56 replies to this topic

[SilverShaded]

also is desulfurized NG used for reformer fuel, or pipeline undesulfurized NG as such?

There will be a guard bed

[daraj]

silver, thanks.

 

what is the current state of the art efficiency of a reformer?

 

I have calculated heat duty of reformer as  = heat provided in reformer+heat provided to preheat natural gas fuel and boiler feed water +heat loss to flue gas. Now what should i divide this value with to get overall duty? (top fired reformer). 0.8 or 0.9? what is the overall efficiency loss expected?(not including heat loss via flue gas as that is already accounted for)

[SilverShaded]

silver, thanks.

 

what is the current state of the art efficiency of a reformer?

 

I have calculated heat duty of reformer as  = heat provided in reformer+heat provided to preheat natural gas fuel and boiler feed water +heat loss to flue gas. Now what should i divide this value with to get overall duty? (top fired reformer). 0.8 or 0.9? what is the overall efficiency loss expected?(not including heat loss via flue gas as that is already accounted for)

I'm not sure what you mean.  Where is this heat being lost to?
 

[daraj]

Hi silver, let us say iam burning 100MW worth of natural gas fuel. Ot of this certain heat is used for SMR rections and certain portion forpreheating fuel, water, etc. However, some portion of this heat is lost along with hot flue gas. some portion of it is also lost as irreversible losses in radiation/convection possibly. so if we assume these losses total 10MW, then only 90MW is useful use of the heat input. so efficiency = 90%. Thats my understanding. So given this, what do we generally take as current day overal efficincies? is it 92 or 93 or 90 or what?

[SilverShaded]

Hi silver, let us say iam burning 100MW worth of natural gas fuel. Ot of this certain heat is used for SMR rections and certain portion forpreheating fuel, water, etc. However, some portion of this heat is lost along with hot flue gas. some portion of it is also lost as irreversible losses in radiation/convection possibly. so if we assume these losses total 10MW, then only 90MW is useful use of the heat input. so efficiency = 90%. Thats my understanding. So given this, what do we generally take as current day overal efficincies? is it 92 or 93 or 90 or what?

In terms of a simulation you don't need to account for radiation losses, real units can be modelled accurately without assuming 'parasitic' losses.

[daraj]

silver, for purposes of energy balance I need to. Cannot ignore any losses. Need to know how muchfuel i need to burn. whne licensors say efficiency is 92%, does that mena remaining 8% is all lost to hot flue gas? I believe this 8% includes irreversible losses in each section. Please clarify

[SilverShaded]

What percentage have you got going up the stack?

[PingPong]

silver, for purposes of energy balance I need to. Cannot ignore any losses. Need to know how much fuel i need to burn. when licensors say efficiency is 92%, does that mean remaining 8% is all lost to hot flue gas? I believe this 8% includes irreversible losses in each section. Please clarify

Efficiency depends on stack temperature, which again depends on whether furnace has air preheat or not.

Normally an SMR furnace has air preheat, unless it is very small.

 

For example: suppose the stack temperature is then 145 ^oC, that is 120 ^oC higher than the 25 ^oC on which the LHV of the fuel is based.

Divide that 120 ^oC by 20 ^oC/% and you get 6 % stack loss.

Add to that 2 % heat loss in the whole furnace and you get a total loss of 8 %, so furnace efficiency is then 92 %.
 

[daraj]

Pingpong, i have seen PFDs in the past where steam generation has been shown to recover heat with SMR flue gas as well as by cooling the reformed gas. It can be modleed as a series of heat exchangers through which boiler feed water flows and gradually turns into saturated and then superheated steam, which is then used to drive compressors. The question I have is, in some of those PFDs I dont see a steam drum whereas in some of them i see a stem drum attached to the syngas cooler and also attached to the heat exchangers within the SMR convection section. for modeling purposes i dont really need a steam drum. I can have steam flowing through each successive heat ex, gradually turning into superheated steam.  But in reality I wanted to know when is a steam drum really needed. do you needed wherever you are generating saturated steam? what purposes does the steam drum serve? (plese note that in my process iam generating superheated steam which is being used to drive compressors)

[PingPong]

The question I have is, in some of those PFDs I dont see a steam drum whereas in some of them i see a stem drum attached to the syngas cooler and also attached to the heat exchangers within the SMR convection section.

You need a steam drum.

 

I don't know the PFD's that you have seen. Were they real PFD's made by professionals, or merely simplified flow schemes in textbooks, articles and websites? Simplified flow schemes always show only part of the required equipment and are not suitable to base a real design upon.

