Hydrogen Production

Attached is a PFD. I would appreciate help with it. 

I have a heat exchanger that is a WHB with a steam drum (process stream to pre-reformer/reformer and export steam). Should that be the first heat exchanger after the reformer?

I have a final heat exchanger that will act as a condenser.

If I have a single HTS, I will need another heat exchanger to cool down the outlet prior to entering the condenser. How best to integrate it?

I am basically not sure how to integrate it all in the best way. 

Also, is the process otherwise acceptable? Is it capable of start-up as well as operation once started?

Edited by CEEXPD, 09 November 2018 - 11:33 AM.

Your PFD is very rudimentary.

Moreover it seems to have two shift reactors although you state to have a single HTS..

A lot has been written about heat integration in the other topic. I suggest you read that first.

I have a heat exchanger that is a WHB with a steam drum (process stream to pre-reformer/reformer and export steam). Should that be the first heat exchanger after the reformer?

Yes, that has to do with avoiding metal dusting.

Edited by PingPong, 09 November 2018 - 05:34 AM.

In what ways is it rudimentary? It is a conceptual design, so simplicity is to be expected. Or is it a problem more than mere simplicity? I would appreciate your thoughts on the PFD.

We are unsure if one or two shift reactors will be included, which is why it has two at the moment. 

I am going to change the ZnO beds to add another one in a lead-lag formation. 

Are you able to please check the attached? I am unsure about the situation for the mass balance regarding the HDS and ZnO. I have a lower mass flow out of the ZnO than into it. 

Edit: I know 100% removal isn't the reality, but at this concept stage it has been deemed acceptable. We will adjust accordingly during the detailed design phase.

Edit 2: Also for the pre-reformer, there is the breakdown of hydrocarbons, as well as a methanation and WGS reaction. What sort of conversion is expected? I can't find literature for the conversion, and we have been told to use literature-based conversion for the conceptual design phase (i.e. 75-80% for reformer at 800 Celsius and 20 bar, 90% for shift, etc.).

Edited by CEEXPD, 09 November 2018 - 12:51 PM.

Most published articles on the subject contain a simplified flow scheme that is more detailed than your PFD, for example the one above.

Your PFD is not really a PFD. There is no heat integration shown, no indication of how the stripper/deaerator is operated (open steam, or reboiler, or...), not even a hydrogen recycle compressor.

The ZnO beds are always operated in a lead lag arrangement, with a piping and valving arrangement that allows each bed to be the lead or the lag..

When the first bed breaks through it is bypassed and its spent ZnO replaced, and then placed behind the other bed.

As H₂S is converted into H₂O there will be a small decrease in gas mass flow as O has lower mass than S.

The outlet composition of the Prereformer depends on feed composition, temperature, pressure, and amount of steam. I don't think you will find usable conversions on the internet. You can calculate it using calculation methods from that other topic.

A reformer outlet temperature of 800 ^oC is extremely low. In that other topic the students used 850 ^oC which is the lower limit for design. I suggest you use at least 880 ^oC.

Edited by PingPong, 12 November 2018 - 03:44 AM.

I will try and update the PFD this weekend and offer an improved setup. It is mostly that I was unsure about how to integrate the heat best and so left it partially completed. Although I did include a hydrogen recycle compressor. Perhaps the symbol isn't what you would expect for a compressor, but is what I have been told to use.

I have suggested a lead-lag configuration, but someone is suggesting that we don't need two because the sulfur content won't be sufficient to need switching out prior to the shutdown we will have every two years (shutdown based on reformer tubes).

Will it need changing more often that every two years? 

They have calculated H₂S as 0.850 kmol/hr with a total 578.48 kmol/hr flowing out of the HDS and into this single desulfurisation unit. 

Although we don't have an amount for catalyst, from your experience and knowledge, does it seem unreasonable/unrealistic to only have one. 

I will calculate it as per the other thread and also adjust the reformer outlet temperature to 880 degrees Celsius.

Edit 1: With regards to the NG fuel (as opposed to feed), should it be run through the treatment? We currently don't as it doesn't contact with catalyst or end up downstream at all, but will therefore end up in the atmosphere. 

Edited by CEEXPD, 14 November 2018 - 02:13 PM.

Only a foolish client would accept only one ZnO bed.

As I said before: the amount of sulfur in your natural gas is excessive.

If you want to remove that with only one ZnO bed with a runlength of 2 years you would need an enormous ZnO vessel.

They have calculated H₂S as 0.850 kmol/hr

And how much H₂S would that be in 2 years?

And how much ZnO would that require?

http://www.psbindust... Data Sheet.pdf

NG fuel for burners is not passed through HDS&ZnO.

Edited by PingPong, 15 November 2018 - 03:29 PM.

