Explosion While Replacing Generator's Hydrogen for CO2/Air


Thread Starter


First of all hi to everybody, i'm new here. Thank you in advance and sorry for my english.

So this happened few months ago.
While purging the generator we had and explosion and the rotor has to be replaced because of the damage, bad news. They will have to work in the stator a little bit as well.

The mechanical team purged first the hydrogen with CO2, (checking the purity) and once the purity of H2 was low enough, they started to pressurize with air.

The thing is that once they started to pressurize with air, we had no pressure of CO2 in the generator. Was it correct, or they screwed up?

Thank you very much. By the way, machine is 7FA.
In my understanding of explosive atmospheres, one can have a volatile mix of gases (H2 and air), but there needs to be a spark or an ignition source of some kind (most gases--some just ignite when exposed to air or some component of air). The ignition source might be something with a temperature sufficient to ignite the explosive mixture, but simply having a volatile mixture of gases won't result in an explosion. Remember: fuel, spark and air--just like when establishing flame in the combustors of the gas turbine. And they all have to be in the proper proportions for flame to be established.

Most synchronous generators produced by or for GE-design heavy duty gas turbine power generation applications have internal space heaters which are energized whenever the generator breaker is open. This is to prevent condensation inside the generator casing when current stops flowing in the stator- and field windings. Even hydrogen-cooled generators have internal space heaters.

One of the very common mistakes made when purging and/or charging GE-design heavy duty gas turbines is that the proper valve sequencing is not followed. CO2 is heavier than H2--and air. So, CO2 is supposed to be admitted through a manifold in the bottom of the generator when purging H2 from the generator, and H2 is supposed to be vented through a manifold in the top of the generator. Then when trying to purge CO2 from the generator the valve positions are reversed--air is admitted through the top manifold and CO2, because it's heavier than air, is pushed out through the manifold in the bottom of the generator.

The location of sampling, and the method of sampling is very important, also. Sufficient time and venting must be allowed for the gas sample to be representative of the gas in the generator casing. This is another very common mistake.

The biggest mistake made when purging/charging is haste--if the gases are admitted too quickly into the casing, even through the proper manifolds, there will be a lot of mixing of gases in the generator casing and sampling will be adversely affected.

Another mistake made when purging is to shut off Seal Oil, or not to watch the casing- and Seal Oil pressure and -flow-rate. H2 can leak out of the seals when purging and into the load- or collector compartments where they can be ignited. Seal oil needs be maintained, and care needs to be taken to prevent oil from getting into the generator casing as casing pressure drops.

Finally, a portable gas analyzer was usually provided with most GE-design hydrogen-cooled generators--it had the capability to be used to sample various places (seal drain enlargement tanks, for example) and to measure CO2 and H2 mixtures. It, does, however, have to be maintained and calibrated--neither of which usually happen at most sites.

Sorry to hear about the accident, Odrog; hopefully no one was injured or maimed, or worse. Training is critical to safety--especially around hydrogen-cooled generators. Anyone performing purging/charging operations should be thoroughly trained, have adequate visual aids for the systems (P&IDs at a minimum; pictures of valves and manifolds, preferably), and be using a written site-specific procedure reviewed and approved by Operations Management, Site Management, and any safety officer/manager. Such documents are living, breathing documents--meaning they are not static (written once and never reviewed or revised)--rather every time an incident like this happens, or even a near miss identified during a procedure, the written procedure should be immediately reviewed by all involved (and even those NOT involved) and revised as appropriate. Sometimes, more training is advised, or a review of previous training is in order.

Be safe!

(I was always taught that industrial accidents don't happen--they are the result of some human error or misperception or haste or misunderstanding. And, I have come to believe this is true. Doubtful the readers of control.com will ever know the exact details of what happened at the original poster's site, but it would be a great learning experience if we did. Leaving the generator casing space heaters on is not an accident; shutting off Seal Oil is not an accident; purging/charging through the wrong manifold is not an accident. These are all the result of human error. Even something like damaged insulation on a space heater wire termination inside the generator casing--which may have occurred in the factory during assembly--is not an accident. This should have been caught by quality control prior to final assembly/shipping. Safety accidents don't just happen, and they are all learning opportunities for those involved, and those not involved.

"Good judgement comes from experience, experience comes from poor judgement."

