4 replies to this topic

[RobAD]

I found this site today, and I came away with the impression that it could be an indispensable resource, so I was quick to apply for a membership.

I'm neck deep in my final design project for my course, and I'm discovering that there are many things a chemical engineer needs to know that we weren't taught in class. One of these things is how to figure out where to put relief valves on a compressor, and what size they should be.

The compressor in question is a four stage centrifugal compressor that takes a 3:1 hydrogen to nitrogen feed (with 1% argon) at 50 bar and 25 oC and compresses it to 190 bars before it enters the ammonia synthesis reactors.

The first three stages compress 2.6 kmol/s of gas to 180 bar, so each stage has a compression ratio of 1.53. In the fourth stage, a flow of 5.4 kmol/s recycled synthesis gas enters (also at 180 bar) for a combined total of 8 kmol/s of gas being compressed to 190 bars.

There is an inter-cooler after each of the first three stages that returns the temperature of the gas from ~81oC back down to 25oC

I hope that is enough background on the compressor to help answer some of my questions.

I've been told I need a PSV after the third stage, and I'm starting to assume I will probably need one after the 4th stage too. However, (this is more for my own general knowledge than about this compressor) how do you tell when and where you need a PSV?

The real heart of this problem is I don't know what the criteria are for designing a PSV. From reading this site this afternoon, I know it has something to do with the maximum allowable working pressure of the compressor, but I don't know how to calculate that number.

And, as if that weren't enough there is one more detail (that I know of...) to consider. The molecular weight of the synthesis gas with a 3:1 ratio is 4.6 kg/kmol, but I've also been told that the ration can fluctuate between 2.5 and 3.5 (which means MWs of 5.1 and 4.2 kg/kmol respectively) To be conservative, I'm assuming we use the density that will give us the bigger valve. Is that the only consideration we'll have to make (in terms of designing a PSV) because of this changing ratio?

Thanks in advance for any help, or pointers in the right direction you can give me.
Rob.

[RobAD]

I've got a solution to the problem, I'm hoping someone could tell me if my reasoning is sound.

Using the Bernoulli equation, assuming no work is done, no friction, and no height change. I realize in reality most of those things will be present, but if this is a worst case, senario, all of these things will be working against a maximum flow.

dP/rho + (V1^2-V2^2)/2 = 0

This is for the case where the discharge of the compressor is blocked in, so V1 = 0.

The most dense the gas will get is 65.43 kg/m^3, or 6.681 kmol/m^3.

We're venting to atmosphereic pressure, so P2=1 atm.

Solve for V2 = 799 m/s

The flowrate in is 2.61 kmol/s, or using the molar density, 0.391 m^3/s

Dividing the two numbers, A=Q/V2 = 489x10^-6 m^2

Which workes out to a diameter of 0.025m, or .984 in.

The closest inner diameter of a scheduel 40 pipe 1", with an ID of 1.05 in.


Does this seem like the right track to be on for calculating this sort of thing?

[Art Montemayor]

Rob:

Your two questions are:

1. Where to put relief valves on a compressor?
2. What size they should be?

The first question is one that must seem arcane and veiled to you. I can’t fault you; the process logic involved in deciding where Pressure Relief Valves are called for within a process is actually learned and mastered out in industry under mentorship and constant learning of standards and codes such as API 520 and API 521 – stuff that you probably haven’t even heard about yet. Don’t worry. You demonstrate a lot of resourcefulness and moxie, so I’m sure you will do just fine when the time comes to confront those challenges. In the meantime, let’s attack your problem at hand.

You place your relief devices on a compressor – in this case, a Centrifugal compressor – where they will be the most effective. Never fail to identify WHAT type of compressor you are dealing with; this is very important as you will find out with time. The type of machine sets the engineering logic into gear. A centrifugal machine has a performance curve that clearly shows that the machine can only reach a maximum, “dead-head”, pressure – and go no higher. Therefore, if your compressor driver is a constant-speed motor (or something similar) than you can actually design the discharge of the compressor and all downstream components to safely withstand the maximum discharge of the machine and be done with it --- no PSV is required! In other words, it is virtually impossible for the compressor to surpass the dead-head pressure, so if the system can withstand that pressure, it is safe. However, this may not be your case. Therefore, you must sit down and analyze the process with the in-depth knowledge you have about the compressor and the process characteristics. Here is where your research and study of the Unit Operation and its major equipment characteristics comes into play. This is why it is so important to carefully study and visualize how the machines and vessels actually work – on the outside AND on the inside. You should have a complete understanding of the centrifugal compressor and how it functions. Knowing this, you contemplate all the possible, credible ways the compressor can undergo an excess discharge pressure. I believe you will come to the conclusion that a high pressure can be generated on the discharge of the compressor if one of two things occur:

1. The compressor is over-sped – either through instrument failure or by accident;
2. The discharge piping of the compressor is subjected to being blocked off by a shut valve downstream – again, either through instrument failure or by accident.

