6 replies to this topic

[lp2]

I have a question on transferring liquid NH3. I work at a plant where the NH3 storage tank is partially refrigerated - pressure about 55 - 57 psig at the top and temperature about 40'F. The pressure is controlled with a compressor drawing vapors off the top, returning condensed liquid to the storage tank. The pressure and temperature of the railcar deliveries vary with ambient temperature - pressure is typically between 100-150 psig. A side stream off the storage compressor discharge (120 psig - 170 psig) is used to pressure the transfer of NH3 into the storage tank.

I am planning to install some new unloading piping and will have a longer piping run into the storage tank. The goal is to achieve a certain unloading rate. I'm trying to calculate the unloading rate for the proposed new piping system.

My questions concern the flow and pressure drop calculations. How do you calculate what the unloading rate will be with the new piping system, considering the high pressure drop from the railcar to the storage tank? How do you prevent excessive flashing with this high pressure drop (around 170-60=110 psig in summer). There is currently not a backpressure control valve at the inlet to the storage tank to maintain the NH3 as a liquid during the transfer.

Without a backpressure control valve at the storage tank inlet, do you have to assume two phase flow? The two phase flow calculations I have seen assume that you know the percentage flashed. I have done an adiabatic flash calculation using upstream and downstream vapor and liquid enthalpies, going from 170 psig to 60 psig, which gives me a weight percent flashed of 10.7%. I also have a spreadsheet using the Friedel Correlation for two phase flow to calculate pressure drop. What is the right approach for this type of flow problem?

Thanks for your help. I've attached a sketch of the system.

Attached Files

•  NH3_Railcar_Unloading_Sketch.xls   23.5KB   154 downloads

[Art Montemayor]

lp:

Please study the attached modified copy of your Excel Workbook that I have taken the liberty to add on to in order to further the work effort that addresses your query.

I have prepared the additional work to show how to arrive at the resolution of defining all the line sizes in your project.

However, I have fallen short of the calculations I had planned to do last night. An unfortunate death in my family makes it imperative that I take the next few days to dedicate to family matters and I regret I cannot finish the workbook at this moment. I will try to get on it possibly this weekend.

In the meantime, I believe you can understand where I intend to go with this product.
 NH3_Railcar_Unloading.zip   104.54KB   147 downloads

[lp2]

Art,

My sympathies to you and your family.

No rush with your response. I'll review what you have attached and look forward to hearing from you next week.

Thanks.
LP

[djack77494]

LP,
I have never done a transfer like this, but have designed anhydrous ammonia railcar unloading stations. Always went into a higher pressure ambient temperature ammonia "bullet". In that case the ammonia arrives as a saturated liquid, and not a two phase mixture.

Seems to me that you could achieve something similar if you could vent vapors from your railcar into your condenser. Let the condensed vapors drain into your storage tank, and draw down the pressure of the railcar to equal the storage tank pressure. That should also drop the railcar temperature to equal the storage tank temperature.

If that is feasible, then at this point you could resume with a "standard design" tank to tank transfer system using a fairly small differential pressure induced by the compressor to transfer liquid between the railcar and tank.

[Art Montemayor]

Lp:

Attached is my Rev1 of the workbook we are using.

Please follow the calculations for the various streams. Note that what I have assumed is that you know the capacity of your ammonia compressor(s). The size I picked based on my experience, seems to be about what you would be using if railcar unloading were to be the maximum capacity that would be fixed in the scope of work. Normally I would divide the total maximum work among at least two compressors. If possible, I would use one compressor only when I had to unload railcars. There are various ways to configure this type of unloading and I'll leave you to tell us what you have. I would expect that you have stand-by, spare compressor capacity available in the event your main compressor goes down or needs maintenance.

To determine the pipe sizing, more basic data is required.

There are some pressure relief scenarios that need to be addressed and other safety factors that are inherent with this type of ammonia unloading. I won't go into those details since you have not raised the issue. I have to presume you know what you are doing and that you are aware of the advantages as well as disadvantages of this type of system - as opposed, for example, to a liquid pump used to transfer the unloaded product.

I hope this helps you define what you have and what you can do.
 NH3_Railcar_Unloading_Rev1.xls   1.16MB   177 downloads

[lp2]

Art,

Thank you for your detailed analysis. I certainly appreciate the time you spent explaining your calculations. After reviewing it, I still have some questions. I also have some clarifications on my system. I hope you have the patience to bear with me on this issue.

My compressor capacity is given in tons of refrigeration. I can multiply by 12,000 Btu/hr/ton and divide by the latent heat of ammonia at condensing conditions and get a capacity in lbs/hr, then convert that to ACFM at suction (storage) conditions. But for the compressor in your example case, your 214 TR does not give me 454 ACFM. Am I missing something here?

I think you have a typo on your line 87 of the Calculation Pad. You calculate 581 lbs/hr vapor from Expansion Valve #1, not #2.

My system has no compressor discharge regulator. I have a rotary screw compressor (and backup) with a slide valve for capacity control.