And, this is true when learning from the mistakes and experiences of others--so that we don't have to make the same mistakes of judgement leading to bad experiences, or worse.)

So, the upshot is--there needed to be a source of ignition (heat; spark; etc.) as well as a volatile mixture. The sampled gases either weren't properly representative of the actual casing gas composition, and/or there was something that was hot enough to ignite the volatile gas mixture. Pure hydrogen won't ignite. So, some mistake, somewhere, sometime, led this unfortunate incident. (Again, there are no safety (accidents.")

Live and learn--and thanks for the lesson!
Hey there:

From my experience in performing various generator purge a 9FB, and my English translator of google, I want to comment the following:

All the purging and charging process is performed by the operating personnel, maintenance mechanic changes only change the Removable Spool Pieces. We have a working technical instruction designed to prevent explosive mixtures, for emptying and filling the generator. During the whole process has to follow a checklist, the emptying process consists of about 35 steps, a key step is the change curve analyzer H2 H2 in CO2. We injected CO2 until CO2 H2 concentration in less than 5%, always keeping the pressure generator on 10 - 15 psig.

As for the cause of the ignition of the explosive atmosphere only add a few comments to those already made by CSA:

Mixtures of hydrogen/air the may inflame through ignition sources that contain very little energy. The least amount of energy needed to ignite hydrogen is 0.019 mJ which is only 1/10 of propane gas. For example, oxide particles, which have been transported by a rapid stream of hydrogen can develop the ignition spark through electrostatic charging or striking an object, can cause an ignition.

I have no experience of hydrogen cooled generators. But curiously, I was reading this post.

As you told the hydrogen will accumulate at the top part of the vent. Is not there is a method to vent out the accumulated hydrogen by opening the vent during purging? this will ensure venting out of hydrogen from the top most part? a more practical approach.

> So, CO2 is supposed to be admitted through a
> manifold in the bottom of the generator when purging H2 from
> the generator, and H2 is supposed to be vented through a
> manifold in the top of the generator. Then when trying to
> purge CO2 from the generator the valve positions are
> reversed--air is admitted through the top manifold and CO2,
> because it's heavier than air, is pushed out through the
> manifold in the bottom of the generator.

Again, there two manifolds inside GE-design hydrogen-cooled generators--one at the top, and one at the bottom. There are manually-operated valves (and solenoid-operated valves)--and a manual three-way valve to alternately connect one manifold or the other to the atmospheric vent. This is how the various gases used in purging/charging are admitted and vented, using the relative weights of the gases to aid in the processes. Mixing of gases during purging/charging is a real problem--as likely evidenced by this post; there had to be some hydrogen and air in an explosive mixture inside the generator casing--along with an ignition source--for an explosion to have occurred. So, that's why the designers put two manifolds inside the generator casing (one at the top and one at the bottom) so that, with proper caution and gas sampling, the purging/charging processes can be done with as little mixing of gases as possible--again, as long as proper care is used.
Thank you very much for the replies guys! Very very helpful.

I'm telling you now, there was no procedure, no checklist, and no proper training. The thing is that they had done it few times so they though they knew how to do it (maybe they were just lucky first times). GE was there as well, as advisor, so everybody became overconfident.

Other thing is that we do commissioning and often we try to rush to save some time. Not a good idea.

Thanks again

GE-design heavy duty gas turbines supplied with hydro-cooled generators that were packaged by GE USA <b>ALWAYS</b> had generic procedures for purging, charging and operating the generators. Those procedures should be the basis for any commissioning activity and plant-specific operating procedure for purging and charging. I would be extremely surprised to learn of a GE-design heavy duty gas turbine package which included a hydrogen-cooled generator, regardless of packager, that did not have a generic procedure that commissioning personnel weren't using and that al least someone in the plant operations or safety department didn't know about and refer to.

I and many others have commissioned units for decades--and was almost always under pressure to complete the commissioning in less time than had been allotted in the project schedule, many times before construction was anywhere near complete! Project and Commissioning Managers always say, "Just do the minimum to get us generating power--and then we'll shut down and finish everything as it should have been done!" Which is a lie, because once the revenue watt-hour meter starts turning and the money starts coming in--they can't shut down, and the project will take months longer to finish. (I know of one combined cycle plant where we did that--just the minimum required to be able to generator steam and electricity. It took two years longer to finish because they had to have the steam and couldn't shut down and more than half the motor-operated valves in the plant weren't commissioned and had to be operated manually. That plant was tripping several times a week--sometimes a day--and the control system was always blamed. But it was Construction and Commissioning that were to blame. And I kept having to return on nights, weekends and holidays (many of my wife's birthdays) to do little bits of work that should have been done during commissioning, and always being told how bad the control system was.)