In either of these events, you as the design engineer must see to it that no hazard is created to the personnel operating the process or to the equipment. I don't know who told you that a PSV is needed on the 3rd stage discharge, but I don't see the rationale behind it. You must furnish safe over pressure relief. You do this by installing a Pressure Safety Valve (PSV) at the discharge of the compressor's last stage. This PSV is set to relieve at a pressure that is below the lowest safe-rating pressure in the system. This safe-rating pressure can be the Maximum Allowable Working Pressure (MAWP) of the compressor itself or that of the connecting piping, valves, instruments, gaskets, flanges, etc. The weakest component in the system is the one that sets the MAWP for the system and the PSV set pressure. I don’t think this has been explained to you in this manner at your university and that is why I’m very verbose and taking the pain to explain in detail for your (& other students’) benefit. This logical study and analysis of ALL the possible and credible over-pressure scenarios is the most important and critical step in the process of designing a proper, safe relief capacity. If you don’t do this first important step correctly, you will fail in protecting your operators, your plant, and yourself.

The next step, that of identifying the relief capacity required to maintain the process safe, is much easier to carry out. By logic, you know that the compressor has a certain capacity at the conditions it will find itself under should an over-pressure occur. You know this capacity since you have performance curves available to you. If you don’t have the performance curves (as is your conceptual case), then you have the given (and stated) maximum capacity of the compressor: 2.6 + 5.4 = 8.0 kmol/s at 190 bars(g) of pressure. If your compressor is running at constant speed, then that is your maximum capacity and you can then concentrate on picking your weakest pressure-rated component to set the PSV relief pressure. Your job is now practically done. Note that you justify the maximum relieving capacity because you cannot accumulate any Syn Gas while the compressor is turning and your discharge is 100% shut off (blocked off). I can assure you that this is a very credible scenario and, unless I miss my guess, the one that yields the largest flow requirements for relief – and thus, the one to design for.

The third and final step is to carefully and accurately specify the relief conditions and data to a PSV supplier so that the appropriate PSV can be calculated and supplied to you. That’s the normal way the correct PSV is identified in the real-world. You can also calculate the appropriate orifice diameter for the PSV (which is the means of how the external, physical size of the PSV is determined) using supplier-furnished software that employs the related equations and algorithms. You can also calculate the orifice size using manual methods – either way is acceptable and correct. This step is nothing but grinding out numbers using existing equations, so it’s a no-brainer and is obviously the easiest step of the total operation.

That’s it; you have gone through a typical exercise of how a pressure relieving device is identified in order to avert a potentially hazardous and dangerous situation caused by a mal-function of your process or by an error/mistake in your operations. In any case, you should detail out your rational and logical reasoning for placing PSVs in your process and this should form part of your report to your professors who expect this kind of engineering logic from you and other students.

Needless to say, your second posting was not on the right track, nor was it the correct manner of analyzing the situation as I understand it.

I hope this helps you out in further understanding the engineering steps and criteria that await you in the future. You are on the right track if you interpret what I told you to be nothing more than using your God-given horse sense for 75% of the right answer --- typical of most successful engineering solutions. Calculations and computers typically play a minimal role in time and effort for reaching a successful engineering solution to a given problem. The hardest part is the human "brain cell" effort - and it is the one thing that can't be done by anyone else but human engineers.

P.S. - Our website has suffered at least two hacker invasions that have been warded off. However, some posts were deleted in the process and Chris Haslego, the owner, is working hard to safeguard us from future attacks. You - and others - may have suffered some deletions in your posts.

[RobAD]

Thank you Art, that was actually quite illuminating.

It does leave me with a few more questions.

The first one likely has a really obvious answer. You say:

You can also calculate the orifice size using manual methods – either way is acceptable and correct. This step is nothing but grinding out numbers using existing equations, so it’s a no-brainer and is obviously the easiest step of the total operation.

While this should be a no brainer, do you know a good source for PSV design equations? Weather or not I end up using software to calculate actual orifice dimensions, I think it would be helpful to have a good concept of how the diameters are arrived at.

My second question is about when to install a PSV. You site two situations in which one is necessary:

I believe you will come to the conclusion that a high pressure can be generated on the discharge of the compressor if one of two things occur:

1. The compressor is over-sped – either through instrument failure or by accident;
2. The discharge piping of the compressor is subjected to being blocked off by a shut valve downstream – again, either through instrument failure or by accident.

