14 replies to this topic

[K J]

we have a p1,000,00 NM3/Hr capacity hydrogen plant. It has topsoe designed side fired reformer. 80 to 85% heat duty of reformer is delivered by PSA off gas coming from final hydrogen purification. now problem is that this reformer is very sensitive with respect to weather condition change. normally at morning 9 to 10am and in evening 5 to 6pm sudden changes are likely to taking place in atmosphear which reduces reforme gas temperature 6 to 7C(normaly remains 831C). this disturbs PSA operation and PSA off gas generation increases. this whole phinomena takes place within 15 min and disturbes entire reformer. plant location is gujarat(India) where normally ambiant temperature remain 30 to 35C.

 

My question is that  is any body faces same kind of problem, if yes what solution they have adopted.

[PingPong]

Trouble shooting is difficult even when standing in the plant. Via short messages on the internet it is almost impossible.

Maybe somebody can help if you provide more information, such as:

 

Do you have a flowscheme, or DCS screen print of the unit (assuming that is not confidential) ?

 

.... p1,000,00 NM3/Hr ....

What exactly do you mean? 1,000,000 or 100,000 or ....?

 

What is the feedstock for the unit?

 

What is the fuel to the heater, apart from the PSA offgas?

 

Does the heater have air preheat?

 

normally at morning 9 to 10am and in evening 5 to 6pm sudden changes are likely to taking place in atmosphear

What kind of sudden changes do you mean? Sudden temperature changes? Sudden atmospheric pressure changes? Sudden rain? Sudden wind? ....... ?

[K J]

100000 normal meter cube per hour that you can see in PSA screen shot

feed stock is NG with approx composition shown in NG distribution shot

Apart from PSA off gas fuel to reformer is NG itself is fuel

Yes, air is preheated to nearly 255C

sudden change means may be humidity level change because same problems were faced when rain is about to start and humidity level increases suddenly.

Attached Files

•  ASGP.xlsx   1.87MB   89 downloads

[clarkr3000]

I assume you are steam reforming NG to make H2 and CO. I am not sure how the ambient temperature would affect the refoming process other than your main air/fuel to your burners having more moisture resulting in causing you to have to fire more to get the same RG temperature. If you are getting more water in your PSA off gas then you have a cooling water temperature problem in your feed gas assuming you are scrubbing it somewhere. The solution to this is obvious.

[thorium90]

It might not be very believable (to engineers and managers! Not so to operators...), but the reason such variations occur is simply due to the angle of the Sun's rays hitting the plant. The off gas drum is a tall piece of equipment made of metal with no insulation and simple paint protection.

 

When the sun rises at about 9am in Gujarat, the sun heats up the drum rapidly, causing an appreciable increase in off gas temperature going to the burners, causing the tube temperature to rise.

You might then wonder why the reason for the drop in temperature is consequently not at sunset at about 8pm. The reason is primarily because of the layout of the plant. Pipe racks and other equipment block the sun's rays and hence at about 5-6pm, the sun's rays are mostly blocked by other equipment. The secondary effect is due to the isenthalpic expansion of the gas from the PSA assisting in the cooling down of the offgas drum, thereby bringing down the temperature faster the moment the drum is out of the sun's view.

 

http://www.suncalc.n...013.12.06/18:56

 

In fact, such interesting conditions are commonly seen in all plants. Rain will cause even larger variations in the plant, including cooling down the steam drums, pipes and ambient air.

 

There isnt really alot that can be done to prevent these disturbances. Operators should be more alert during these times, thats all.

Edited by thorium90, 06 December 2013 - 06:03 AM.

[PingPong]

In any steam reformer a properly designed, and tuned, control system should be able to maintain a constant reformer outlet temperature, no matter what the ambient temperature, or its moisture content is.

 

However when I look at the controls of this unit there seems to be no reformer outlet temperature control. There is TIS2156A/B/C but that seems to be only a 2oo3 voting system for furnace safeguarding (or similar). There is no TIC2156 connected to it to control the fuel firing.

In the fuel control scheme there is no signal coming in from the process to set the fuel fired. The firing only seems to be controlled by PIC2309 and FIC2309. Neither of them gets its setpoint from the reformer outlet temperature. I have the impression that the firing is simply controlled by a manually set duty of 120 Gcal/hr.