But, shortcuts on hydrogen-cooled generator purging/charging activities was not a risk anyone was willing to take. (Sometimes Plant- and/or Commissioning Managers had to be reminded of the explosive nature of hydrogen, repeatedly,, but usually they would exercise good judgement and order safe procedures be followed (grudgingly in some cases).)

Commissioning--and operating--power plants is really all about risk management, because there are always risks and with proper understanding and good judgement a surprising amount of risk can be tolerated. But, risks with hydrogen--that's not worth lives and lost generation. That unit is probably out of service for weeks if not months--at a great loss of revenue (income) not to mention the expense of repairs and payroll for people while there is no income being derived from that machine.

The time waiting for parts and repairs should be utilized in training and procedure review. Yes, it's yet another additional cost--but next time it could be someone's life or their livelihood if hurt and unable to work.

In many places in Europe and North America a person can't take one single voltage measurement without filling out a Job Hazard Analysis form and having it approved by at least two people if not a committee. Why? Because untrained people were injured or killed taking a simple voltage measurement, or injured or killed others doing so. That's how lawsuits get filed, lawyers get rich, and onerous safety laws get written At some sites I've worked at no circuit can be energized when troubleshooting or maintenance activities are taking place. And I was on-site to troubleshoot or perform maintenance on these circuits because a project schedule had been written and reviewed weeks in advance but no provision was made for de-energizing the circuit. LOTOs had been written and couldn't be changed (they were inflexibly developed and written). But it wasn't until the controls guy arrived to do the scheduled work did anyone realize that the circuits had to be energized --and then the Safety Manager refused to let the work occur after reading the Job Hazard Analysis.

And all because some where ,some time people had taken ill thought-out risks--and people were hurt or killed. Trace these "safety procedures " back far enough and such sad occurrences will always be found.

And many Safety Managers these days have no actual work experience, but have to write procedures as part of their job. Who wouldn't like to work on de-energized circuits? But in the controls business, most often that's not possible. Or, to do so some process or the plant has to be shut down. And it's always the controls guy who isn't flexible and can't find a way to do the work that's the cause of the schedule delay. Not the people who wrote the schedule or reviewed it--but the controls guy.

It's a serious problem--and getting worse. Untrained power plant operators, and supervisors not ensuring people at least have and follow minimal procedures with minimal supervision. Hydrogen, and hydrocarbon-based fuels--is hazardous to life if one doesn't understand the risks and relies too heavily on automation.

And <b>THAT'S</b>the real mistake being made hiring, and allowing, untrained people to operate power plants: automation is being unrealistically expected to protect equipment and people. Automation can do a lot--and can in some cases do more. But in today's world if it's programmed to protect against every possibility then it will shut down or trip the equipment so often that the control system will be deemed unreliable. Operators need to have training to understand systems and risk and take appropriate action to avoid excessive trips and shutdowns that only an extremely intelligent automation system (which would be extremely expensive) could hope to avoid.

But, managers making these hiring decisions are paid to manage risk--and many of them believe the automation is better/smarter than it is.

And it's not.

And when it is, then people won't have jobs operating power plants.

I agree with the point of admitting air into the hydrogen header (while purging CO2 out of the machine) given the difference in gas densities. My concern is that our GEK instructs us to admit the air into the CO2 header to ensure there is no chance of air mixing with a "pocket" of hydrogen in the top header. Their justification is that air is relatively cheap and you should be able to use large flow rates to combat mixing. This has always been hard to convince operators at our site to do but the GEK is where the procedure was derived. Any thoughts?

> Any thoughts?

I have a lot of thoughts; on this topic I have several.

First, I don't have access to the hydrogen/CO2 P&IDs for the unit in questions, nor the GE instructions. So, I can't comment any more than what I already have on the subject of purging and charging. How many manifolds are in the top of the generator casing, and in the bottom of the generator casing?