 

If you would make a design in which the BFW is completely vaporized and superheated in a convection coil then the salts that are always present in the BFW will be laid down inside the coil at the location where the last water is completely vaporized, thereby gradually plugging it and moreover the coil metal would become overheated at that location and may fail.

 

A steam drum has a blowdown (continuous as well as intermittent) to remove the salts by purging some drum water.

 

The only way to generate steam without a separate steam drum would be to use a kettle type steam generator. In that case the kettle acts also as a steam drum and also has blowdown.

[daraj]

Thanks. So steam drum will be attached to the exchanger where you are making the saturated steam.

 

for a quick sizing of horizontal steam drum, any guideline is tere? like residence time based on rate of generation of steam? let us say iam making 120 tons/hr of saturated steam at 100 bar. are there set dimensions for steam drums of varying sizes?

 

 

also you mentioned

"For example: suppose the stack temperature is then 145 ^oC, that is 120 ^oC higher than the 25 ^oC on which the LHV of the fuel is based.

Divide that 120 ^oC by 20 ^oC/% and you get 6 % stack loss.

Add to that 2 % heat loss in the whole furnace and you get a total loss of 8 %, so furnace efficiency is then 92 %."

 

didnt follow thiscalculation fully although i think it will be useful for me. why divide the 120C delta by 20C? is it a thumbrule for every 20c hotter flue gas, you lsoe 1 efficiency percentage?

[PingPong]

for a quick sizing of horizontal steam drum, any guideline is tere? like residence time based on rate of generation of steam? let us say iam making 120 tons/hr of saturated steam at 100 bar. are there set dimensions for steam drums of varying sizes?

No.

It depends on how much holdup time you want for the drum water, which again depends on the reliability of the BFW supply. Client may have requirement for hold-up time to be used.

Also depends on mechanical design pressure of the steam drum: the higher that is the higher the optimal L/D ratio of the drum will be for economical reasons.

 

 why divide the 120C delta by 20C? is it a thumbrule for every 20c hotter flue gas, you lose 1 efficiency percentage?

Yes, 20 ^oC/% is a very rough rule of thumb.

Exact number depends on fuel composition and % excess air to burners.

 

[daraj]

Pingpong, for energy/ower balance, often we need to know efficiency values for steam turbine(that drive compressors) aswell as efficiency of compressors themselves(such as syngas compressor). do you know if 70% for steam turbine and 75% for compressor overall is conservative enough? vendors claim higher efficiencies(like 80% for compressors)but i have heard that those numbers are not real. My duty values are within 20MW for these equipment.

[PingPong]

That question had better been asked in your "Steam Turbine Efficiency" topic, so I will post a response there.

 

This topic is to be about "Steam Reformer Simulation" (see title) so best not to contaminate it with other stuff.

[daraj]

OK, pingpong, tlaking of SMR, iam struggling a bit with heat balance. if my understanding it right when i simulate it, i will describe the steps i follow:

 

1. i have a NG and PSA tail gas stream coming in as fuel to a RGIBBS type combustion reactor. i also send in enriched air sligtly preheated at 60C

 

2. outlet of combustion reactor, after complete burning , it will calculate a temperature. Lets call it T1.

 

3. now i cool the above stream in a cooler block  and extract just about enough duty which i need within the reformer tubes. after i extract this duty the hot flue gas is now entering the convection section where it is about to preheat NG, steam, air etc. at this stage the tmeperature is around 1050C. (my reformer exit is around 860C)

 

so far is it correct?

 

my problem is my T1 value righ after combustion is high as preeiducted by aspen around 1950C. is tis realistic? will it be this high in the radiiation section?

[daraj]

my NG is coming in at 25C only, and air at 60C. But after RGIBBS combustion reactor, the outlet stream reaches a temperature value of 1950C. Now if i reduce amount of fuel burnt, to reduce this temperature, then after providing heat to the reforme r tubes, the stream may cool down to 950c or 900 and that is too low. i know bridgewall is around 1000-1050c. and then it goes to convection bank, prheats everything else and finally leaves as stack gas at 110-120C.

 

in my reformer  reactor tubes, the NG+steam mixture enters at 550C and leaves at 850C. the endothermic heat duty is what is provided by the hot flue gas after the cobustion RGIBBS reactor.

[PingPong]

The T1 = 1950 ^oC that RGIBBS calculates is merely the theoretical adiabatic flame temperature, not the temperature in the radiant section.

T1 can very well be around 2000 ^oC in a gas fired furnace. It depends also on the amount of excess air that is used.

