5 replies to this topic

First post ever. Here goes:

My question relates to the necessity of a bypass valve with a reciprocating compressor design. By "bypass valve", I am specifically referring to a valve on a piece of pipe connecting the suction and discharge piping, inside of the block valves.

As an electrical engineer by trade, my old job had me automating large CAT driven recips for natural gas compression on pipelines. These recips had a bypass valve as described above. Typically, we would open bypass, open suction block, start engine, open discharge, then close bypass. This included large gas volumes, high flow rates (50MMSCFD), and large pressure separations from suction to discharge (500psig to 1000psig during operation).

A system we are designing has been proposed with two recips (each with their own suction, discharge, purge valves) and there is a recycle valve that connects the common suction and discharge headers (read: after the unit block valves). The pressure spread is likely 200psig to 400psig (suction to discharge) and the flow rates are much lower (I'm not sure of this value at present). There are two buffer tanks per unit, one inside the suction and one inside of the discharge.

Do I need a bypass valve for initial loading of the compressor? What considerations are there for this valve to be there? Will the buffer tanks suffice during startup? What calcs can I do? How worried should I be about deadheading?

Thanks for everyone's time.
----Brian ----

[Art Montemayor]

Brian:

I think I can help you out with your concerns. In my 48 years of engineering, I've done a lot of projects related to reciprocating compressors – industrial gases and natural gas. I compressed gases to 5,000 psig through 5 stages of compression. I believe I understand your concerns very well because I've been there and done that many times. Allow me to summarize what I believe are answers to your queries, but not before I draw a flow diagram of what I believe is your proposed 2-compressor setup. Please refer to the attached Excel Workbook and look at the flow diagram. Feel free to correct or modify it as you feel is needed in order to confirm that it is a TRUE depiction of what you are proposing. This is important to consider because the compressor configuration, piping, and controls all play a significant role in safely starting, stopping, and operating the system. Based on what I have envisioned as your setup, my comments are:

1 What you have is a recycle capacity control valve – NOT a bypass valve. A bypass valve is something totally different and for a different application. Look at the flow diagram and confirm that the proposed hook up is basically as shown by me. Note that I show a manual bypass valve around the automatic recycle control valve. This is the standard way to pipe up a reciprocating compressor when you need to have capacity control via a recycle valve.

2) A reciprocating compressor is a positive displacement machine. Once it starts moving gas, it will continue to do so – regardless whether you shut the discharge valve on it or not – unless you equip it with a capacity control system. There are several ways to control the capacity of a recip. You are proposing the cheapest and simplest way. It also is the costliest way with respect to energy consumption – but that is your choice. Another way is to furnish clearance pockets on the cylinders. This is the most expensive, but wastes no energy. Valve unloaders and speed control are the other two ways to vary the amount of gas a recip compresses. A reciprocating compressor is like an electric generator: it will keep on generating gas at a pressure unless you put capacity controls on it. All the recycle valve does is that it works to keep the suction pressure in the 1^st stage suction drum at a pre-set value by recycling high pressure gas from the discharge back to the suction drum. This means that you waste the energy you put into the same gas to compress it while expanding it back to low pressure. But it keeps the cylinders filled with gas and that is what the positive displacement machine understands when it is run at constant speed.

3) The reason you use a recycle to start (and stop) the recip is that a positive displacement machine – like an electric motor – can't just start and stop in a split second while it is trying to carry a load. It must be unloaded while starting and stopping. The recycle valve allows the recip to unload itself by simply recycling all the product gas (inefficiently) back to its suction – in other words, it is doing no work, just literally spinning its wheels while it is put on line or taken off line. The manual bypass valve around the automatic recycle valve allows you to do this as a startup or shutdown procedure.

4) You haven't given us much information on the characteristics of your compressors, so I've had to assume some things in order to get started. I assume both compressors will be identical, 2-stage machines (at least). Both machines will be driven by a diesel or gas engine at constant rpm. You will unload the recips prior to starting or stopping the engine drivers. You will be capable of running either one machine at a time or both in parallel operation. You will probably use air-cooled intercoolers between stages of compression and separate out any condensibles – like water or condensate. I am not showing any details on all the necessary or desired instrumentation, relief valves, alarms, purges, blow-downs, and controls. This can be shown later.

5) You need a manual bypass (as I show it) for unloading the machine – NOT for initial loading. Note that all recycle gas is cool gas - NOT hot gas.

6) I have already explained the considerations for the valve to be there.

7) The "buffer tanks" are really either separators or pulsation drums; either way, they are required for different reasons.

8) You basically don't have to do any calculations except that you must know the capcity of the compressors in volumetric flow and be able to specify to a control valve supplier the conditions and specifications of what the valve is expected to do. Fisher Controls is usually the supplier of choice for oil patch equipment such as what you are describing. Contact them and start identifying what you have to specify to them for supply of appropriate control valves.