 

Maybe I overlook something, but it seems to me that the firing control of this unit is no more than a duty control, in which the reformer outlet temperature does not play any roll. Moreover the heating value of the PSA tailgas is manual input, and consequently does not change when the composition of the PSA offgas changes.

 

I have never seen a steam reformer design in which the firing was not controlled by a TIC at the reformer outlet. That TIC can be the master of a duty controller, or of a PIC or FIC on the fuel gas system.

 

Also I have never seen a steam reformer with an outlet temperature of only 830 ^oC. I have seen designs ranging from 870 - 930 ^oC. The lower the outlet temperature, the higher the methane slip.

 

I wonder, did Haldor Topsoe really design this unit without a TIC to control the reformer outlet temperature?

And did they really recommend to operate this unit at a reformer outlet temperature of only 830 ^oC ?

Or is this low outlet temperature simply the result of a too low manually set duty at FIC2309 ?

Edited by PingPong, 06 December 2013 - 06:29 AM.

[thorium90]

The load controller is not shown in his screenshots but the temperature TI2156 does go to the controller that controls FY2309. The amount of fuel gas flow is calculated by the controller using TI2156, TI2094 and the theoretical duty required. Dont worry PingPong, its not your overlook, the OP just didnt show it. See the green diamond pointing a white dashed arrow towards PIC2309 as well as the grey box FY2309 on its right that is clickable? That could be the controller that if shown to you will clear up your doubts.

 

I believe the reason the plant is operated with such a cool syngas temperature of about 830 could be due to hot tubes. Notice how hot the flue gas temperature is? TI2094 is 1085.9C. Also, note that the methane slip is a mere 6.35%, seems pretty good for just 830C. Also note that his fraction of hydrogen injection is higher than needed for this load, possibly causing the lower outlet temperature of 830C (WGS is endothermic in the direction of CO and H2O). Burner duty of 120Gcal/hr does seem high for this load, although I will need to see more numbers to confirm that. The difference between the two O2 readings AI2094A/B are too far apart to make much judgement. It should be calibrated.

 

In response to the OP's question about battling weather effects on plant operation. Other than having alert operators and better controllers, there can be simple non-engineering solutions to this. Having a shade for the PSA equipment will help. A shelter for all the bottles might not be so feasible but sheltering the valves skid from the sun and rain would be good for the valves in the long run. Having the drum painted white and maintaining the paint job will also help in reflecting radiation and reducing thermal variation

Edited by thorium90, 06 December 2013 - 09:26 AM.

[PingPong]

I don't see a signal from TI2156 to FY2309, and I don't see a signal from TI2094 either.

 

Anyway, my point was: there is no (visible) temperature controller at the outlet of the reformer. There seems to be nothing present to maintain the reformer outlet temperature constant at 830 ^oC.

 

If there were a temperature controller TIC at reformer outlet to maintain that 830 ^oC then it would be impossible that the reformer outlet temperature drops 6 - 7 ^oC simply because the weather changes. The TIC would then immediately increase the firing to maintain that 830 ^oC.

In older designs the TIC would reset the setpoint of a FIC or PIC on the fuel gas. Nowadays it is fashionable the have duty control where possible, so then the TIC should reset the duty setpoint. Note however that duty control only makes sense provided that the duty can accurately be calculated, which is not the case here. As long as its setpoint is reset by a TIC it does not really matter, but unfortunately that is not the case here.

 

 

PS: I did not realise until now that this unit is in the same plant as the N2 compressor in http://www.cheresour...gal-compressor.

Also there the process control system was rather questionable.

Edited by PingPong, 06 December 2013 - 08:57 AM.

[thorium90]

Haha, KJ actually has a pretty large plant with many units. What he has shown us (in different threads) is just a tiny portion of what he has. You are right, the DCS screen doesnt show everything. KJ appears to be learning the plant from the DCS (Not the best way IMO, but thats not up to me). Perhaps the P&ID on this portion will clarify the control mechanism of this part.

 

The controls are not immediate. It would take a while for the drop in the offgas drum temperature to affect the gas temperature, then the furnace firing, then the tube and catalyst temperatures and the subsequent syngas and then for the controller to make adjustments to compensate for that. The off gas drum is pretty large. The average space time for an SMR of about 100000NM3/hr is about 10+mins, consistent with his observation of about 15mins to stabilise the plant.

