Yesterday, we attempted to start one of our GTs, everything was fine from 0% TNH to 95% TNH during the startup. As we are starving for adequate gas pressure and due to the national power crisis, it has been a long time that we are accustomed to starting our GTs with a very low gas pressure. Yesterday it was around 12.7 bar!! It was noted that there were no abnormalities in the MarkV Speedtronic panel, even there were no serious diagnostic or process alarms and all cores were in A7 status. GT took around half an hour to reach 98% TNH from 95% TNH but could not reach FSNL. For more clarity, our DLN GTs are equipped with SRV, GCV and GSV. SRV remains 100% open due to low upstream gas pressure while GT is in operation. Finally we had to stop the GT as the speed was saturated at 95-98% TNH. At this moment, we don't understand what could be the cause of this abnormal behavior. Did anyone face such an event then please share your thoughts and suggestions to understand the cause and find a solution to this problem?

 

What is the design P2 gas pressure? If the SRV was all the way open before even reaching rated speed, that is your problem.

Did you have any exhaust / backup or other "temperature control curve active" alarms? These would stop the GT from increasing the fuel.

 

@SaberMPS,

As @MattyIce says if the SRV is wide open before reaching FSNL this could be the problem.

You didn't say what Frame size machine you are working on.

What is the gas fuel supply pressure before you initiate a START?

You didn't say what position the GCV is at during the failure to reach FSNL when the machine is at 95% TNH.

There is usually a Y-strainer upstream of the SRV and if it's plugged/choked/dirty that could make low fuel supply pressure problems worse for the machine. Has anyone checked the Y-strainer to see if it's dirty/needs cleaning? (It's usually just a "rock catcher) but I've seen large pieces of coalescing filters, bolts, nuts, washers, rags (of many different varieties), duct tape and rope in the Y-strainer. (It's amazing what the gas transmission companies leave in the pipeline when they are performing maintenance or repairs.

And, fuel supply is not just about pressure--it's also about flow (see the above). Generally if there isn't enough flow the pressure will drop, but sometimes there is sufficient flow available but the pressure is too low. If there are gas filters upstream of the SRV and they are dirty that could make the low fuel supply pressure problems worse. Under this condition, even if you reached FSNL and were able to synchronize the machine, you probably wouldn't be able to raise the load very much, if at all.

12.7 barg is about 185 psig. The gas fuel pressure downstream of the SRV (and GCV) has to be greater than axial compressor discharge pressure (CPD) in order for gas fuel to enter the combustors. If I remember correctly, the machine is on Start-Up FSR until it reaches FSNL at which point the start sequence is complete from a fuel standpoint. My best GUESS is that there is some start-up Control Constant that is limiting the GCV. (You could look at the FSR bargraph display to check this most easily, I think.) BUT, even if that's what is causing the machine not to reach FSNL, and you manage to synchronize it's highly unlikely that you will be able to load the machine very much if at all--and if the gas fuel supply pressure is fluctuating/oscillating the load will also fluctuate/oscillate.

Without knowing what Frame size machine you are working on it's difficult to say how much fuel gas supply pressure (and flow!) you would need to keep the SRV from reaching 100% open during starting.

We also don't know what kind of gas fuel you are burning. It's not uncommon for gas fuel make-up (methane; BTU; etc.) to vary as fuel gas pipelines switch suppliers/fields. A low quality gas fuel could also be contributing to the problem.

Some machines with DLN-2 combustion systems need a pretty significant gas fuel supply pressure during starting because they transition combustion modes before reaching FSNL. You mentioned a GSV (Gas Splitter Valve) so it's most likely your machines use DLN-I, not DLN-2.

 

Sorry; I didn't read the title before I clicked on 'Post reply'. My bad.

So you have a GE-design Frame 6B heavy duty gas turbine, which means it will have a DLN-I combustion system and because it doesn't have a GTV (Gas Transfer Valve) it is a transfer-less DLN-I system.

But, it's MOST likely the fuel flow-rate/supply pressure that's causing the problems. And, even if you DO manage to get synchronized, it's not likely you will be able to carry much load (produce MW) unless the gas fuel pressure/flow magically increases when the generator breaker closes.

My experience with Frame 6B machines is that P2 pressure must be above about 14-15 barg to transfer combustion modes reliably AND the SRV will most likely be at about 30-50% stroke at Part Load. This means the gas fuel supply pressure must be above 14-15 barg, more like 17-19 barg.

Anyway, we still need information. (See the bold text above.)

By the way, finding and changing that start-up Control Constant that limits fuel flow during a START up to FSNL is NOT recommended. It's there to protect the machine from excessive fuel flow during a START when the gas fuel supply pressure is at or near normal. Even if you did get the machine to make FSNL, if the gas fuel supply pressure/flow is still low the machine WON'T produce much load (unless the gas fuel supply pressure/flow magically increases after the generator breaker closes...).

 

Thank you very much for your valuable contributions and I am sorry for being late.

@MattyIce

​

During startup at 20% TNH the SRV opens and at 87% of TNH it's position becomes 100%!!! Because of low gas pressure (12.5 to 14 bar).

