STEAM JACKETING FOR HEATING A GAS STREAM

[gelsi]

Hi all
I need design information/criteria for operation in object.
I have following data:
3" line jacketed with 4" tube;
350 kg/h of GAS (similar to air)@ 6barg & 25°C , to be heated to 140°C;
satured STEAM 150°C.

I try with double tube unit with shell condensation, but it's very difficult to develop a calculation sheet from zero.

It would be better a xls file, if there is someone in the web. Help!!!!

P.S. I think the better way to heat a stream is an heater equipment (like S&T), while the jacketing make only mantainement of temperature, what do you think?

 

[MJCronin]

It sounds like the decision to use jacketed pipe was made by others.... perhaps your boss.

Getting "locked in" to a certain type of HX usually means that someone who developed the proposal for the client ( or the client themselves) had to look competent at a meeting. Perhaps the jacketed system already exists and it was to be reused.

Jacketed pipe can be used in this application..... but it may be very long and expensive, especially when an accompanying steam and condensate system is included. Don't forget steam traps) A S&THX will probably have a cheaper CAPEX

If a jacketed piping system with steam is considered, do not forget the extra cost because of a vacuum rating necessary for the jacket. This increases the jacket wall thickness and cost.

Steam heated S&THXs have been in use in the natural gas industry for quite some time.

http://www.slideshare.net/pinarpina...wards-sustainability-3857869?next_slideshow=1

MJCronin
Sr. Process Engineer

 

[gelsi]

right guessed, MJCronin.
I want propose an S&THX but nonetheless I must estimate the minimum jacketing length.

 

[MJCronin]

Aha ! I knew it !

Once you have performed a detailed calculation to demonstrate to your MBA boss how his solution will be too expensive. He will then demand that you rush size, design and get prices for a brand spanking new S&THX.

The quotes will come back too as expensive and too much fabrication time...

After that, the MBA will demand you get prices and perform an evaluation for all used S&THXs within a 1000 mile radius.

If you find a used unit that is acceptable, he will go into his boss and tell him about this and all of the money he has saved.

Next, the MBA will bring you into his office to demand an apology and a reason why your task took too long......

MJCronin
Sr. Process Engineer

 

[gelsi]

neglecting h0 (annular side) i find that:
nr 3 jacket pipes 3"/4" of 6 m give about 12000 W, that are needed to rise air temperature from 25 to 145°C with sat steam @ 150°C
Anyone can say if this is an abnormal result?

 

[LittleInch]

12 KW looks good from my perspective and your numbers as a gross heat input into the gas.

Hence 4kW per 6m long pipe - average ~700W/m. A little high, but achievable at 150C.

I think the issue is a more practical one of how you get that amount of steam / condensate along the annulus

You need to allow something for external heat losses, but otherwise the issue could be incomplete complete mixing.

what is the Actual gas velocity in one of the three 4" tubes? I could work it out, but you have more data than me.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[Latexman]

I agree with MJCronin. S&T HX is better than jacketed pipe.

Have you considered a P&F HX?

Fill in a datasheet and ask for quotes for both types. No need to size it. The vendors will. Ask the to send calcs. Check their duties and sizes against each other.

If you HAVE to use jacketed pipe, I had good results from Controls Southeast Inc.

Good luck,
Latexman

Technically, the glass is always full - 1/2 air and 1/2 water.

 

[gelsi]

...
my Ud = 66 W/m2K by calc
i find a reference range 25 - 50 W/m2K
so i would reduce the Ud to 40-45.

LittleInch, velocity of gas is about 7.2 m/s (into 3" pipe)
(For major clarification, 4" is the size of the jacket arround the 3" pipe)

 

[LittleInch]

Hmmmm, to get 4 kW into a gas in less than a second..... Just doesn't sound right to me, but not sure where you would calculate that.



Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[marty007]

Continuing our conversation from the other thread...

Just out of curiosity, I ran your numbers through my HTRI software. I won't give you exact numbers, but I'll just say that 6m seems a little short...

You have a pretty tight temperature pinch at the air-outlet...

Cheers,
Marty

 

[MJCronin]

Hhhhmmmmmm..... anyone vote for 12m of jacketing ? How good is the heat transfer between the jacket and the pipe ?

As I recall, this crucial heat transfer was always done by some mysterious black glop.... only available from the jacket vendor !

12m of jacket will still give less than 2 seconds residence time ?

Does the STHX sound any better ?

MJCronin
Sr. Process Engineer

 

[gelsi]

marty007,
my result is n°3 pipe of 6 m (tot 18 m jacketed), with
- about 66 Ud,
- sat steam @ 150°C,
- neglecting h0 (annular side),
- target: to rise (6 barg compressed) air temperature from 25 to 145°C.