Edited by thorium90, 06 December 2013 - 09:27 AM.

[K J]

@ ping pong, you are correct, N2 compressor was also part of the same plant. Actualy our old plant was LSHS gasification(Texaco design) based ammonia plant with dirty co shift conversion(COMo catalyst), rectisol CO2 removal and N2 wash unit for final syngas purification. But as part of enrgy reduction new syn gas generation plant was establised. In new plant operation old plant asset was utilised(ex LP N2 from ASU, Syn converter loop etc). our old plant is still in operation at low load ( approx 35% of total NH3 generation) as our CO based products depend on gasifier CO rich gas.

 

@ thorium, for learning purpose P&ID is good but DCS graphics are good to explain real operating condition sweet and short.

 

this reformer unit has not actualy such kind of TIC2156 which controls duty and reformer duty is curently adjusted by manual entry. actualy this unit has been comissioned just before 9 months even GTR is also pending. we dont have experience of side fired reformer operation. we have proposed topsoe to provide such kind of TIC.

Edited by K J, 06 December 2013 - 02:39 PM.

[thorium90]

 

this reformer unit has not actualy such kind of TIC2156 which controls duty and reformer duty is curently adjusted by manual entry. actualy this unit has been comissioned just before 9 months even GTR is also pending. we dont have experience of side fired reformer operation. we have proposed topsoe to provide such kind of TIC.

 

 

Im not so sure about that. I've worked with a number of HT SMR units and I have seen the control designs. Perhaps if you posted the P&ID on that part it will ease everyone's mind.

Edited by thorium90, 06 December 2013 - 06:44 PM.

[clarkr3000]

Again the solution is obvious.

[PingPong]

.... even GTR is also pending. we dont have experience of side fired reformer operation. we have proposed topsoe to provide such kind of TIC.

Does GTR mean Guarantee Test Run ?

 

You should look up all supplied unit documentation. That should contain the guaranteed numbers, the original design heat and material balance, and the reformer outlet temperature that the design heat and material balance and guarantees are based on.

[K J]

After checking number of things reason we found was funny. as you will see in PSA P&ID off gas flow (which I said provides 80% heat duty) is controled by FIC4308. it gets PT compensated signal from FT4308. now here problem starts. range for temperature compensation is defined 28.5 to 51C. during night time in winter ambient temperature falls to 17C which also reduces TT4308 indiaction below 28.5C. during transition from 28.5C FT generates PT compensated and PT uncompensated signal intermitently. during uncompensated signal FT4308 generates raw value signal which has large indiaction differnce than than  earlier. flow indiaction difference thus reported was around 2300NM3/Hr. after 15-20 min TT4308 drops below 28.5C and transition over and final raw signal will results. we have now force TT4308 signal going to PTC as 29C as solution.

Attached Files

•  fuel control.pdf   83.68KB   42 downloads
•  NG distribution.pdf   102.47KB   43 downloads
•  psa.pdf   93.87KB   43 downloads
•  reformer radiant.pdf   84.08KB   52 downloads

[PingPong]

we have now force TT4308 signal going to PTC as 29C as solution.

Nice detective work, but it does not change the fact that the most important process variable, the reformer outlet temperature, is not controlled but is simply the result of what an operator sets in that duty control. The duty setting at 120 Gcal/h, the heating value of the PSA offgas (set at 4400 kcal/Nm³), even the heating value of the natural gas fuel (set at 8700 kcal/Nm³) are all manual inputs.

The reformer control is presently no more than glorified hand control.

 

 

I used the DCS data that you posted in message #3 to estimate the material balance and stream compositions.

These are my observations:

Almost 30 % of the methane from the natural gas feed is not converted and ends up in the PSA tailgas. That's why the PSA offgas is such a huge portion of the total fired duty.

The heating value of the PSA tailgas is more likely to be 4700 kcal/Nm³ than the manually set value of 4400 kcal/Nm³ and as a consequence the fired duty of the PSA offgas is about 110 Gcal/h.

The fired duty of the natural gas fuel is about 20 Gcal/h.

The total fired duty is therefor in reality about 130 Gcal/h instead of the manually set 120 Gcal/h.

Edited by PingPong, 09 December 2013 - 03:10 PM.