Now, in the situation I am visualizing, the compressed gas stream leaves the compressor between each stage, and enters an intercooler to be cooled back down to 25 oC. Are there no credible cases where there might be a valve that is accidentally blocking off the discharge of the 2nd or 3rd stage of the compressor, and in such a situation wouldn't you also need a PSV?

I would imagine you would need to be able to isolate these intercoolers somehow when the plant was shut down and repairs were being done, I don't know if those would be the only valves between the stages in a compressor, but they are some that come to my mind.

What happens in the case where the discharge on an intermediate stage of a compressor is blocked?

Thank you for all the help so far.

Rob.

[Art Montemayor]

Rob:

Thanks for the response. To paraphrase an old joke: “I’m glad you asked that!”

After you read the following, you’ll probably regret what you asked --- just kidding.

Your first question: where does one go for a good source of PSV design equations? Why, where else? The Google Search Engine, of course! For starters, go to:

http://www.iceweb.com.au/PressRelief/prv.htm
http://www.mercerval...cts/sizing.html
http://www.spiraxsar.../block.asp?id=9

You can easily spend the rest of your youth reading about Pressure Relief devices on the Internet. I admire your thinking and outlook; it is most certainly helpful (I think it is an OBLIGATION for an engineer) to have a good concept of how the PSV diameters are arrived at. It is by knowing and mastering the basic equations and algorithms that an engineer can determine if the correct PSV has been specified, purchased, and installed. Never, never forget that regardless WHO selected the “correct” orifice size on a PSV on YOUR project, YOU are the responsible person and the engineer of record in the event something goes wrong in the future due to a flawed PSV size. That's why it is vitally important to be able to check the calculations - regardless who did them - and another one of the reasons your future employer is going to pay you the big bucks.

I like your second question also, because it shows you are using that God-given brain to think beyond the words and descriptions that you have been told. You are starting to think as an engineer – always looking for the weaknesses and potential hazards in a machine or in a process or challenging the logic of the design. You are absolutely going in the right direction when you critically analyze your compressor and its workings – looking to find a potential hazard in the way it is operated or installed. It just so happens that I described the main characteristics of a centrifugal compressor and as I inferred, the machine is a dynamic one – unlike a reciprocating type. Therefore, each stage of a multi-stage model (such as you are applying in your project) is totally wide open to the next, subsequent stage – all the way through to the last stage of compression. This is so because you normally only have one feed stream to the unit and one discharge stream. It is rare to have side streams to compressors – although this is not to be expected, some process compressors DO have some side feed streams entering/leaving intermediate stages. These types of machines are special and are typically spotted on P&ID drawings. You haven’t stated you have side streams, so I have to assume you don’t. That’s why I emphasized that each compressor has to be analyzed on its own merits. NEVER, NEVER EVER generalize when you are doing a process PSV scenario analysis. You must specifically study each application and judge it on its own merits and weaknesses. That’s how you arrive at deciding what is the worse-case scenario for which you must design the relief case. To answer your question specifically, there normally is no credible case where there might be a block valve on the discharge of the 2nd or 3rd stage of a centrifugal compressor, and in such a situation you wouldn't normally need a PSV there. HOWEVER, it is your responsibility to analyze the situation in detail (usually through an “as-built” P&ID) and ensure that that is indeed the case before deciding whether a blocked discharge condition is credible and possible. In the case where the discharge on an intermediate stage of a compressor is blocked, you should analyze the scenario and make the obvious conclusion that a possible over-pressure hazard is not only credible – but also possible, and a PSV is required. I expected you to make that horse-sensical good engineering judgment.

You normally do not need to be able to isolate the compressor intercoolers when the plant is shut down and repairs are required on the same. Sit down with a cup of coffee and think about it, Rob! Use that horse sense one more time. If the plant is shut down, so is the compressor! And if the compressor is shut down, you are totally free to work on any component to your heart’s delight. Therefore, why invest in block valves within the stages and cause a potential hazardous condition by allowing for some operator or other human to mistakenly close these block valves while the compressor is operating or prior to starting up the compressor? The answer is obvious: You don’t do that nor do you go there. It makes no common horse sense. This is the type of needed, valuable, and important analysis that I’ve been alluding to.

I hope that I have helped in making normal engineering design logic more logical to you and other students who may be also reading this. I hope you enjoy this type of engineering thinking and judgment because if you succeed in getting your degree (as I suspect you will) you will be doing a lot of it during your future career.