Start Up of Propane Vaporizer

[2ply]

I was wondering in anyone has experience with the start up procedure of propane vaporizers. I have an issue with the heat load at start up. The vaporizer is for 25000 kg/hr super-heated propane @ 25 bar. LP steam is used as heat transfer input with condensate side control.
Start up load seems to be very load for this...Peak load is 4.5MW and start up has a very small load plus gas sent to flare.... How is it usually done?

 

[georgeverghese]

Check the thermal design of the vaporiser - the minimum steam side tube skin temp should be well above 0degC at max heat load, so water doesnt freeze up and stop heat transfer.

 

[2ply]

It's not freezing that I am afraid of. I am wondering if propane pressure will rise up too fast since there is no temperature control. There is only condensate side control (with level control pot) and I am not sure how the vaporizer will react on start up with no load. The propane must be 25 bar @ 100 deg Celsius.

 

[ApStef]

Do you not have pressure control downstream of the vaporizer?

 

[2ply]

The propane line is delivering gas to a series of turbines. There is no control valve as such, pressure is controlled from the heat transfer rate via altering the surface area of the heat exchanger. Gas is flared until superheat and pressure match turbine conditions.
What I am unsure about is how fast the vaporizer will built up pressure with the condensate side control system during startup.

 

[Compositepro]

The pressure in the vaporizer should be controlled by the supply pressure. If the pressure builds in the vaporizer it will simply push liquid back to the supply tank and there will not be any liquid in the vaporizer.

 

[2ply]

That is the usual case with small vaporizers, in this case we have a check valve on the liquid inlet side (and of course a check valve on the pump discharge) and a check valve on the gas outlet. My worry is how fast this thing builds up pressure on start up and how long it takes to adapt to a decreasing load.

 

[Compositepro]

So where is your pressure safety valve? Your description sounds like an accident waiting to happen. Perhaps you are missing something?

 

[2ply]

There is a fire rated pressure safety valve opening at 35 bar going to safe place plus ESD in case of high pressure (cuts steam, closes condensate pot control valve and activates purge to flare).

 

[georgeverghese]

From the discussion so far, sounds like pressure control is managed at the supply pump, and this sounds sensible.

Shouldnt there be a temp controller on the vaporiser exit that controls steam condensate level - how can you ensure the exit propane is at 100degC otherwise?

 

[2ply]

Of course there is one... We have a FF control on the condensate level that manages pressure and temperature. The issue I'm worried about is the start up with the condensate side control. How is this usually done. It seems to me like you start up with maximum heat input since you have no level in the pot. Also I'm a bit worried of the response when the load is reduced. We managed to shift the level point based on load(vertical S&T have a linear relation) but still not sure if the pressure will be suddenly rising....

 

[georgeverghese]

I see the problem now - at turndown, there is a risk of pressure building up in the feed to the turbines, if the pump dicharge check valve is closed. Would guess the only way out of this is to enable an overpressure line to flare from downstream of the vaporiser. Dont know how you can make a PC and TC control loop work by operating level control only at the vaporiser.

 

[2ply]

I guess it is the same principle as a steam boiler boiler but instead of burner you have steam ass heat input. The heat supply causes evaporation and as a consequence a rise in pressure. The level control is used to change the exposed area to steam by altering condensate level.

 

[georgeverghese]

Had a think about this overnight, and I can see how it is possible to manage pressure and temp on vaporiser exit all on a single HX:-

a) High set PC (set at say 27barg) on vaporiser exit to reset propane level LIC on the shellside. This LIC in turn resets the pump discharge PC set at approx 25barg.
b) Steam condensate level in condensate pot (and hence HX tubeside) is reset by vaporiser exit TC set at 100degC.

Since this is not a U tube HX, thermal stresses at startup need to kept at a minimum. Am guessing this propane is stored in a pressurised tank, and so propane temp is at ambient (15-25degC)?

An FF scheme may be possible for either one of the primary loops (a) or (b), but not both, if there is a flow feedback signal from vaporiser exit.

Does this match up with what you've got now ?

 

[2ply]

Yes you are actually right FF is on propane flow TC/PC has a matching condition (not FF as mentioned below with a LL override on the condensate although IMO FF could apply to both assuming there is no load variation). The part you actually got right is the tube side needs to start up slowly and I was wondering if it should start with the pot full of condensate... Also knowing that the dynamics for this arrangement are fast when you have increasing load (lower the level of condensate to increase steam space in the HX) but not so fast when there is a decrease in load...

 

[georgeverghese]

On loop (a), think it may be better if we had the PC on vaporiser exit at normal set of 25barg, and the pump discharge high set PC at 27barg.

If the loops are configured properly and the steam condensate letdown valve(s) and the FT have adequate turndown, then there shouldnt be a problem with low load operation.

Agreed, the FF scheme would be better set up for the PC resetting propane level loop, since the turbine fuel gas supply pressure usually has a high pressure trip but not a high or low temp trip. Moreover, the bulk of the heating load would be for vaporising and less for superheating.

 

[georgeverghese]

Again on loop (a), you may get more turndown if you used both condensate level and tubeside steam pressure to vary heat input.

 

[2ply]

I thought the same thing but it seems that we can get a lower turndown by lowering the level of propane (BEM HX) so it still has some control flexibility... still do not know how fast will be the response on a sudden decrease in load.

 

[georgeverghese]

But wouldnt the propane level LC setpoint on the shellside be directly reset by the vaporiser HX exit PC on cascade control ? - then you wouldnt have to reset this LC manually. The LT sensing propane level should be fitted with diaphragm seals on both liquid and vapor side sensing legs leading to the dp cell.

Agreed, for cases when a turbine trips, the controls may not react quickly enough to prevent overpressure on the supply to the turbines, due to the latent heat stored in the tube bundle metal mass - for this there should be a higher set PC to relieve excess pressure off to the flare header.

 

[2ply]

Well the level system is a bit more complicated than that and the loop is FF but yes my worry is as you said "due to the latent heat stored in the tube bundle metal mass"...how do I calculate that?