Dissolved O2 removal from boiler feed water 9

[gugubarra]

Our company has signed a pledge to reduce energy consumption. I am looking at using waste steam to pre-heat the feed water to the deaerator. I undertand that there is a general belief of a maximum temperature limit- about 20 oC temperature difference between the incoming feed water and the outgoing deaerated boiler feed water has to be maintained for efficient O2 stripping with steam. I wonder why this has to be so. If the feed water is preheated to a higher temperature than normal boiling point of 100 oC, it should flash (to the extent dpendent on the temperature difference) as it comes off the sprays. This steam flashing should be as effective as using raw steam to strip the O2. It should eliminate the use of raw steam for stripping. Anyone has experience in this please comment.
Thanks.

 

[pmover]

the removal of o2 should occur in the dearator and not in the piping system upstream of the dearator or by flashing the condensate in the dearator. the action taken place in the dearator, that is lp steam mixing with condensate stream, is effective is removing o2 from the condensate/boiler feedwater. i recollect (not too certain) that it is optimum to remove o2 at the operating temperatures and pressures of the dearator and not some other value.
i do not believe that what is described, as i understand, will leave boiler feedwater in the bottom of the dearator - that is preheating condensate and then flashing in dearator. what will cool the steam to produce boiler feedwater in the lower portion of the dearator?
-pmover

 

[EMKWR]

Gugubarra,

you have to get the energy from somewhere in the boiler to preheat the feed water to above 100 °C. This energy is released while flashing. I think it won't be that efficient. Edwin Muller
KEMA Power Generation & Sustainables
Arnhem, The Netherlands
E-mail e.f.muller@kema.nl
Internet:

http://www.kema-water.nl

 

[feather]

Your method wil reduce the number of stages in the deaerator section to 1, which will significantly reduce the O2 removal. Draw a Thiele diagram to see the difference it makes. There are spray/tray designs, but it sounds as if you are trying to reduce the G in the L/G ratio beneath the trays, and that won't work.

You can raise the pressure in the DA if you like to use that spare energy.

 

[jproj]

gugubarra:

First of all, flashing liquid through spray valves is a good way to ruin them. If you break the spray valves, you will have near zero performance in the unit, no matter how much steam flow you have. FYI, 90% of deaerators are operated at 5 psig. Saturation temperature at 5 psig is 228ºF (108.9ºC). If you are bringing the water in at 212ºF (100ºC), it's not going to flash anyway.

Secondly, pre-heating the water releases dissolved gases. I'm guessing your pipe upstream of the deaerator is not stainless steel. If it's not, the liberated gases are going to cause corrosion and eventual pipe failure (dangerous and expensive). There is a reason plants remove dissolved gases IN the deaerator .... this is one of them.

Thirdly, why are you using the waste steam in a pre-heater? A deaerator is nothing more than a direct contact heat exchanger. If you want to use the waste steam to its full potential, use it IN the deaerator.

To answer your original question, a 20ºF temperature difference between the inlet and outlet of the deaerator is desirable because it allows for enough steam flow through the unit. Steam flow through the spray and tray section disrupts the liquid film as it falls. This breaks the falling film into droplets (more surface area). The smaller these droplets are, the easier the dissolved gases are to remove (less distance for diffusion).

jproj

 

[TBP]

jproj - Many plants have tanks where condensate and make-up water is collected/mixed prior to being pumped to the DA. Unless a stainless steel tank is used for this collection/mixing, the carbon steel pits to death in as little as 18 months. The rationale for this arrangement is that it reduces the peak steam loads on the DA. It does, but the price to be paid is a corroded tank & transfer pumps/piping and a huge amount of rust pushed through everything, and on into the boiler.

I have NEVER liked this arrangement, and pointed out to those that would listen, the separate connections for make-up water, and condensate on the DA. I always thought it unlikely that a DA manufacturer would go to the expense of welding on another connection, just for the sheer fun of it. There's also the issue of taking make-up water at city/plant pressure, dumping it into a vented tank, then pumping that same water into the DA. Why deliberately lose the initial pressure, only to have to kick it up again?

I'd appreciate your thoughts on mixing condensate & M/U water in a common tank, then pumping it to the DA. Is there something I don't understand?

Thanks

 

[jproj]

TBP:

I 100% agree with you! There is absolutely no advantage (except for padding the supplier's wallet) to mixing condensate and make-up in a tank prior to the DA. Like you said... you loose water pressure, which requires you to use a pump to get the water to the DA.... more money down the drain. It is much easier to just mix the streams (in the pipe) prior to the DA inlet. Then there is still only one nozzle and you eliminate the need for a stainless steel tank and the associated pumps. Control valves can be installed to control the amount of water the DA sees and the ratio of condensate to make-up. From what I've seen, condensate is allowed to return at will and the make-up is controlled by the level in the storage tank.

