Generator Hydrogen Cooling System


Thread Starter


I want to know how a hydrogen cooling system for GE frame6 generator works. Anyone who can provide me drawing or a brief tutorial will be helping me.

And my email address is seshu_inst [at],
The concepts of a hydrogen-cooled generator are very simple. As current flows in a conductor, heat is generated. A generator has a lot of conductors and a lot of current flowing through the conductors, generating a lot of heat. If that heat isn't "removed" then the windings will be damaged (insulation "blisters"; conductors grow and elongate causing clearance and balance issues; etc.). In addition, in a synchronous generator (alternator), there are high currents flowing in the rotor windings, which also generates heat which must be "removed".

Air can be used to cool a generator, by circulating it through the generator to absorb heat and then exhausting the air to another area outside the generator. A continuous flow of air from outside the generator, through the generator, to another area outside the generator will cool the generator and rotor. (The presumption is that the air entering the generator is cooler than the generator.)

Another way to cool the generator is to use hydrogen gas circulated through the generator and around the rotor to cool things. Hydrogen is seven to ten times better at transferring heat than air. That is, hydrogen is much, much better at absorbing heat and then at giving up that heat to another medium/area than air. This means that for the same size generator, if it's cooled with hydrogen versus air that more current can flow in the stator and rotor windings which means that more power can be produced. Or, thinking about it a little differently, the same about of power can be produced with a smaller generator cooled with hydrogen than one cooled with air, which is the typical reason for using hydrogen cooling--to reduce the physical size (and cost) of the generator.

The "problem" with using hydrogen as a cooling medium is that it's explosive when mixed with air and exposed to an ignition source. However, if the purity of the hydrogen is maintained at a very high level, meaning there is very little or no air in the generator casing to mix with the hydrogen, then even if there was a "spark" there won't be an explosion.

So, to use hydrogen to cool a generator it's necessary to prevent air from contaminating the high-purity hydrogen used to fill the generator casing after displacing the air. And that's important, displacing as much of the air inside the generator casing before filling it with high-purity hydrogen.

The hydrogen gas inside the generator is usually at a pressure of approximately 2 barg (30 psig), which means that air cannot leak "into" the generator casing where the hydrogen is. In effect, this is the primary method of preventing air from getting into the casing and contaminating the hydrogen.

The hydrogen is circulated by fans on the ends of the generator rotor, and as it's circulated around the generator it passes over coolers which have water circulating through them. The heat which is absorbed by the hydrogen gas as it passes through the generator and around the rotor is transferred to the water in the cooler. As the hydrogen exits the coolers, it's recirculated back to the generator and rotor, in a continuous cycle.

Another important consideration is to keep the hydrogen from leaking out of the generator, mixing with air and causing an explosion- or fire hazard where it might leak out of the generator.

The generator rotor rotates where it passes through the end-shields and that is the area that must be sealed to keep the hydrogen in and not allow it to leak out. A hydrogen-cooled generator uses "seals" to keep the hydrogen gas inside the generator.

The hydrogen seals are on the two shaft "ends" that pass through the generator end-shields. Oil is typically used as the sealing medium, and is sprayed on the shaft around the entire circumference of the shaft. The "seal oil" is at a higher pressure than the hydrogen inside the generator casing. Some of the seal oil flows out of the seal area along the shaft to the "air" side of the generator and some of the oil flows out of the seal area along the shaft into the "hydrogen" side of the generator.

The oil that's used as the seal oil is generally the same lubricating oil that's used for the bearings. That oil is normally in contact with air when it's in the lube oil tank and the bearing drains. So, air (in the form of small bubbles) can be entrained in the lube oil, and when sprayed on the generator shaft that air can be liberated from the oil that flows into the hydrogen side of the seal area. That air, if not "removed" somehow, can continue to collect inside the generator casing and reduce the purity of the hydrogen, and cause a safety concern.

So, because the hydrogen inside the generator casing is at a higher pressure than outside the generator casing, air can't leak into the generator. And, because oil, which will have entrained air in it, is used as the sealing medium, the air released from the "seal oil" that flows into the hydrogen area can reduce the hydrogen purity if not removed. So, the primary source of air to reduce hydrogen purity (contaminate the hydrogen gas inside the generator) is air liberated from the oil used to keep the hydrogen inside the generator from leaking out along the shaft.