@WTF?

It is frame 6B machine. You are right, it might happen from insufficient flow of supplied gas.

What is the gas fuel supply pressure before you initiate a START? around 12.5-14 bar.

You didn't say what position the GCV is at during the failure to reach FSNL when the machine is at 95% TNH. I can't tell you exactly how much GCV position was, sorry for that. But as per shift engineer "It was 56% open!! Later on 18th May, 2024 we had a successful start, before that start we maintained a little higher gas pressure than the normal practice. Everything was smooth.

Has anyone checked the Y-strainer to see if it's dirty/needs cleaning? Strainers are new.

We also don't know what kind of gas fuel you are burning. We use natural gas where Methane percentage is 95-96.

A view file of a successful start is uploaded herewith for your reference.

 

SaberMPS,

I can tell you with with a LOT of experience commissioning new machines that the amount of trash found in y-strainers during initial operation is very substantial. Especially if the gas fuel supply line was new to the machine, and even more especially if the new connection was done with a "hot tap" method. There can be LOTS of trash and debris in the pipeline leading up to the y-strainer if it was not properly cleaned after the construction--and construction crews ALWAYS believe they leave the piping clean. (HAH!!!)

Anyway, the information you are providing isn't very good or very complete. Get some more gas pressure is all we can say.

 

The 9E unit on our site had the same problem twice in history. The speed stopped rising after reaching 95%. When it happened for the first time, we consulted an expert, who told us to stop the unit completely, wait until the speed reached zero, and then restart the unit. We stopped the unit according to the expert's method, waited until the speed reached zero, and restarted it. The unit started normally. A few years later, this situation occurred again. The same method was used as before and it started normally.
I don’t know whether the problem in this case has been solved. What is the solution?

 

The main reason of the issue has been identified; however, it remains unsolvable. Insufficient gas pressure prevents further acceleration from 95% to FSNL. We encountered this issue in May 2024 with one of our GTs, but in the last few days, we have faced the same thing happening with our remaining three machines one by one!! At startup, if the pressure is lower than 188–190 psi (GE Frame 6B), it hovers about 95% of TNH and doesn't accelerate any further. The area, where our power plant is located, there is a severe natural gas pressure crisis.

I appreciate all of your insightful comments and experience sharing.

 

Lots of rather strange information. I've seen many Frame 6Bs start with natural gas supply pressures of around 190 psig and successfully reach FSNL. HOWEVER the y-strainers and any filters upstream of the y-strainer were all clean--verified clean (as in the natural gas lines had been blown down (to atmosphere) and THEN the strainers (and filters if any) were all pulled before reassembling the piping to send clean fuel to the machine through the y-strainer (there is almost ALWAYS a y-strainer upstream of the SRV (Stop-Ratio Valve).

At light-off, the P2 pressure being maintained by the SRV upstream of the GCV should be around 30-40 psig--regardless of natural gas supply pressure! Typical natural gas supply pressure for a GE_design Frame 6B with conventional combustors is usually around 300 psig, and the SRV reduces the supply pressure to around 30-40 psig during firing (that's a rough estimate). To do this the SRV is usually open only a few percent of full stroke if the natural gas supply pressure is at or near rated. If the natural gas supply pressure is less than rated the SRV would have to open further for the P2 pressure to equal the P2 pressure reference. But--there is a limit to the opening of the SRV during firing, and it's usually about 30% stroke. And the alarm that is usually annunciated if the SRV stroke is above about 30% is START-UP GAS FUEL FLOW EXCESSIVE and the machine is usually tripped on this condition (SRV open more than about 30% during firing). The Mark* doesn't actually look at the flow-rate of natural gas during starting--the presumption is that if the natural gas supply pressure is at or near normal (or even at about 65-75% of normal) AND the SRV is open to more than about 30% of stroke then the gas fuel flow-rate WOULD BE excessive.

A few of the sites I commissioned did not have rated natural gas supply pressure (and flow) during commissioning and, of course, EVERYONE is anxious to see the machine start and run up to FSNL. So, often I was pressured (no pun intended) to start the machine with less than rated natural gas supply pressure just to see the machine run up to FSNL to make all the managers and owners happy. More often than not if there was sufficient flow even with low supply pressure the machine would make it to FSNL, but there were a couple of sites that, while they had about 50-60% of rated natural gas supply pressure there wasn't sufficient flow to reach FSNL.

It's NOT JUST pressure--it's also flow-rate. If there's some reason that as flow-rate increases during starting and acceleration that the pressure starts dropping as speed increases then there's often something restricting the flow of natural gas such that even a low supply pressure can't be maintained. And, that something is usually something like natural gas compressor(s) that have yet to be commissioned, or plugged strainers/filters (upstream of the y-filter) or a dirty y-strainer, or a valve in the supply piping that isn't fully open, or simply there's not enough flow from the natural gas supplier to maintain even a low supply pressure. So, you can monitor the supply pressure during starting and if it starts dropping from about 190 psig as the machine speed increases then it's a good bet there's some kind of flow restriction OR the natural gas supplier simply does not have the gas (yet) that they should have by this time in the job.

