Max Disch Temp for H2 Rich Service 2

[sshep]

Several of our make-up compressors in H2 service have high compression ratios resulting in discharge temps almost to 400F. This situation has existed for many years, but this was not original design temp. It is primarily the result of changing to higher purity H2 supplies (higher Cp/Cv). The discharge conditions are 750 psi, 99% pure H2.

API 618, section 2.3.1 (using 4th ed, Jun95) cites "Predicted discharge temperatures shall not exceed 275F (135C) for hydrogen-rich services." Lately this has become a topic of dicussion for us. The equipment has obviously been serviced many times with no noted problems.

My question is: What is the basis or concern which has prompted the 275F limit to be written into API618?

 

[Montemayor]

sshep:

First I'm going to assume that we are dealing with reciprocating compressors and that they are equipped with conventional, cast iron rings and are lubricated with common cylinder oil - like Mobil's DTE 103.

The industry started to get concerned about allowing discharge temperatures on oil-lubricated compressors to increase due to higher compression ratios being used (to economize on number of stages) long before the API got concerned and issued their recommendations. This started around the late 1950's. The basis for the concern was not the Hydrogen application, but that air compressor cylinders were exploding (since these are mostly of cast iron construction, they were literally "grenades") and killing some people. This was predominantly happening in air separation plants where high pressure air (3,000 psig) was generated as per the state of the art at that time. I know because I started in air separation plants in 1960 and saw some of these happenings. There is a lot written about this and a committe was formed by the AICHe to investigate and report yearly on Safety in Air Separation and Ammonia Plants.

When we examine the thermodynamic equations and relationships involving compression, one can see how the discharge temperature can climb exponentially due to higher compression ratios. On a more serious level, one can also investigate the carbonization and explosivity of cracked lube oil at temperatures exceeding 350 oF and the result will be a clear explanation of why explosions can occur in air compressors lubricated with oil: Oxygen in the air will combine with any cracked lights from the oil and spontaneously ignite them. One Cooper Bessemer cylinder in the Mexico City suburb of Ecatepec leveled the process building around 1965.

The high temperature effect on cylinder oil is well appreciated by me since I've had to literally scrape and dig out handfull's of coked and gummy deposits from compressor cylinders firsthand. This is not only an occurance in air compressors but has also occurred to me in Hydrogen service. I never tolerate any reciprocating compressor specification that calls for a design discharge temperature higher than 275 oF. I fully expect a recip to operate no higher than 250 oF. And I control this by the compression ratio imposed on the cylinder.

You may have a "Teflon" ring (or other material) compressor with either no lube or mini-lube. You haven't stated this. However, a high temperature will degrade your non-metallic rings as well and further complicate the entire situation if you have double-acting pistons. You are in no danger of a cylinder exploding due to lube oil breakdown in a Hydrogen application, but you will damage the cylinder and its components should the situation continue long enough due to the solid, hard carbon coke formed by the oil breaking down.

I don't understand your explanation about the k (=Cp/Cv) for Hydrogen. The k for air is the same as that for Hydrogen (1.400), but I don't know what the properties of your impure H2 are. It may not impact you very much now, but if you are operating your oil-lubed cylinders above 350 oF for sustained periods, I can attest to the tendency to form carbon deposits and possibly damage your machine(s). I find it hard to believe that a Cp/Cv value change would increase the discharge temperature significantly. I would rather suspect that the machines have a built-in compression ratio that is higher than normal; but I would need far more data to speculate further.

I hope the above helps explain your question and concern.

Art Montemayor
Spring, TX

 

[sshep]

Thanks Art, I have been waiting to get one of your detailed explanations.

To clarify this is a recip compressor. The original design was for a heavier offgas containing less than 50mol% H2, and having a k=1.28. This was 1965-ish vintage, and that design seemed more interested in finding the max HP case than in any discharge temp limits. The difference in adiabatic discharge for a 200psig to 750psig compression of 100F gas is fairly significant (if I did my math right): 345F(k=1.4) vs 279F(k=1.28)

Your comments about air are well taken and easy to understand. Air/O2 is discussed briefly in API618, although the guidence is not as qualitative as desired by the seriousness- i.e. we are told to "consider" lowering discharge temp limits (from 300F) for air, but no actual recommended value.

The basis for API's H2 guidance is still not well understood because it applies to both lubricated and non-lubricated systems. We would also think that MOC could be chosen for H2 applications such that no hard limit would need to be cited- by now you probably know that most common H2 handling concerns are known to me.

Anyway, I will gladly take further comment as to what established the 275F discharge temp limit for H2 service. Thanks, sshep

 

[Montemayor]

sshep:

I forgot to add in my lengthy explanation that non-lube service in Hydrogen (or any gas, really) would have a temperature limit set on it due to the thermal degradation that Teflon is limited to - usually around 400-500 oF.

I also forgot to mention the ancient concern that all recip compressor operators take to their beds every night: beware of losing your cylinder clearances! In other words, if you overfeed the oil lube - as I've found some operators doing in order to make up for excessive ring wear- you start to create a lot of solid or gums from the degraded oil in the cylinder(s) and the clearance between the piston and the cylinder end wall starts to diminish -- possibly getting down to "zero". This would be analagous of the other nightmare: allowing liquids to enter the compressor cylinders. I've seen the results of "zero" clearance as well (in the case of liquids entering the cylinder) and the results in one case were benign: the piston rod bent and the crank's bearing failed. No steel or cast iron flew around; the machine merely stopped dead in it's tracks without hurting anyone.

I thought you might be interested in these other two effects that could result from excessive discharge temperatures.

Art Montemayor
Spring, TX

 

[25362]

In one refinery I worked with they switched from highly refined specially selected lubes to polyalkylene glycols for hydrogen compressors for temperatures of up to 200^oC to the operators' satisfaction.

It appears that beyond this temperature, and even without any protecting additive, PAG lubes decompose without forming any kind of deposits in service.

It is claimed that by not been washed away or diluted by any hydrocarbon, dry running and viscosity drop in the pressurized section, respectively, were thus prevented. This is being considered one good reason to use PAG lubes for the compression of hydrocarbon gases, e.g. natural gas and refinery gases.

 

[sailoday28]

sshep (chemical) For my information. You final temperatures are based on isentropic compression of perfect gas. Can one use real gas with some sort of compression efficiency?