If one purges the generator casing of hydrogen properly it's not very likely that there will be much air in the hydrogen header--for two reason. First, a proper purge (one done using the portable gas analyzer) will be done until there is no hydrogen present and 100% CO2. This is because the CO2 is being admitted from the bottom of the generator because it's heavier than hydrogen and pushing the hydrogen out of the hydrogen pipe at the top of the generator. So, all of the hydrogen should be pushed out of the hydrogen piping by the CO2, and only CO2 should be present in the piping (hydrogen pipe and vent piping). The presumption is the portable gas analyzer is connected to a sampling port where the concentration of air is likely to be the highest during purging.

Next, if the generator casing is to be entered for some reason the CO2 is purged from the generator casing using dry air admitted through the hydrogen header--because the CO2 is heavier than air and it is vented out of the CO2 header in the bottom of the generator casing by pushing it down and out using air admitted from the top of the casing (usually through the hydrogen header).

Prior to charging the generator casing after a maintenance outage, all of the air is to be purged from the generator casing using CO2 admitted from the bottom of the generator--which should push all the air up and out of the manifold at the top of the generator. AND, before admitting hydrogen, the portable gas analyzer should indicate no air in the casing (which includes the manifold at the top of the casing).

So, there <i><b>should be</b></i> no air in the manifold at the top of the generator casing when hydrogen is admitted to the casing through the manifold at the top--as indicated by the portable gas analyzer measurement when purging the air with CO2, and vice versa.

Finally, there needs to be an ignition source--a spark, or a very hot surface--to ignite an explosive mixture of air and hydrogen. Admittedly, hydrogen in air requires a very low energy ignition source, 1/10th that required for gasoline in air. So, there is always an elevated risk with hydrogen. <i><b>That's why purging is SO important.</b></i>

Most sites have seen calculations of how many bottles of CO2 are required to purge a generator--and that's what they have available for purging. And that's all they use. They ass-u-me that using the calculated volume of CO2 will result in a complete purge of air from the generator casing and all associated piping. Many sites don't even use the portable gas analyzer during purging--that's right, they've purged and charge "...so many times..." they <b>know</b> it's going to be correct and safe. (I frequently see the portable gas analyzer rusted and dirty and obviously damaged (physically and by water), laying in a corner of the area reserved for hydrogen/CO2 cylinder storage. It couldn't be used even if someone wanted to use it.)

Familiarity breeds complacency, and complacency leads to "accidents" (by the way, it's widely regarded that there are no such things as industrial "accidents"--there's just ignorance and lack of training and when the two combine, the results can be seriously dangerous and even deadly).

Connection of the portable gas analyzer to the sampling port/location where there's most likely to be air is the single best way to ensure there is no air in the generator casing and associated piping. Full stop. Period. A properly working and -maintained portable gas analyzer is key to performing proper generator purging and charging. (Which implies it will be used to check for air/CO2 during the purging and charging processes....)

Finally, since there is a very large difference in the weight of CO2 and air I find it very hard to understand how air can be used to push CO2 up and out of the generator casing <i>unless it is admitted <b>SLOWLY</b> into the bottom of the generator while CO2 is being vented out the top of the generator casing</i>. Unless there is another manifold at the top of the generator casing which I'm unaware of if air is admitted through the manifold at the bottom of the casing then the only way for air to get out of the casing is through the manifold typically used to admit hydrogen at the top. Granted, dry air is much less expensive and more readily available (though it usually comes from instrument air systems--which means it's been compressed and dried which isn't free but is low-cost) so as long as the air is admitted <b>SLOWLY</b> into the bottom of the generator casing as CO2 is vented out of the top of the generator casing it's probably in the best interests of preventing even a small chance of a fire/explosion.

But, I've never seen a site <b>SLOWLY</b> purge CO2 out of a generator casing with air. Never. Ever. And, rarely do they use the portable gas analyzer to measure the concentration of CO2 in air (or air in CO2)--if a working portable gas analyzer could be found. I did see one site in South East Asia hook up the portable gas analyzer to the sampling port, but they didn't plug it in. Which means it was useless. The procedure just said to connect the portable gas analyzer, and that they did.