 

You moreover say that you use enriched air, so less nitrogen ballast, that will also increase T1.

 

And air preheat increases T1 too.

 

However 60 ^oC is hardly a realistic air preheat temperature to the burners of an SMR furnace. To bring the stack temperature down to 120 ^oC (as you mention) the preheated air temperature is likely to be in the order of 400 ^oC or maybe even more.

 

You must first calculate the required duty to heat the process gas from the XOT (you mention 550 ^oC) to the COT (you mention 860 ^oC) and taking into account a realistic ATE. Say that is Q1.

 

Take into account that there is a heat loss of about 1 to 1.5 % from the radiant cell to the environment, so the duty to be delivered by the flue gas is actually slightly more than Q1, let's say Q2.

 

Then you determine the amount of fuel and associated combustion air (including excess) required to deliver Q2 while being cooled to the bridgewall temperature (you mention 1000 - 1050 ^oC).

Edited by PingPong, 17 December 2021 - 10:11 AM.

[SilverShaded]

Dont forget to subtract the low calorific value gas (from the PSA) from the calculated total fuel duty.

[daraj]

Thanks to both of you.  Yes, PSA tail gas is already part of the fuelstream that goes to my burner

[daraj]

Pingpong, the exchangers within SMR in the convection bank, used for preheating NG, steam etc. cann they be modeled as shell and tube(shortcut method) in aspen? with flue gas on the shell side? in reality these are just tubes over which hot gas flows. for a somewhat realistic simulation of heat transfer, i want to knpow what would be the approahc temperature i can use and what type of exchanger (shell and tube?) should i use in the simulator? in conventional S&T, i use approach temp of 10C.

[PingPong]

The MITA (minimum temperature approach) of all coils in the convection bank will be much higher than 10 ^oC.

 

Designing the convection bank of an SMR furnace is not simply applying some typical MITA to each coil.

It requires experience and insight that I can't teach you in a few forum posts.

 

If you are familiar with Pinch Technology you could apply those principles to position all streams requiring duty (NG preheat, NG+steam preheat, steam generation, steam superheat, air preheat) against the fluegas that is cooled by these.

[daraj]

"If you are familiar with Pinch Technology you could apply those principles to position all streams requiring duty (NG preheat, NG+steam preheat, steam generation, steam superheat, air preheat) against the fluegas that is cooled by these."

 

that is what iam trying to do in a simulator. I am having hot flue gas at 1000+C, progressively gettiing cooled as it delivers heat to each of these recovery exchangers. but given the options available in my simulator(aspen) i am wondering how to calculate a somewhat realistic heat duty transferred in the convection bank. I know reality is complex, but . I am not exactly designing the convection bank here but just trying to do an energy balance to calculate fuel that i need to fire. for each of the stream(NG, ait, steam etc) that is being preheated in the convection bank, can the outlet temperature of those streams be say 15C less than the outlet temperature of the flue gas after it provides the heat? or even higher?

[PingPong]

Fuel fired is determined by heat balance of radiant section, not convection section.

 

Read my post #42 again and again and again until you understand that.

 

For positioning the various cold streams against the hot flue gas use a grid diagram as in Pinch Tech. Can be done simply by hand.

[daraj]

"Fuel fired is determined by heat balance of radiant section, not convection section"

 

Pingpong yes iam following what you suggested in post #42. only question that remains is the air that is used for combustion is actually preheated in convection section making it a bit cyclical.. so when iam calculating how much heat the flue gas can give off before being cooled to bridgewall temperature, the temperature of flue gas predicted by simulator is actually affected by air preheat temperature. should i consider air to be preheated for calculating radiant section duty? if i consider preheated air, then obviously i can recover more heat in radiant section before being cooled to bridgewall temp. of 1050C. But to determine how much air preheat is possible i need to complete the convection section as well. It is all interconnected making it a bit tricky.

Edited by daraj, 22 December 2021 - 01:19 AM.

[daraj]

to elaborate, i found the duty. lets say it is 250MMBTU/hr(including losses) in radiation section. Now the hot flue gas, after providing duty of 250MMBTU/hr, needs to reach a temperature of 1050C.  which means before it provides this duty it may be at a teperature of 1900C or 2000c or lower. this again depends on temperaturre ofair iam using in y BURNER block in simulator. if i dont preheat air, for the same radiation section duty, i end up with a bridgewall of just 975 or so. if i use preheated air, i can end up with 1050. so we may need to assume a certain possible preheat first and then obtain that preheat later on in convection section, it looks like