The only reason I could think of for having a condensate return tank would be if you had a spray type DA where you needed to run it at near constant load / inlet temp.... but I'm not too fond of spray type DA's either!

jproj

 

[KeItSiSt]

A condensate tank before a deaerator is commonly called a surge tank. There are only two reasons that I can think of to need one: 1) If you have gravity returns coming back that cannot overcome the deaerator pressure then it would need collected and 2) If you have several condensate stations throughout the plant that are independent of each other and could pump back at the same time causing more flow into the deaerator than what it is designed for. A surge tank does not have to be stainless steel. Many times they are lined with an epoxy coating that is much less than purchasing a stainless steel tank. The tank should be protected in some way.

If you are needing a surge tank for reasons above then the make-up water should also be added in this tank with the return condensate. This helps even the temperature before entering the deaertor and letting the deaerator operate at a more consistent load. This is more critical for a spray/scrubber deaerator but also doesn't hurt a spray/tray deaerator either. Hope this helps.

 

[TBP]

Thanks jproj - you've confirmed what I thought. Many DAs have seperate connections for M/U water (usually through a level control valve) and another, separate one for condensate. The condensate just dumps in unregulated as the various pumps in the plant send it back. I have seen several with separate connections again for high and low pressure condensate. Many of these units appear to have been piped-up randomly. (Whatever you do, DON'T read the installation instructions!) Amazingly, they still more or less worked...most of the time...pretty much.

KeItSiSt - My experience with coatings on tanks, piping, etc that are subject to thermal expansion/contraction is not good. The coatings/linings I've seen eventually crack, and allow water between the coating and the carbon steel. The subsequent corrosion between the layers just goes nuts. Perhaps it's because the ones I've seen used most often are as a band-aid, after initial damage has already occured. My experience with stainless steel is that it's not that much more expensive than carbon steel, after labour costs are subtracted. And if SS is the correct material to use, there's no savings to be had by trying to make carbon steel work.

I've always run the DA normal operating level with a bit of space between it, and the overflow connection to allow for condensate pumps sending returns back in a surge. Even if the DA level goes a little high for a few minutes, and bit of water goes down the overflow, it's not normally enough to worry about.

 

[KeItSiSt]

TBP - I agree on operating the DA giving a cushion. I was referencing before the water reached the storage area. If I have a DA designed for 20,000 #/hr (for example) but my condensate is collected at different stations and all comes back at once my DA connections and internals are not sized to handle that much water to be deaerated at once. Most condensate receiver systems are designed to pump the water back quickly. Too many of them at once will flood the deaeration portion of the DA. As far as lining we have not had much trouble using linings in atmospheric surge tanks and is quite a bit less expensive than stainless for us to supply and manufacture.

 

[TBP]

With the condensate coming back relatively hot, I always figured that deaeration requirements would be quite minimal for condensate. Also, if the DA level is high and rising due to several condensate pumps operating at once, the M/U water level control valve should be closed anyway.

On the other hand, I've seen plants with the HP condensate coming back to it's own connection on the DA have problems, even though things were connected correctly. They had some HP traps failed open, and this was keeping the DA pressurized. The PRV that was supposed to be supplying steam wouldn't open, as the DA pressure was made. They were busy pitting their storage section until they fixed the traps.

I come at this as someone who has worked in operations and maintenance for a number of years as opposed to a DA designer. I have no doubt that some plants benefit from using a surge tank, however, I have had far more problems on systems with surge tanks, than on systems without them.

One MAJOR problem I've noticed in several plants lately is that the safety valve protection on DAs is often GROSSLY UNDERSIZED. (A safety valve that would pass 1,200#/hr was protecting a DA, and a customer supplied steam PRV that was capable of feeding 8,500 #/hr to it.) People (at least one of whom was the PE who "designed" this installation) have looked at the dinky little safety valve that comes mounted on the unit from the factory, and have assumed that this is all that's required. I think that DA manufacturers would do better to not put any safety valves at all on their equipment, but leave a larger connection available. The one DA that I refer to above had the factory mounted safety installed on the only availble point on the vessel, the safety inlet was the same size as the connection to the vessel. That connection wasn't even CLOSE to being large enough. We installed an additional safety valve in the piping downstream of the steam PRV, and dropped the size of the PRV from 2" to 1".

 

[KeItSiSt]

I agree with you on the relief valve and PRV comments. Anytime a PRV is supplied the RV HAS to be able to relieve the full capacity of the PRV if stuck wide open (per ASME and common sense!) THe PRV and relief valve should be supplied by the same company. THere is no way to supply a RV without knowing something about the PRV. A common practice now is to have the relief valve mounted between the PRV and deaerator instead of directly on the deaerator tank which is actually preferred by most code people mounting the safety device "as close as possible" to the PRV. This also allows making the connection the size required for that particular installation instead of relying on the manufacture to "guess" what maximum size would be needed. The RV inlet also changes drastically with set pressure which is not always what the design pressure of the unit. Another big mistake on this type of installation is the bypass valve being sized incorrectly around the PRV.