Safety regulations are being written these days by people with little or no idea of how actual processes are accomplished. I'm not saying we don't need safety regulations--I <b>AM</b> saying we need better training, and more training. Even more redundant training (regular "refresher" courses). I have rarely seen the people who write safety regulations and procedures go out and observe , audit or verify that the safety procedures were being used or that they were relevant and correct. Mostly, they write the procedures, expect people to blindly adhere to them, and then they get involved in the review and analysis of what happened and how after someone gets hurt when it's learned someone with little or training was not following the procedure. So, it's the individual's fault--for being sent to do a procedure without understanding the risks and being trained on the equipment and being expected to be safe just by reading and following a procedure without understanding it or the equipment or the process.

Sorry--but your question presumes a lot of things, as most written instructions from a certain OEM do as well. A thorough knowledge of and understanding of the systems and processes is required to be able to "interpret" OEM instructions and develop a site-specific procedure which is applicable and relevant and takes into account all the variables (like using the proper sampling ports when using a portable gas analyzer during purging and charging).
This site has 7FH2 generators. I actually entered a PAC case on this issue because I thought the GEK was wrong when it instructed customers to admit air in the CO2 header and vent the CO2/air mixture out of the top H2 header. Excerpt from GEK 10709e is below:

"Placing Air into Generator
The Removable Spool Pieces are designed to inhibit the introduction of air while there is hydrogen in the generator and provide a means of absolutely preventing the inflow of dangerous gas while personnel are inside the generator performing maintenance.

1. Air Can Have a High Flow Rate

Air is relatively inexpensive compared to H2 and CO2. Therefore, the generator and the purge procedure are not designed to minimize the quantity of air needed. So that there is no chance of mixing H2 and air together, air is introduced into the generator through the CO2 distribution pipe on the bottom of the generator interior. Thus the lighter air is below the heavier CO2, and the gasses thoroughly mix during purging. Since mixing cannot be avoided, a very high flow rate of air can be used."

I almost feel obligated to follow the GEK since this is about the closest thing that GE will provide for a procedure. PAC case verified that the instructions above are accurate, admit air into the CO2 header and vent mixture out of the H2 header. I just wanted to get your thoughts. Thanks.

By the way, I agree with your comments on training needs as well. Most of my career has been commercial nuclear and I have seen that a systematic approach to training and good procedures can help operators be successful. I will research the number of manifolds.

Your Danger Ranger (Safety Manager/Supervisor) is going to pretty much INSIST you follow the GEK. I haven't read any hydrogen-related GE instruction manuals lately, and I will try to look up the reference doc's I always used back when I was field engineer. There used to be four P&ID drawings for purging and charging in the Operations and Service Manual which I always used for instructing the site personnel during commissioning which had some notes on them (OEM and personal) which I will look for and reply with any comments.

I wouldn't ever expect the PAC to advise against anything written in the manual--unless it was quite obviously wrong. And, in this case--it's really a judgement call. (And remember: Good judgement comes from experience; experience comes from bad judgement.)

Is the portable gas analyzer used at your site when purging/charging? (It's kind of telling that you didn't mention it--and I did, a LOT.) And, if it gets used, where is connected (which sampling port) when purging CO2 with air?

In my personal view, if one wanted to use CO2 to purge the hydrogen in the header when purging CO2 with air prior to entry into the generator, purging <b>SLOWLY</b> for 10 minutes or so with air to push CO2 into the upper header, then switching to admit air into the upper header and vent CO2 out of the lower header would be a good idea.

Disclaimer: I've been wrong in the past, and I will be wrong again in the future. I was taught how to purge and charge by a very respected field engineer--back when field service personnel were actually engineers trained to think and analyze and act on their experience and judgement. I did always read the manuals, (study them, actually) but I always deferred to respected people with a demonstrated track record of good judgement and critical thinking skills. Admittedly, we didn't have the World Wide Web back then, and I've been doing some reading on the subject, especially after the comment above about the very low ignition energy requirements of hydrogen in air, and this is affecting my thinking. But, there still must be an ignition source. And, any hydrogen in the header must be a relatively small amount if purged properly with CO2. And, again, sampling in the proper location with the portable gas analyzer is key to knowing for certain what gases are or aren't present.

I'll write back with any information I can find. I don't remember the four P&IDs for hydrogen-cooled generator purging and charging being supplied with early F-class turbines (I copied my drawings and handed them out during commissioning--they were VERY useful).