As usual, Ping-Pong is absolutely correct in his comments and recommendations.

I would add that only a naive and incompetent engineer would suggest or recommend one sole ZnO vessel for service in a steam reformer Hydrogen plant.  How did you arrive at the decision to use only one vessel?  Common sense certainly was not employed when considering the need to dump all the spent ZnO without shutting down all hydrogen plant facilities and those downstream.  A serious and thoughtful practical engineering meditation over a nice cup of coffee would surely convince you that the lead-lag design for two ZnO vessels is the logical and practical way to design the process for continuous and dependable operation.  Attached, please find a quick sketch of what I believe is the core of Ping-Pong’s comment on the subject.  I have placed the piping and valving as I recall designing it 43 years ago.  I would be grateful if you could look at it and give me the assurance that you understand how you would operate such a unit in order to keep the hydrogen plant operating continuously in spite of having to dump and replace the ZnO charge in either of the two vessels.  Oh, and also look into how you could use one vessel to lead and the other to follow (lag) in order to maximize the H₂S loading on any one of the two beds.

I hope these comments help you understand what an experienced and proven professional like Ping-Pong is communicating to you regarding engineering design and making sound and practical engineering decisions.  The more basic data, drawings, and calculations you can furnish our Forum members like Ping-Pong, the more rewards you will collect in the form of sound, specific, and logical advice and answers.

Good luck.

 Lead-Lag Vessel Configuration.xlsx   18.84KB   35 downloads

Only a foolish client would accept only one ZnO bed.

 

As I said before: the amount of sulfur in your natural gas is excessive.

If you want to remove that with only one ZnO bed with a runlength of 2 years you would need an enormous ZnO vessel.

 

They have calculated H₂S as 0.850 kmol/hr

And how much H₂S would that be in 2 years?

 

And how much ZnO would that require?

http://www.psbindust... Data Sheet.pdf

 

 

 

NG fuel for burners is not passed through HDS&ZnO.

I have convinced them to include a lead-lag configuration. 

We are not passing the NG fuel through the burners, but I am not sure the reason for this not being done. Is it because the amount of sulfur is so low into the atmosphere (i.e. below legal limits) that the cost of running the NG fuel through the beds to remove the sulfur wouldn't be worth it?

Edit: Apologies for this question not flowing, but I am planning a schedule for the front-end engineering design phase, and I am unsure whether the HAZID should be before or after the General Arrangement drawings. Once I have completed the Equipment Specification, I can begin drawing the General Arrangement. But I am not sure whether I can simply HAZOP after the P&ID and prior to the GA, or whether it has to come afterwards. And if so, why?

Edited by CEEXPD, 17 November 2018 - 08:54 AM.

I have convinced them to include a lead-lag configuration.

Is it because the amount of sulfur is so low into the atmosphere (i.e. below legal limits)

We are producing three processes for comparison and evaluation (SMR, ATR, POX) at this concept stage, and the person doing POX was refusing to include it as they said it wasn't necessary. Sorry for not being clearer. 

I am trying to understand the reason for not running it through the treatment units, and so figured it was essentially a legal limit issue (i.e. if it isn't above then don't as no cost-benefit regardless of conditions to atmosphere).

I don't want to keep asking without more detailed information furnished, which I will only have during the next (front-end engineering design) stage, as I appreciate what Art said. 

However, I appreciated the mention of metal dusting for steam drums, as it is something I had never heard about. It is something I can now include in the HAZID I am performing on the process. My understanding so far is that a mitigation for this is to line the vessel. Is that correct?

We are producing three processes for comparison and evaluation (SMR, ATR, POX) at this concept stage, and the person doing POX was refusing to include it as they said it wasn't necessary.

In an SMR sulfur must be removed to avoid poisoning the reforming catalyst.

In a POX unit there is no catalyst so one could leave it in and remove the produced H₂S and COS downstream by amine absorption, or by absorption in a physical solvent, or in a mixture of both. A POX unit would not have HDS&ZnO.

However, I appreciated the mention of metal dusting for steam drums, as it is something I had never heard about. It is something I can now include in the HAZID I am performing on the process. My understanding so far is that a mitigation for this is to line the vessel. Is that correct?

Read my earlier remarks again. Metal dusting is never an issue for steam drums but can be for heat exchange tubes.

It would be an issue if the reformer effluent at 880 ^oC would be cooled to HTS inlet temperature of say 350 ^oC by heat exchange with for example the NG or NG/steam mixture. But that is not done exactly to avoid metal dusting.

By cooling the hot reformer effluent by steam generation at 250 ^oC the exchanger tube metal temperature is below the limit for metal dusting. It may look inefficient to use an 880 ^oC stream to produce a stream at only 250 ^oC but the reason is practical. You do not need to mention this in a HAZID.

