for:

I have the system together and operational. Despite my budget constraints, the system I could afford seems to be working well and faithfully. Here is the basic setup:
Connected a I-R 20 HP 4-stage intercooled unit directly to a HP air dryer made by AIR-CEL, cabable of -40 deg F dewpoint (I only needed -10). I selected this unit based on price, promise and several references. The unit is a Regenerative Desiccant Type using Activ Alum (just like Monte advised to me early on. Thanks Art) and I am extremely pleased with the unit. This dryer then is piped into a large HP air flask from the last stage and then from flask out to the network. It took awhile to purge through and get the AirCel unit working but it is running dry now. How dry? I still don't know and probably never will in my career, but the quality is leaps and bound above what we suffered with before.

My piping consisted of 1/2-inch nominal cu-ni pipe, sil-braze fittings and CPV brand face seal valves. I applied the NAVY Sil-Braze criteria for workmanship, qualification and inspection simply because no one in the ASME B31.1 or 31.3 world had any clue as to what I was talking about.

I am now doing the same style of piping on another bigger system, the only difference will be I am going to use two parallel Bauer 20-Hp units with built in desicant dryers.
It's been fun and painful but also very sucessful.

Regards to all who have given me your advise, thoughts and interest.

Tom