Anyway, I'd be very curious if the portable gas analyzer is used at your site during purging and charging. (I think many others reading this thread would also be very interested to know.)
We use the installed eOne dual hydrogen control panel analyzers for monitoring gas purity online and during purge activities. We do not use a portable analyzer. The representative gas sample would be the same for our unit because the portable analyzer sample valve (always left closed) and the sample read by the eOne analyzers are from the same line. The valves are only inches apart so I would think all is well on that concern. We monitor gas flow across the analyzers to ensure we don't invalidate the readings as well. I'm told if flow is too high it could drop cell temp and voltage and that will cause readings to be off slightly. (we use P&IDs 190D5487)

I haven't seen the most recent eOne gas analyzers. Do they monitor for CO2 and air as well as hydrogen? (That's why the portable gas analyzers were required with the hydrogen control panels provided for decades--because the analyzers on the panel were ONLY capable of monitoring hydrogen purity--not for CO2 or air, as the portable gas analyzers were capable of.

I've been away from the new units for too long (imagine that--only a few years can make such a difference).

And, I can also imagine the valving on the control panels is improved, too.... (That's a presumption--<b>not</b> a statement.

I haven't yet found my hydrogen piping system P&IDs.... Guess I didn't label the boxes very well....)

Yes, the analyzers read percent CO2 in air and percent H2 in air. Our site only uses the install analyzers. We will probably continue on the "backwards" path and admit air in the CO2 header based on the GEK and PAC response. It has worked at least twice already at this site.

Thanks for your input.

If it helps, in a 9FB we emptying procedure generator, the sequence is as follows:
1-Purging H2 with CO2
Change curve analyzer, measure concentration H2 in CO2.
Keeping the pressure in the generator 10-15 psig
Stop injecting when H2 in CO2 concentration is less than 5%
Remove/stop oil seals.
Remove spool SP-CO2 / AIR
2-Purging CO2 with Air
Change curve analyzer, measure concentration CO2 in air.
Install spool vertical position SP-CO2 / AIR
Stop when the concentration of CO2 in AIR is less of 5%, with pressure in the generator 10 to 15 psig
Stop the air injection.


You make some very good points, which resonate with my experiences as an Operator, C&I Technician and now as a C&I Engineer, regarding Generator de-gas and re-gas operations.

My experience has shown that people do not give Hydrogen the respect it deserves. Everybody is aware and respects the dangers Natural Gas presents - great. But the same respect and appreciation is not granted to Hydrogen. This despite Hydrogen having a much wider explosive range (4% to 77% by volume)and being a Group IIC gas, it requires minimal ignition energy to ignite a flammable mixture. Often with devastating effects. Hydrogen does strike a chord with me personally as I had a near escape with a Hydrogen explosion once - an Electro-chlorination Plant - if I conducted my surveillances 10 minutes later, a Storage Tank Roof would have landed on my head. And I'm not joking.

On top of a very vigorous, detailed and thorough Operational Procedure, the following are pertinent points to consider for a de-gas and re-gas Operation:

(1) Maintenance of the Hydrogen Control Cabinet (HCC)- ensure all coalescer filters, dessicants are in good working condition, replace if required. Have sample lines been blowdown to ensure no presence of Seal Oil? Check to ensure solenoid valves are not passing - calibration gas may be diluting sample streams - I have seen this previously! Do not just rely on the solenoid valves for calibration and de-gas operations, use the manual valves as instructed.

(2) Calibration of the (HCC)- conduct a calibration of the HCC a week before the Outage to ensure it is at its optimum accuracy before the de-gas operation. All the gasses required should be available at this stage, so calibrate for all 3 measurements, not just H2. Much easier to do pre Outage, rather than someone question the accuracy of the HCC during a de-gas or re-gas, whilst the H2 CO2 and / or Air are tied up in an LOTO isolation as part of the procedure.

(3) Establishing and Maintaining Correct HCC Sample Flows - this one annoys me the most. A cursory glance at the start is not enough! They should be regularly monitored and adjusted by Operations during the process. I would suggest every 30 minutes. With changes to Generator Frame Pressure and the changes in characteristics of gas flowing through the needle valve of the VA flowmeters, the flows will not stay the same and no assumption should be made that they will stay the same!