 

[LittleInch]

The key point is how you get to a U value of 66 - please post your calc/assumptions to see where this comes from.

've had a look at double tube heat exchangers and can't really see a typical value hopping out, but for plate type steam/air values as low as 15 appear.

Given your desired outlet temperature is close to the steam temp, a longer tube would seem to be a good suggestion based on something closer to 15 to 25 U value. if it starts coming out too hot, shut one of the tubes off...

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[gelsi]

Assumptions
- chosing the minimum pipe of 6 m long, I run only 3 sections,
- properties are estimated at overall media temperature and not at single pipe media temperature
- Nusselt correlation used is SAIDER TATE even if Pr = 0.68 < 0.7
- ho = infinite
- fouling steam side = 0
- fouling air side = m2K/W 0,000352243

These are just a little more detailed results...

DATI COLD STREAM PIPE 1 PIPE 2 PIPE 3
step 1 FIXED A Tcold out - LMDT CALCULATION
tc OUT °C 108,76 136,61 145,88
LMDT °C 76,1 25,3 8,4
step 2 hio HEAT TRANSFER CALCULATION (SAIDER TATE) & Ud CALCULATION WITH ho = INFINITE
Twall °C 150 x 150 x 150 x
Nu - 200 200 200
hio w / mq K 67,9 67,9 67,9
Udirt w / mq K 66,5 66,5 66,5
step 3 FIXED LENGHT OF SINGLE PIPE IN SERIES - DUTY CALCULATION
L m 6 6 6
q W 8479,0 2820,0 937,9
step 4÷6 Tcold out CALCULATION AND CHECK THE ERROR
Tc out °C 108,7 136,6 145,8
errore -0,0001 0,0000 0,0002
tot DUTY W 12236,9

 

[marty007]

I took a look at your numbers again, and the math appears to be correct. One issue I would consider is your step size; 6m pipe lengths really aren't small enough to correctly capture the very tight temperature pinch on your outlet.

Using your numbers, I made a small excel spreadsheet. I used your heat duty data and matched the model with 6m step increments, and at 18m (3-steps) had a discharge temp of 145C. I then increased the resolution to 0.5m increments (easy to copy/paste excel rows), and it now required about 28m.

Be careful, and consider your step size... and don't take my word for any of this...

 

[MJCronin]

Another factor to consider is the complete reworking/reconfiguring the piping system.

Aside from rack piping, it is not easy to find a 12m straight section in any piping system, much less 28m. I am guessing here, but a reconfigurated system appears to be a necessity. Don't forget the new pipe supports, thermal stress analysis and steam traps.

CAPEX ?

Has the MBA asked yet how much this reconfigured piping system will all cost ? (and why you are not already done ? )

MJCronin
Sr. Process Engineer

 

[LittleInch]

marty,

shows why a diagram is essential... I initially read it as three parallel 6m heaters. I now think you mean three series heaters??, i.e. essentially a heated pipe 18 m long with three heated sections all connected to the same steam line, hence all at 150 C. correct?

For space you could then have two elbows between each so you get flow going in the same direction.

That makes much more sense, but even then a residence time of 2 and a bit seconds doesn't sound enough to me. That the first pipe goes from 25 to 108 means it is taking over 50% of the steam flow / heat input.

If you make an allowance for four pipes then you're probably covered, just don't connect the steam supply to the last one if your equation is correct!

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[marty007]

This would essentially be one big double pipe exchanger. Each pass wouldn't need to be 6m exactly, as long as you have the correct total heated length. It could look something like the pictures on this website: Link (not my photos, just found in google).

The connection of the steam supplies isn't all that critical as long as the velocity/pressure drop in the jacket isn't crazy and the condensate can freely drain. The nice thing about condensation is that it occurs at a constant temperature, so co-current/counter-current are meaningless.

As an aside, I have to think that a standard S/T exchanger is still going to be your better option. These types of double pipe exchanger have their place, and our shop has built quite a few for various clients (up to 20-30 stacked sections for overall effective lengths of 400+ft). The place where they work is in situations where you have a significant temperature cross, require true counter-current flow, but have very low flow-rates.

 

[gelsi]

Hi
I tried to reduce lenght of a section to 1 m but my result is the same
each section is different to each other only for DTML
where am I wrong?
(see attach)

 

[gelsi]

using properties vs media(T) of section:
gas rise to 145°C at 17.7 m without dependance of sections number (tried 3 sections of 6 m and 18 of 1)