For your example, I understand the adiabatic flash across the expansion valves and the calculation to determine the quantity of vapor produced. I also understand that the compressor discharge stream is equal to the ammonia vapor created by the heat leak into the tank + the expansion vapor from the two liquid inlet streams. But I’m having difficulty understanding why I have to subtract the vapor created from Expansion Valve #1 and vapor created from the heat leak into the tank system from the total compressor capacity to get the compressor capacity available for unloading.

In your calculation, you take the entire compressor capacity and flow it through the condenser back into the tank to calculate the vapor created by the flash across Expansion Valve #1. If all of the compressor discharge is being recycled back to the tank, what vapor stream is left to unload a railcar? Are you just doing a worst case vapor generation here?

My system has no specific expansion valves. In place of your Expansion Valve #1, I have a level control valve on the outlet of an Ammonia Receiver, which receives liquid ammonia from the condenser on the compressor discharge. When this level control valve opens, the liquid flashes across it, so it acts as an expansion valve. But this level control valve closes when the level in the Ammonia Receiver drops to the set point, so I don’t agree that I have to subtract the vapor created by flashing across this valve from the compressor capacity available for unloading. When there is no unloading going on, heat leak into the tank system creates vapor, which is sent to the compressor suction, all of which is sent to the condenser and receiver, and flashed back into the tank through the level control valve. But when a railcar is unloading into the tank, there may be no ammonia vapor flow through the condenser with liquid flowing back into the tank. Why can’t all of the vapor flow go to the railcar during unloading? I have no back pressure regulator forcing flow through the condenser during unloading, and my level control valve does not stay open.


I also have a question on the heat leak. I can see that the heat leak vapor + the vapor generated from other liquid streams entering the tank cannot exceed the compressor capacity. For my system, in the winter when there is very little heat leak into the system, the compressor can be shut down for extended periods of time with no pressure increase in the tank (no vapor is being generated). Then, to unload a railcar, the compressor must be started, but it is in an unloaded condition, the discharge pressure is much lower than in the summer, and it takes longer to unload a railcar. In fact, we plan to vaporize a small stream of liquid ammonia to feed to the compressor suction for winter unloading – we will be doing the equivalent of creating a heat leak to help with unloading. The vapor from the new heat leak plus the vapor from the unloading liquid flash into the storage tank will be compressed to unload the railcar. In the winter, the railcar may be at only 100 psig, and the vapor generated from Expansion Valve #2 is much less (about 0.051 lbs/lb feed), so we are not running near the compressor capacity. How do you handle your unloading rate calculation in winter conditions?

I have no Expansion Valve #2. In my unloading line, I have manual valves, check valves, and emergency slam valves, but I have no metering or control valve that is equivalent to your Expansion Valve #2. I think I can still do the adiabatic flash calculation for vapor created during unloading (going from railcar conditions to storage conditions), but I can’t define exactly where in the unloading line the flash is occurring. Where the flash is occurring and how to avoid two phase flow in the unloading line was what I was trying to ask in my original post. I'm concerned that if the flow from the railcar flashes in the unloading line, versus right at the tank, it will slow down the unloading rate. Do I need a new expansion valve located at or near the storage tank inlet to maintain liquid phase conditions up to that point and maximize the unloading rate?

Thanks again for all of your help.
LP

[Art Montemayor]

LP:

I’ve now had the opportunity to answer all, if not most, of your questions and comments in your last post. I hope this clears the points I have tried to make. The following are my response to your questions and comments:

1. “for the compressor in your example case, your 214 TR does not give me 454 ACFM. Am I missing something here”
I used printed data from a refrigeration compressor catalog that I have and apparently that information is wrong – as you can note in the calculations I have made to check this information out. At the time I used the information, I was in a hurry and didn’t have time to check out the data. I assumed it was correct as published. Your method of calculating the required compressor capacity is wrong. The correct way to convert Tons of Refrigeration is to first establish what is the reference evaporator temperature (or pressure) that the “TR” is based on. There are several temperatures. I used 5 ^oF – as noted in my calculations. You can use any basis, as long as it is the one that the compressor rating was based on. Then you multiply the TR by 12,000 Btu/hr/TR and you divide by the latent heat of VAPORIZATION (NOT condensation, as you say) at the evaporator’s referred condition. The whole idea of Tons of Refrigeration is to get a measure of the refrigerating effect – which always takes place at the evaporator, not the refrigerant’s condenser.

2. “you have a typo on your line 87 of the Calculation Pad”
Once again, you caught me. However, this was not a typo. As I stated, I was in a big hurry and didn’t check my work in order to get it out, so I used a “copy and paste” maneuver from the previous calculation. This is actually a “pasto” error instead of a “typo”.

3. “My system has no compressor discharge regulator”
If you don’t, you have a situation where you have to manually adjust the pressure bled off and into the railcar. I would not do it this way because I don’t know your installation and I don’t trust manual regulation of pressure into a variety of railcars. This regulated gas is totally dead-ended at the rail car and is needed only at the same rate that you are extracting the liquid ammonia out of the railcar. As you can see by my calculations, this amounts to only about 1.5% of the liquid extraction rate. You are not using that much gaseous ammonia to drive the liquid out of the railcar. And I would use only enough pressure to drive the liquid out at the desired rate. I would not take a chance on putting the total compressor discharge pressure on the railcar because the eventual result, once all liquid is transferred, is that that same high pressure ammonia will be injected into the storage, raising the storage pressure. Not having any real complete basic data with regards to your installation, I consider this a potential hazard and that’s why I go with a regulated flow.