(4) Admission rates of Gases - too many people try and be heroes and "roar" the gas in to speed up the process - completely counter-productive. A good GE TA once told me to imagine you are creating a blanket of CO2, moving it upwards or downwards in the Generator, trying not to separate it, in a steady, controlled rate. Rapid addition will only cause the gases to vortex and mix inside the generator, creating the possibility of an non-homogenous sample to the HCC. Also the rapid addition can promote short circuiting, not blanketing. I once saw one Operator roar in CO2 at such a rate, he emptied the CO2 Cylinder Bank of gas pressure quicker than the CO2 liquid could vaporise, freezing the Cylinders...

(5) Control of Generator Frame Pressure - fluctuations in frame pressure will disturb sample flows as mentioned previously. Needs to be keenly observed and not allowed to fluctuate wildly.

In my opinion, an Operator should be given adequate time to manage a de-gas and re-gas operation as a solitary task, and give it the time, observation and focus it deserves. Unfortunately, with restrictions on resources and even more pressurised shutdown and re-commissioning windows, Operators have to take on significant tasks to prepare Plant for restart, and struggle to allow full focus on a de-gas or re-gas operation. All whilst under time pressure and, depending on the number of Generating Units, keeping other Units operational!

Interesting. Your recommendations are put into practice.

Could you tell me what was the root cause of the Hydrogen explosion once - an Electro-chlorination Plant. We also have a system.
. In one tragic accident 3 workers died when a leak in a generator stator they were maintaining detonated unexpectedly.

The investigation uncovered a number of issues, but two that stay in my mind are:

1) most of the workers had a poor understanding of the behaviour of gas mixtures including density effects and Brownian diffusion,

2) instrumentation is not always simple and straight forward. In this case, decades ago, the instrument used for measuring gas concentrations used gas thermal conductivity as the sensing medium. Engineers and the surviving workers testifying to the investigators expressed the view that they thought the settings marked 100% Air, 100% CO2 and 100% H2 actually identified the gases involved in the operations and purging procedure involved. But they didn't.

There were two important other factors leading up to the accident. Firstly, the leak that the team were investigating had been particularly difficult to find. The team had to repeat their efforts some twenty times, so on this last occasion there was a degree of familiarization that may have led to some laxity. Second, just before the accident, the team stopped for lunch allowing a time interval of approximately one hour in the course of the purge process.

The work involved shutting down a large electrical generator that used a hydrogen cooled stator, where hydrogen was circulated around the fixed wiring of the generator in the generator casing. The first step in removing the hydrogen to allow safe access to the stator compartment was to displace the H2 with heavier CO2. The CO2 was introduced through a nozzle at the bottom of the compartment while the lighter H2 vented out through a nozzle at a high point. The gas exhausted was measured with an instrument with a reading that moved from 100% H2 to 100% CO2.

After this step, the CO2 was displaced by air introduced at the high point and the heavier CO2 exhausting through the bottom vent. The instrument would then indicate that the exhaust gas would change from 100% CO2 to 100% air. In addition, the procedure required the number of gas bottles used to be recorded to show sufficient gas used in the purging.

Now the reality was that while H2 is lighter, less dense, than air while CO2 is heavier, the gases do not completely stratify but diffuse over time due to Brownian molecular movement. Workers in the power station had observed elsewhere that released hydrogen would rise to the roof and ignite on lighting fixtures with a relatively harmless “pop”. Brownian movement however was not fully understood.

During the purge procedure some pockets of gas remained trapped. With time these pockets would diffuse, so that with sufficient volume of purge gas and time allowed all the hydrogen would effectively be removed. On the fateful day, however, the CO2 purge cycle was stopped during the lunch break once 100% CO2 was observed at the exit nozzle.

After the lunch break, the workers commenced the next step to replace the CO2 and waited till the exhaust air indicated 100% air. In fact, the chamber contained a mixture of H2 and CO2 with the same thermal conductivity as air, so the instrument saw the exhaust gas as air. In fact there was an explosive mixture plus enough CO2 to cause anoxia.

The first worker then entered the chamber, with a second man part way through the manhole and another just behind when the detonation occurred killing all three. It is thought that the first worker collapsed from anoxia breaking a leadlight and igniting the detonable mixture inside the chamber.

The investigation concluded that pockets of H2 remained when the lunch break was taken and these pockets diffused in the CO2. The resulting mixture of H2 and CO2 had a thermal conductivity approximately equal to that of air such that during the last purge cycle with air a reading of 100% air was reached prematurely when, in fact, there was H2, air and CO2 in a detonable concentration.