Reciprocating compressor discharge pressure 4

[Renoyd]

Hello,

In reality, how do operations increase the discharge pressure of a reciprocating compressor? Thanks.

 

[Montemayor]

Reynoyd:

I think you are suffering from a misconception of how a reciprocating compressor develops a discharge pressure.

The pressure is NOT built up by the compressor. The pressure is a characteristic result of the SYSTEM’s resistance to flow. The compressor merely moves (or displaces a volumetric rate to overcome the resistance to flow inherent in the system. The “system resistance” can be a very long pipeline or just a throttle valve. In the case of the pipeline, it is a fixed resistance of the system; in the case of a throttle valve, YOU vary the resistance by either opening or throttling the valve. The reciprocating compressor’s discharge pressure rises to the occasion to meet and overcome that resistance while displacing the same volumetric gas rate (…..well, not really – but close to it).

 

[zdas04]

Good post Montmayor. The compressor is simply reacting to discharge conditions, but the amount of work (and heat) developed is a function of the ratio of suction to discharge pressure so while the machine will almost magically rise to acheive any given discharge pressure it will only do that while the work is less than the maximum work the driver can do and the discharge temperatures are less than the high temp kill.

David

 

[Renoyd]

Good points, Montmayor and David. Just wondering if there are other ways, like the cylinder discharge valve. For the spring loaded valves,if a higher pressure rated cylinder discharge valve is used (Gas has to be compressed to a higher pressure before discharge happens), this will increase the compressor discharge pressure, right? How often do people do this (change the discharge pressure by change out the cylinder valve)?

 

[nickelkid]

Renoyd,

You will just increase the pressure inside the cylinder to overcome a stiffer valve spring, not in the system you are trying to increase the pressure of (also waste power, increase discharge temp., increase rod load, etc...). Resistence is needed to obtain pressure in your system as Monemayor explains.

 

[25362]

I'm not an expert on compressors but as I understand the discharge valves tend to flutter, an effect called valve pulsation, several times during one cycle.

Spring tension is used to reduce the amplitude of these pressure peaks.

 

[zdas04]

As far as that goes, you can put a discharge back pressure valve after the compressor to make the machine achieve any discharge pressure it is capable of. This is a reasonably common band-aid fix to perceived problems (which are usually simply a lack of understanding of the capabilities of the installed compressor). The only time I sanction a back pressure valve in one of my designs is when there is a need for a higher pressure stream (e.g., line pressure tends to be around 200 psig, but a gas lift stream needs 500 psig and the machine is configured for 500 psig it is reasonable to hold the back pressure and pull off a side stream ahead of the back pressure valve, this "waste" of fuel is justified because installing a stand-alone gas-lift compressor is very expensive and would use a substantial percentage of the wasted fuel anyway).

Stiffer discharge valves on the cylinder would truly be a bad idea and I've never heard of anyone trying to get more than an insignificant dP that way (i.e., to try to solve a valve flutter problem).

David

 

[25362]

By reducing the spring tension one reduces the amplitude of the pressure peaks, the number of peaks isn't altered, but the valves are more prone to leakage.

From A working guide to process equipment by Lieberman and Lieberman, McGraw-Hill:

The amplitude (i.e., the heights of the peaks) is a function of the valve spring tension. The frequency of the peaks (i.e., the number of peaks) is a function of the speed of the compressor and the geometry of the suction and discharge piping.

 

[Montemayor]

A reciprocating compressor produces a repeated, batch-type of cycle. In other words, the compressor takes suction gas and compresses it by submitting it to a reduction in volume imposed by a reciprocating piston. That's ONE cycle. Then, the other, identical cycles are repeated - over and over again.

I have a lot of respect and admiration for Norm Lieberman's knowledge and experience - especially for the way and manner in which he liberally and pragmatically distributes it. I join him in my complaints & disagreements with the Carnot Cycle as applied to reciprocating compressors. Norm may have reduced the valve tension at the Hebronville Compressor Station (as he states in his book) and obtained a reduced power consumption, but I can almost bet that he did this only after investigating or consulting with the compressor manufacturer's settings of the discharge valve spring tensions. There is an inherent and dangerous risk one runs in reducing the discharge valve spring tension: Valve leakage immediately is formented or initiated if done excessively. And the consequences are an immediate exponential increase in discharge gas temperature due to severe re-compression of discharge gas. This out-of-control increase in discharge temperature can cause severe damage to a recip.

One can never avoid the inherent pulsation experienced with a reciprocating machine. It comes with the territory. You can alleviate it just so much and then you have to live with it. As stated above, the recip deals out a series of BATCH processes. And in-between each batch there is a discontinuity. This will never be resolved and is a fact of nature.