So, there is a system to remove the air which is liberated from the seal oil that flows into the hydrogen side of the generator seal. That system is typically called a "scavenging" system. A small amount of gas is allowed to be vented from the seal oil enlargement tank, where the entrained air liberated from the seal oil is hopefully contained. That vent is normally piped to a safe area of the atmosphere away from any ignition source since it will have hydrogen gas in it as well as air.

Because a small amount of gas (air and hydrogen) is continually being vented to atmosphere through the scavenging system, the pressure inside the generator would decrease if nothing else were done. However, there is a pressure regulator from a source of high-purity hydrogen that maintains the pressure by flowing a small amount of high-purity hydrogen into the generator casing to maintain the pressure.

There is usually a hydrogen purity monitoring system which can be used to detect changes in purity during operation. Again, the presumption is that the hydrogen purity is high to begin with (after the generator casing is purged of air and filled ("charged") with hydrogen). But because of the air which can be liberated from the oil used as the sealing medium to keep the hydrogen from leaking out of the generator, it is necessary to have a means of monitoring the purity, usually at multiple points in the generator, to be sure that the purity is maintained to prevent a possible explosion or fire.

The scavenging flow rates must also be set and monitored to prevent excessive hydrogen consumption (to replace the hydrogen lost with the air that's being vented to atmosphere).

The seal oil that flows into the hydrogen side of the seal area is usually directed to an area called and "enlargement tank" and that's where it's hoped the entrained air is liberated and "contained", and vents from the enlargement tanks are directed through calibrated flow-meters to atmosphere (the scavenging system). The hydrogen purity monitors are usually capable of monitoring the purity of the gas in the top of the enlargement tanks that's vented through the scavenging system.

The hydrogen purity monitor can also be set to monitor the purity inside the casing (which should be higher than the enlargement tanks, theoretically).

Usually, when the hydrogen purity monitor indicates the purity is decreasing, then it increases the scavenging flow-rate to try to increase the purity, but venting more "contaminated" gas to atmosphere and replacing it with high-purity hydrogen.

If the hydrogen purity drops below a certain level (around 80% or so, depending on manufacturers' recommendations), then usually the generator and prime mover are stopped and then the generator is purged of hydrogen. (CO2 is usually used as the medium for purging air from the generator when filling with hydrogen, and also when purging the hydrogen from the generator.)

Presuming the hydrogen being used to maintain the pressure while scavenging is high purity, and because the casing pressure is usually two barg, the most likely way that purity can be decreased is from the air entrained in the seal oil. If the seal oil flow-rates (usually monitored with a single (unfortunately) flow meter) increase, then the air liberated from the seal oil will increase and that is the primary source of contamination and decreased purity. (This presumes that the purity monitors are working correctly.) So, it's very, very important to monitor seal oil flow-rates to detect an increase which would usually result in a decrease in purity. But, with only a single seal oil flow-rate meter for seals at two ends of the generator, it's difficult to detect which seal is degrading and consuming more oil which is liberating more air (but the decreased purity in a particular enlargement tank is the key ).

That should be enough of a description for a newbie or novice. If you have specific questions, we'll try to help. If you're working at a site with a hydrogen-cooled generator, you should be able to find the seal oil piping schematic (P&ID) which should be helpful in understanding the seal oil system. If the generator was manufactured by GE, it will usually have some drawings in the Instruction Manuals that are very helpful in understanding the systems in use on the generator at the site (though most are very similar, there can be subtle, but important, differences).

I read your message and simply superb. I am having small doubt that "What are the basic reasons for improvement of seal oil flow of the generator to maintain the same DP across the seal oil and the Hydrogen in side the Generator"

Please help me..

Sir, I sincerely hope you do <b>not</b> have a doubt. You have a <b>question</b>.

Oxford English Dictionary defines doubt as a feeling of distrust or disbelief; to question the truth of.

Nothing we say here is intended to mislead anyone. And it's perfectly acceptable in polite company to say, "I have a question", or, "I would like more clarification." But, to express doubt is to express distrust or disbelief.