9) You should be VERY WORRIED about the possibility of deadheading a couple of reciprocating compressors that are designed to continue to compress until they are either relieved of the load, overpressure the system, or they destroy themselves. That is what they are designed to do and that is what they will do because they are good at it. The normal thing is for them to find relief from any system restriction or obstacle that causes an increase in their discharge pressure – like some idiot blocking off (deadheading) their discharge line while they are operating. I think this is probably one of your main concerns.

I've addressed you specific questions, so I'll await any further comments or questions from you.
 Natural_Gas_Compressors.xls   854.5KB   913 downloads

Art-

First off, thanks for the reply. I have seen that this is not your first extended explanation and I sincerely appreciate your time. Frankly, it's awesome.

I went ahead and added another worksheet to you excel file. Not being in compression design exclusively, and being fairly young from electrical background, i apologize if there is anything lost in translation.

I agree that we have an analog recycle valve at present. What I am wondering, is whether each unit needs a "bypass" or "unit recycle valve" to properly start and stop as unloaded. Is this a clearer explanation?

My former experience lended itself to a pipeline company who utilized CAT engines/Ariel Recips/Pockets, Unloaders, and Variable Speed for loading. They also used station level recycle valves. The current job requires substantially different loading needs and priorities. I am aware of the inefficiency of this setup, but we are using this compression as a redundant setup to maintain flows in a gas purification setup.

Why do you assume that the recycled gas is cool? Why do you say that I do not need a recycle vavlve for loading and I do for unloading?

I am most worried about deadheading when starting up the engine (assuming that the discharge block is closed). I don't think I can start w/ the discharge open (possible 200psig spread). I am not sure at present how long I will need to idle the compressor before I can load it.

Thanks again,
Brian

Attached Files

•  Natural_Gas_Compressors_rev1.xls   976KB   531 downloads

[Art Montemayor]

Brian:

Quickly, before I head for the house:

1) Thanks for the quick, efficient, and detailed explanations;

2) In my opinion, you will have instrumentation control problems if you try to save a few bucks by using one recycle control valve instead of two. I realize that this is the basic oil patch philosophy and that this will only be a backup facility, but when you double the required max. capacity of the control valve, you are really stretching it’s controllability when only one compressor is running. That is the main reason why the two compressor trains are depicted as duplicates in instrumentation.

3) You and I are fortunate in your prior experience with other recips and a variety of control devices. We don’t have to waste time with a primer in compressor controls and you are essentially up to speed with some process control understandings. This is very important and gives you an edge.

4) The way you show your parallel piping and valving does not allow you to stop & start your machines independently – and I think you will need to do that. Note that if you are controlling capacity with one recycle valve, you cannot use that same valve to stop and start one machine independently of the one that is operating 100% on stream. That’s why my schematic shows the connections with allowance to run each machine independently at the start or stop – with ability to blowdown independently (not shown). These are important safety features.

5) I do not assume that the recycle is cool. I insist that the recycle be cool before being recycled back to compression. If you don’t, you will be feeding hot gas into a recip’s suction and the resulting discharge temperature will immediately start to increase exponentially. You cannot allow this to happen. You must furnish intercoolers as I have shown. You haven’t said how many stages of compression you have or your compression ratios. I have assumed you have 2-stage machines.

6) You haven’t stated what your proposed startup procedure is. We have to discuss this – as well as your proposed startup procedure. I have assumed you are going to startup each machine manually and that you have ability to blowdown the startup and the shutdown gas. Is that a fair assumption? If not, we have to have detailed procedures for startup and shutdown.

As I said, there are a lot of details that both of us have left out for now, but this is already a good start and we have succeeded in exchanging a lot of important basic data and details. The graphic ability of Excel gives us the ability to communicate quickly and accurately because we can exchange both calculations (later) and graphics.

I hope this explains where I am trying to help and that you can understand my explanations. Thanks for labeling your latest data as Rev1. This helps to keep the information organized and orderly.

Art-

The forum database seemed to have been down yesterday. Anyway,

The difference with our current setup is that the recips will be compressing what is essentially a dirty hydrogen mix at which time we are compressing from about 100psig to 300psig. You will notice that this is a much lighter molecule and pressure load, in comparison to natural gas pipeline compression.

I was thinking that the combination of the use, as well as the fact that we are driving the compressor w/ an electrical induction motor, we may not need the bypass. There is no "settling/idling" time necessary with the electric motor in comparison with a gas engine, persay.

We do have an intercooler after the discharge of the compressor, so I am not worried about the recycle temperatures.

We do have blowdown valving designed inside of the block valves. I did not draw. I think we are ok on that.

I am thinking that we may be able to purge through the suction block as it is relatively low pressure and open the discharge shortly after the motor is up. I may have been overthinking this, given my experience in natural gas and relative inexperience in hydrogen compression.


Am I missing anything here?

Many thanks again,
Brian

[shan]

Hi Brian,

Where is your purge valves? They are not shown in your excel drawing. Is it possible to release gas from purge valves during start-up.

Regards

Shan