4. “I’m having difficulty understanding why I have to subtract the vapor created from Expansion Valve #1 and vapor created from the heat leak into the tank system from the total compressor capacity to get the compressor capacity available for unloading”.
The compressor takes all the excess gas in the storage vessel – and this excess gas is either generated by liquid expansion (from 2 sources) or from the heat leak coming in from the ambient. This is the compressor’s main function and scope of work. The fact that it produces a high pressure product gas is incidental and used as a means to supply a small quantity of high pressure gas at the railcars. I am not calculating the compressor capacity available for unloading. What I calculated was the amount of compressor capacity required in accordance with the above stated equation. I do not mess with the small amount of ammonia gas that is going into the railcar while it is transferring liquid into storage.

5. “If all of the compressor discharge is being recycled back to the tank, what vapor stream is left to unload a railcar?”
Again, please note my calculations. You can see that if you transfer 150 gpm of liquid ammonia from the railcar into the storage, the amount of regulated high pressure ammonia gas required is small. In my opinion, it is not worth the detail of calculating the over-all effect for now. I’m not trying to develop a final, issue-for-construction design. I don’t have enough basic data to do that.

6. “I don’t agree that I have to subtract the vapor created by flashing across this valve from the compressor capacity available for unloading.”
OK, but I still don’t understand your point. I think the problem I have is that you believe that most of the compressor’s discharge should be for use as high pressure gas going to the railcar for unloading. I don’t believe that is the case – at least not using the 150 gpm assumed unloading flow rate that I assumed. You have not furnished basic data stating the unloading rate, so I had to assume a figure and the calculations show that the gas required to transfer the liquid is only 1.5% of the liquid transferred.

7. “Why can’t all of the vapor flow go to the railcar during unloading?
Because it isn’t needed. What IS needed is that the majority of all compressed ammonia be eventually condensed at the prevailing HP condenser conditions and subsequently expanded to the prevailing storage conditions.

8. “How do you handle your unloading rate calculation in winter conditions?”
I don’t know. You haven’t defined what the winter conditions are. But that should be no problem. Even if you are located above the Artic Circle, you can still easily vaporize approximately 700 lb/hr of stored liquid ammonia to generate the amount required to transfer liquid from the railcar(s). You can use an electric vaporizer. This is what I’ve done in the past when my storage pressures were abnormally low.

9. “I can’t define exactly where in the unloading line the flash is occurring. Where the flash is occurring and how to avoid two phase flow in the unloading line was what I was trying to ask in my original post”
Again, I believe there is a misconception of what constitutes adiabatic expansion and why an expansion valve is employed. Although you haven’t shown or stated it in your sketch, I believe you are not using an expansion device such as a valve. I think this is a gross mistake on the part of the basic design if you want to exercise absolute control over your operations. An expansion valve is used to eliminate the exact problem you are stating: 2-phase flow in a pipe. An expansion valve is always located directly (or as close to) at the target vessel – in this case, the ammonia liquid storage tank. The vapor section of the tank is used (actually designed) as the disengagement volume required to separate the products of the controlled expansion from the valve. I don’t know if you intended to show the railcar liquid entering the bottom of the storage vessel in the bottom section on purpose. I presumed it was done as a convenience for the sake of the sketch. If, indeed, you are introducing a 2-phase flow into the liquid portion of your storage vessel, I can readily state that this is a flawed design. By the use of the expansion valve I automatically cure (or eliminate) any 2-phase flow in the transfer lines because I control all the adiabatic flash directly at the valve and physically locate the valve at the storage tank. Expanded, cold liquid falls into the bottom of the tank and cold, expanded vapor rises to the top of the tank towards the suction of the recompressor. That’s the basic and simplest way to design any refrigeration system and avoid 2-phase flow and liquid entrainment into the compressor. I strongly recommend you employ controlled and instrumented expansion valves in your system.

10. “Do I need a new expansion valve located at or near the storage tank inlet to maintain liquid phase conditions up to that point and maximize the unloading rate?”
I think I answered this in Point #9, above.

It is a good thing that we have covered the basic process design criteria and introduced scope or work themes into this discussion because what I have been anticipating as an outgrowth of this thread will, in my opinion, start to come out - the various instrumentation, safety, and process control issues for the system. There is another version of liquid ammonia transfer from a railcar – but that version is based on the main ammonia storage being at the same, basic pressure and ambient temperature as the railcar. You application is one where the main storage is kept at a lower, colder temperature than the railcar.

I hope that my response has been a positive one that it helps to furnish some answers to your specific needs. I look forward to any further comments or questions on your part.
 NH3_Railcar_Unloading_Rev1.zip   374.23KB   121 downloads