Edited by PingPong, 18 November 2018 - 09:07 AM.

I misunderstood. Thank you for clarifying.

With regards to POX, I will try and get the PFD for it so far to see if it is acceptable or not, as it currently has a HDS in combination with a ZnO.

I am unsure how to approach the HAZID then. Performed one in the only previous design undertaken, and that was slammed for the inclusion of PPE, maintenance and training. It is meant to include only "design engineering solutions". 

At the moment the improvement has been to not mention the above, but move to design margins for equipment and also pressure relief valves if that is the potential process deviation. 

Attached is the materials section which I can do well enough as it's merely SDS. I can then reasonably well identify the hazards in the hazard inventory. But when it comes to identifying the hazards, all I am saying is process deviations that could result in containment failure leading to potential jet fires (in the case of the release of hydrocarbons). Then saying the mitigation in place is design margins and relief valves.

So in terms of the "design engineering solutions" I need, I am stumped. How do I approach this? I am trying, but don't want them thinking I haven't heeded feedback. 

I have never seen a POX with HDS+ZnO.

Usually there is a Sour Shift downstream the POX reactor followed by Acid Gas Removal (AGR) like in this very simplified flow scheme:

Ideas for HAZID can be found by using Google, for example by searching for:

process HAZID

SMR HAZID

steam reformer HAZID

and similar search terms (use your imagination).

Attached is what they have produced so far. It is also pretty much what we are having to stick with for the PFD for POX due to time constraints. 

As for the HAZID, we have the attached. One is a list of possible mechanisms to consider, but the other is what I have come up with. It seems repetitive and that is why I am trying to think of things that would require at least some other form of engineering input (e.g. lining).

Edit:

For the demineralisation water storage tank, we are providing for at least 24 hours operation, as we know we don't want to run out of boiler feedwater. As we understand, if there is no water in the waste heat boiler we can't cool the process fluid and the waste heat boiler will overheat and result in serious damage.

Does that not mean the above can't be included in the HAZID? We have already provided the design such that it is unlikely. If we can add it in, how would we go about doing that?

Similarly for pumps, we know they could fail, but have provided a standby pump with the same specification in parallel operation in case this occurs. So it seems as though we can't include it, or if we can we don't know how to go about doing it given the above is in place.

Edited by CEEXPD, 18 November 2018 - 08:05 PM.

My personal advice, based on actual field operating experience:

Don't ever, ever operate adsorption beds in adsorption service with the feed, impure gas flowing in a downward direction as your flow sheet shows it.  You are asking for a lot of potential trouble.  Study adsorption beds and their design.

For the costing, I am using cost curves from here (http://dl4a.org/uplo...4th Edition.pdf), which start at page 254.

I need to know a way of sizing volume (m³) based on the information available at this stage. Volumetric flowrate is known so I think it can be done as apparently others are doing it and it is the recommended approach for us, but I don't know how to get a m³ for the vessel from that. 

It is also pretty much what we are having to stick with for the PFD for POX due to time constraints.

The PFD of your POX unit does not look like that of a real POX unit.

Or is this supposed to be a combination of SMR and POX? That would not make any sense.

In any case nobody uses a POX unit to produce pure hydrogen from natural gas feed as an SMR is much more economical for that.

Only if pure hydrogen is to be produced from coal or heavy oil a POX is used because steam reforming is not an option for such feeds.

A POX unit with natural gas feed only makes some sense if syngas (H₂/CO = 2) is to be produced. The Shell Pearl complex in Qatar uses 18 POX units to turn natural gas into syngas for Fischer-Tropsch. The FT product is sent to a Hydrocracker which gets its hydrogen from SMR, not from POX. So although the complex has many POX units Shell also built dedicated SMR to produce pure hydrogen.

The same applies to ATR; that only makes sense for producing syngas, not to produce pure hydrogen.

In the HAZID you should also include the high amount of H₂S in the NG feed and HDS effluent as H₂S is very toxic. Part of the SMR unit may fall into a special classification like other parts of the refinery where H₂S concentrations in the process are above a certain level.

CO in the reformer effluent is also toxic so you want to mention that as well.

I need to know a way of sizing volume (m³) based on the information available at this stage.

You should not start a new topic about the same problem, so I will ignore that other topic.

The size of a vessel depends on its service: adsorber, reactor, separator, storage vessel, ........whatever.

So to cost each vessel you first need to size it according to its service.

Edited by PingPong, 21 November 2018 - 06:05 AM.

I don't have much information about natural gas to produce hydrogen and how to make hydrogen gas from the rest of the natural gas, I also need to treat your post I would love to be in the queue so that I also know how to make hydrogen gas from natural gas.