David is precisely correct in his analysis. One should always think twice in applying a designed backpressure on a reciprocating machine. I have applied control valves on the discharge of reciprocating machines for the purpose of liquefying the compressed gas - such as is the case for Carbon Dioxide, Carbon Monoxide, and Nitrous Oxide. These are applications where one must maintain a constant pressure on the condensation section (compressor discharge at critical pressure) while cooling below the critical temperature. This creates a high pressure liquid which is subsequently expanded to form a low-pressure liquid for storage - basically like a mechanical refrigeration cycle. However, you must instrument your process to fail safe and with adequate manual overrides in case of an upset.

The usual and customary way to set the final discharge pressure is to let the system determine the precise value while the compressor works at rated capacity.

 

[Renoyd]

Good points.

If there is a valve downstream of a recip, can I determine the %opening of the valve with respect to the discharge pressure of the compressor? How? Thanks.

 

[zdas04]

Depends on what you are trying to do. If you use a linear valve (like a Taylor Choke for example) then a 1% change in valve position (after flow is initiated) is about a 1% change in mass flow rate. The arithmetic exists to estmate what that change in valve position would do to upstream pressure, but so what?

Typically a backpressure valve is used to maintain a constant pressure upstream of the valve. It does this with diaphrams and springs and doesn't require any input from a PLC. Why would you need to know valve postion at a particular time?

David

 

[Renoyd]

David, I am just trying to figure out, if the flow is constant and we adjust the recip discharge pressure by opening/closing the downstream control valve, what would be the relationship between the valve position (or dp across the valve) and the recip discharge pressure. Probably this may not only depend on the valve itself, the valve downstream conditions are also related?

 

[Montemayor]

Renoyd:

David is right. So what if you could do it with a manually-adjusted hand valve? There is no reason to do this when you can control the end result with a typical pneumatic control valve or direct-acting, spring-actuated regulator (Fisher "Big Joe", for example).

There IS a relationship between a valve setting and the developed discharge pressure. That is not a problem. What IS a problem is the possibility of a miscalculated movement, obstruction, or over-correction. Human error or decision-making is not 100% controllable.

I would not try to control the back pressure on a reciprocating compressor with a manual valve on the discharge. All it takes is one mis-cue, a slip, or a human error and you are at the total mercy of the discharge pressure relief valve opening on time and correctly when it senses the correct set pressure. If it doesn't a lot of bad things could quickly go South (wrong).

 

[dcasto]

You bring back memories. The first Chem plant I went into had a system to keep the discharge pressure higher on their refrig system. The reason was that if the pressure was too low, the valvle that let the refrigerant into the chiller could not dump enough. I said, "get a bigger dump valve and lower the pressure, you'll save 1000 HP in the winter", pardon the pun, but "too Cool".

 

[djack77494]

I think it helps to think of a recip machine as one that generates a volumetric flowrate. It doesn't care what the suction and discharge pressures are. It's just going to suck x cubic feet or meters per minute from the suction system and discharge them into the discharge system. Come what may. The discharge pressure is determined by the response of the discharge system to the situation.

 

[dcasto]

you know, we are at the point where more details could help. Sorry djack, a reciprocating compressor is not a constant volume machine. There is this thing called volumetric efficency. As you rasie the discharge pressure, the compressor throughput drops for the same suction conditions.

I wonder if there is a bow through condition? why don't they control the unit the old fashion wayt, adjust speed and suction pressure or have a recycle valve?

 

[gurmeet2003]

Dcasto is correct. The reciprocating compressor may be called a constant volume machine, but only in a limited pressure range. If either suction or discharge pressure is varied enough, the flow rate will change as a result of change in volumetric efficiency of the first stage. The horsepower will go through a peak and then fall again. In a multistage machine, the effect of the change in pressures will be less. The volume pushed thourgh the compressor will vary more as suction pressure changes, becuase this has a direct effect on the volumetric efficiency of the first stage, which control the flow.

I have read the discussion on compressor valve springs with interest. The compressor valves are pressure actuated. The purpose of the valve springs is to close the valves quickly at the end of stroke to avoid delayed closing, which can lead to high impacts and valve plate failures. Manufacturers use sophisticated computer programs to size the valve spring for operating conditions (gas composition, rpm etc.). If operating conditions change enough the springs need to be resized.

Spring stiffness has a very indirect effect on the pressure loss. The valve loss is much more dependent on the flow area of the valve. The delta p across the valve is a loss for the compressor. Therefore manufacturers always try to minize this.

Speed control is used for compressor capacity control. It depends on the speed variation allowed by the driver. Suction pressure or recyle contorl valves are not used on reciprocating compressors because they are less efficient. Speed control and clearance volume pockets (Unloaders)are more efficient and therefore are popular.

Compressors have a blow through condition. If a compressor is delivering gas at a discharge pressure so that all stages are active (are contributing to raising pressure). Now if we start reducing the discharge pressure, first the last stage will become inactive, then the second last and and so on and finally the first one.

Gurmeet