I will answer the <b><i>question</b></i> of what would increase seal oil flow rates, and that is that the "seal" is actually a brass ring with an inner diameter slightly larger than the outer diameter of the alternator rotor. It is held tightly around the shaft with springs. Oil is admitted through ports and grooves in the seal rings and "sprayed" onto the shaft.

The brass rings and holding springs can wear or even fail over time. Dirty oil can also cause wear of the seal rings. The generator rotor shaft, as all shafts in journal bearing, moves as it is accelerated and even axially with the application of torque and electrical forces (though axial movement is generally very small, as is the radial movement). Unattended high vibrations of the generator shaft can also cause problems, as can induced voltages that are not properly grounded (through the shaft grounding brush). Also, the seal oil pressure regulators do require maintenance and adjustment and can be the source if increased seal oil flow rates.

We are happy to answer questions here at I imagine on occasion some questionable information is provided but never intentionally. The form of your question suggested it was not caused by any doubt, rather that you sought further information and clarification. Which we are happy to provide when asked, but not so happy to provide when requested by expressing doubt.

Luis Phillippe

Depending on the price for H2 bottles, the process will be costly and you have a good opportunity to improve it.

Could you approach any solution for the H2 purity decreased ? Like using a vacuum treatment for the seal oil...

In other hand, instead of solving the main problem, installing a H2 generator connected to the H2 supply line would be a feasible solution. Do you have any experience with this kind of system?

The problem with having a separate seal oil system is that at some point the seal oil and some of the bearing lube oil will eventually mix, and so I believe that's why the same medium (lube oil) is used for both bearing- and seal oil. Treating a portion of the lube oil is going to be costly and difficult, at best.

Decreasing H2 purity is primarily a function of seal oil flow rate, and it can be affected by moisture in the H2 and from leakage from the cooling water used in the H2 coolers. There are dryers and desiccants used to help remove moisture but a lot of machines don't have the dryers and only very small desiccant dryers which are rarely, if ever, properly maintained. Also, oil in the belly of the generator (from improperly charging methods and internal leaks) can also cause H2 contamination.

One has to remember: The casing is usually maintained at approximately 2 barg, or at some pressure above atmospheric. So, air can't leak "into" the generator; H2 can only leak out. So, the only real source of air is from the lube oil used as the seal oil medium. Presuming the purity monitors don't drift (and that's NOT a given with the old manual Wheatstone bridge used in many H2-cooled generator purity monitoring systems) about the only source of air is what's liberated from the seal oil.

Most H2 usage comes not from the scavenging replacement, but from leaks--leaks of H2 out of the generator casing that cause the pressure to decrease. Actually, this should help increase the H2 purity because of the higher replenishment rate due to the leakage, because, again, air cannot leak into the generator.

There are many sites around the world which have opted for on-site H2 generation. It just adds another system to be operated, monitored, and maintained. And in my experience, they were never operated or maintained properly and so most of them were soon shut down and bottled H2 was used for the generators.
Great submission ....thanks.

A question: What can be the possible causes for excessive H2 entrained in the seal oil drain to the enlargement tanks?
Failing hydrogen seals; improperly installed hydrogen seals; failing or broken hydrogen seal springs; excessive Seal Oil pressure are the usual culprits. Dirt/debris has also been known to get into the seals and cause problems, usually the brass rings get scored and sometimes the shaft, also. This is an extreme case, but has happened.

I have a PowerPoint presentation I can send you but I do not know how to add attachments to the site, Do you know?


Your friendly local moderator here. The forum does not do attachments. People can post and include their email addresses and ask for the PowerPoint presentation; we will post the messages to the list.

Peg Ferraro
Dear CSA,

Can we use nitrogen instead carbon dioxide to purge the hydrogen? as both nitrogen & carbon dioxide are inert gas, but the density of CO2 is higher than the nitrogen. but nitrogen is still heavier than hydrogen.

People have tried it before with very mixed results. It depends on several factors, first, the patience of the people purging and charging the generator (which is usually not very good, and neither is that of management waiting for the purge and charge to be completed).

Second, the instrumentation available. Usually one of the meters is set to read CO2 in Air, or something similar, so I believe it's not really expecting <i>any</i> inert gas.