 

[jproj]

It is common to supply a safety relief valve that is not designed to pass full steam flow. The thought is that there should be a safety valve on the steam pipe down stream of the PRV. Many times, however, specifications require the safety valve to be sized to pass full steam flow at the design pressure.

Regarding a DA not being sized to handle a FULL condensate return: This is the SOLE fault of the system designer. Full flow cases must be considered and designed for (forwarded to the deaerator designer). Most DA's can mechanically handle increased condensate returns (to a point). The unit may not be able to fully deaerate under these conditions (if it was not designed to), but return condensate will likely have low non-condesables anyway.

jproj

 

[TBP]

The "system designer" function has pretty much ceased to exist in many if not most project and plants. The individual equipment manufacturers get looked to supply their bit to match the rest of the equipment. This almost never works very well, as the DA guy is not a feedpump guy, who is not a boiler guy, who is not a control valve guy, and so on. Commissioning is another function that has largely disappeared. It's no suprise that many plants don't work very well, and even more amazing that some of them work at all.

I'm not a big fan of bypasses on things like steam traps, and control valves. They help sometimes, but overall, bypasses have caused me far more problems than they've ever solved. The way most plants have downsized their operating crews, there's nobody left to operate a bypass valve anyway.

 

[powerpetunia]

This is an excellent discussion, and reminds me of a problem I ran into a while back. I was looking into the design of a deaerator for a cogeneration plant. The steam to the host was to be used in heat exchangers and would always be on the high pressure side. The condensate would be returned at a fairly low temerature 60°C but high pressure ~1000 kPaa. There would also be a fairly high makeup ~20%. One proposal was made to only deaerate the makeup and to "dump" the condensate directly into the "storage" portion of the deaerator.

My experience has primairly been with condensate from surface condensers so I have no idea if the above idea has any merit or not. I would welcome any comments ...

 

[gugubarra]

To all who responded to my query on preheating deaerator(boiler) feed water with waste steam, thanks. They are useful, especially the discussion on potential corrsion to the deaerator vessel and pipe work.
I still think preheating the feed water above its boiling point(say 110 oC) and then let it flash would remove the dissolved O2 and CO2. We have a rather old deaerator made by Permutit It is basically a sparging/stripping column inside a tank(the deaerator). The feed water enters the sparging column at the bottom. It meets the steam coming off the sparger. The sparger is a perforated pipe of 3/4" holes. I am a bit taken by the simple design but it works. As such, the hot feed water entering and flashing in the current stripping column should in principle work as well. The only difference is the stripping steam is now 'self' generated from the hot feed water. To prevent premature flashing, the feed water control valve can be located as close as possible to the deaerator. If necessary, an orifice of suitable size can be installed in the line to keep the water pressure above its flashing point. And the line spec after should be stainless steel. This measures should prevent corrosion in the water line.
I like more comments on this proposal. All opinions are welcome.
Thank you.

Gugubarra

P/s The waste steam is contaminated with process solids. It can't be re-used directly. Currently, it is condensed and sent to the cooling tower.

 

[25362]

I've heard that a one-stage efficient packed column vacuum deaerating towers would economically reduce O2 to levels of about 100 ppb. Three of these stages would bring down the content to 10 ppb. Vacuum increases as the water moves down the tower!

Subsequent injection of chemical scavengers would still bring the O2 concentration down.

Specialists say that deaerators and feed water heaters are subject to corrosion in various forms: SCC, exfoliation, erosion/corrosion, pitting, snake skin and general corrosion. AISI type 446 ferritic StSt is an alloy practically immune to all of these corrosion mechanisms.

Same sources say that even 7 ppb oxygen would cause corrosion fatigue in carbon steel deaerators, starting with pits as stress risers that ultimately promote the initiation of typical transgranular cracks. This process is enhanced by long residence times ie, contact time cycles between metal and corroder.

 

[jproj]

25362:

The majority of deaerators in service are spray-tray type units with stainless steel tray enclosures and carbon steel head & shell. Many of these very units were installed in the 60's and are still in service!

By the way, who are your sources? It sounds like a sales pitch from a deaerator company who manufactures 446 SS packed tower deaerators.

jproj

 

[25362]

jproj: I've seen an old table published by the Eletric Power Research Institute on Corrosion in Fossil Fuel Power Plants showing various alloys and their resistance to corrosion mechanisms.

 

[KeItSiSt]

I know of carbon steel deaerators that have been in service for 50 years. My experience is that any steel containing carbon is susceptible to oxygen attack. However, at 7 ppb makes the "attackers" very minimal and basically has no effect on the life span of a deaerator if the deaerator has been working properly. Basically, any oxygen is bad in a steam system. In gugubarra's design I would watch that the released oxygen cannot attack the vessel walls. Some deaerators release the gasses but then do not dispose of them properly before they have a chance to attack the vessel walls. Also, any type of continuous flashing near steel can wear on the steel. Dissolved oxygen can still remain in water at 212F at about the 44 ppb range, which usually atmospheric deaerators are rated for.