Third, I think it's particularly difficult to purge the casing of air with nitrogen before charging it with hydrogen.

In my personal opinion, purging and charging (regardless of the gases and instrumentation used/available) needs to be done slowly if for no other reason than to reduce the mixing of gases during the processes. Slow and steady wins this "race": the race to purge the casing of one gas as completely as possible and then fill it with another as quickly as possible with the highest purity possible. Leaving any air in the generator casing when charging, whether CO2 or nitrogen was used for purging, is not desirable.

A lot of people ask why the purity goes down during initial operation after a purge/charge, and it's likely because of two things. First, the scavenging rates need to be adjusted, and second, because it's likely that the purge and charge wasn't as effective as believed because of the rush to complete it. So, it usually takes a little more scavenging to get the purity up, and then it can be adjusted lower.

That's the part that most people don't understand about scavenging: It's b>NOT</b> 'set-it-and-forget-it.' It needs to be adjusted over time to maintain purity, and <b>ANY</b> adjustment needs to be logged for tracking and trending purposes which can be very useful when trying to determine if hydrogen usage becomes excessive. And adjustment can be to decrease scavenging as well as to increase it. Remember: The adjustments are made to maintain purity. And any adjustment is probably going to take several hours to seen any effect; it's not immediate.

So, patience is the key.

But, to answer your question simply, it's your machine and you're free to try it. It's been done with mixed success, and if you're willing to take the chance and live with the consequences, then it would seem to be feasible. But, then, a lot of things look easy "on paper", don't they?

The first question that should be always be asked (in lower-case letters, thank you very much!) is: When did this problem start?

If you're consuming large amounts of hydrogen and have high scavenging rates, it could be the seals and/or seal springs are failing or are excessively worn. I've heard that the float mechanism in the float trap can cause this, but I've no personal experience with that.

I've also seen brand new generators with new L.O. have foaming all during commissioning, and I don't know if it ever lessened or stopped. I think some formulations of turbine oil are more subject to foaming under some conditions than others.

We don't know if yours is a new unit, with new L.O., or if the L.O. was recently replaced with a new or different formulation from the same or a different manufacturer. We don't know if there was a recent maintenance outage when work was done on the generator end-shields or seals.

I also have a suspicion that sometimes the "foaming" comes from some higher pressure somehow downstream of the seal oil float trap that is making it's way back into the trap, but I'm not exactly sure how that could happen. I only say that because I witnessed bubbles rising through the sight glass in the float trap. And that's the only reason I could come up with. But, the bubbles stopped in a couple of days (new unit being commissioned). It could have just been gas trapped in the piping somewhere; it was never investigated completely.

But, it would be helpful to know when this problem first started and how long it's been persisting and anything you might have done to try to investigate and troubleshoot the foaming.
thanks a lot sir,

it is really very helpful for me. sir i have one question, what is the function of liquid level detector (i am fresher in CCPP and working in the Maintence side, ours is ge 9fa machine)? is this any how related with float trap assembly).

thanks in advance.
A liquid level detector is for detecting any liquid in the generator casing.

There shouldn't be any liquid in the generator casing under normal operating conditions.

Sources of liquid could be leaking hydrogen coolers and excessive seal oil flows.

Please find the Generator Hydrogen System P&ID drawing. Also, if you have a unit packaged by GE you should have a section in the Instruction Manuals that describes, in very brief detail, how the system works. It's not much, but it's better than nothing.

Learn to use the available documentation at your site. We can fill in gaps and details here at
Honeywell has made a thermal conductivity analyzer for hydrogen cooled generators for decades. The thermal conductivity cell has not changed since the 1970's.

The electronics box is known as a Triple Range Analyzer because it provides concentration readings for running and purge cycles.
Range 1: 0 to 100 % CO2 in Air
Range 2: 0 to 100 % H2 in CO2
Range 3: 0 to 100 % H2 in Air

The electronics that determine the concentration from the thermal conductivity on-line cell raw data definitely expect H2, CO2 or air. not nitrogen in place of CO2. That doesn't Nitrogen can't be used, but any thermal conductivity analyzer expecting CO2 will read wrong when used with nitrogen.