You mentioned this problem started with one machine and has "spread" to others. In my mind, that says there's either a flow restriction somewhere in the natural gas supply piping (valve(s) not fully open; plugged filters and/or strainers) or the natural gas supplier has some serious problems with the source of the natural gas (too much consumption for the available supply; problems with their supply system piping/valving; problems with pressure regulators; problems with natural gas compressors; etc.). Without being able to see a P&ID for the natural gas supply piping it sure seems like if three machines were operating properly, and then one at a time each of them started developing the same issue then a natural gas supply piping problem (including valving and filters/strainers) ON SITE could be the problem. BUT, if the natural gas supplier admits to having problems with natural gas availability and pressure/flow then there's not too much that can be done, because even if the machines reach FSNL and can be synchronized they won't be able to provide much power (because MW is directly related to gas fuel flow (at rated pressure)).

The data in the compressed file looks an awful lot like Mark* V data--not exactly, but a lot. AND, there's no fuel pressures or valve LVDT feedback values (FPG2; FPRG; FPG1 or FPG3; FSRSR or FSRGR; FSG in the file. Or fuel flow-rate (usually signal name FQG).

AND, this thread kind of sounds like machines trying to be started with natural gas supply problems. Again, in my experience, with about 190 psig a GE-design Frame 6B with conventional combustors should be able to reach 100% of rated speed (FSNL) without too much trouble. AS LONG AS THERE'S SUFFICIENT FLOW "BEHIND" THE SUPPLY PIPING TO THE MACHINE(S).

HOWEVER, even if the machine reaches FSNL, synchronizing and then loading the machine WILL NOT be possible until the flow- and pressure problems are resolved.

Anyway, best of luck in securing rated natural gas pressure and -flow. Because, again, even if you reach FSNL without more pressure/flow the machine won't produce rated power without more natural gas pressure and flow.

The typical GCV position for a machine at FSNL is around 20%--meaning that it requires about 20% of the rated fuel flow (at Base Load on an iso day) in order for the machine to achieve and maintain FSNL (100% of rated speed). The typical GCV position at Base Load on an iso day is around 70-75% (when the natural gas supply pressure is at or near rated with no flow restrictions). The difference in GCV opening between FSNL and Base Load is 50-55% (approx. 75% minus approximately 20%)--and that means that Base Load requires 2.5 time more fuel flow than FSNL requires. And THAT is with natural gas supply pressure at or near rated with no natural gas flow restrictions, or minimal restrictions at worst. I'm trying to offer proof of why, even if the machine reaches FSNL and can even be synchronized--that unless there is near full natural gas supply pressure with minimal flow restrictions the machine simply will not be able to produce anywhere close to rated power. Power is, again, directly related to fuel flow, and in the case of natural gas that includes flow-rate AND pressure.

But, also again, if this problem has affected running machines one at a time on the site and the natural gas supplier hasn't admitted to being unable to provide normal rated pressure and flow then it's time to start looking at on-site piping/issues (such as faulty pressure regulators (pressure reducing valves; valves not in the proper positions; plugged natural gas filters and/or strainers). If a natural gas supplier is having problems maintaining pressure/flow it's entirely possible that there are mechanical impediments causing the problems (if it's not excessive consumption by all the loads supplied by the natural gas supplier). Which could mean sand, dirt, silica, rust, weld slag, scarf, rags, etc., causing problems and eventually making their way to site to cause problems.

Anyway, that's all I got for you. Hope everything works out soon.

 

I'm pretty sure there is a TIL out (forget the number) that also should trip the unit if the SRV goes too far open during startup like yours did. If whatever was causing the restriction upstream broke loose during the startup with the SRV at 100%, there will be damage.

 

I have always believed that this issue is not just about insufficient gas pressure. If it were due to insufficient gas pressure, then why do different units, and even different models, all maintain their speed at 95% rather than at other speeds? There must be some logical signal that is restricting the speed command from increasing further。

For example, at our site, this has happened twice. Each time, the unit when the speed does not increase beyond 95%, and then we shut down until the speed reaches zero. After restarting the unit, it can normally reach FSNL. During this process, the gas pressure did not change.

 

@DGY,

What turbine control system does the machine at your site use?

While almost anything is possible if the conditions are right, stopping at precisely 95% speed during starting/acceleration is unusual to say the least. Usually the machine is accelerating because the IGVs have opened and many machines close the compressor bleed valves at L14HS--95% speed, which means more air is now flowing into and through the machine which improves the efficiency of the machine and there is more mass flow through the turbine section being converted to torque.

Also, with most GE Mark* turbine control systems beginning with early versions of the Mark* V the acceleration rate of the machine is a function of an acceleration reference (signal name TNHAR). BUT, as always, there is the limiting factor of BOTH exhaust temperature (FSRT) AND an acceleration FSR limit (if I recall correctly the Control Constant name is something like FSKSU_AR). After the IGVs start to open, however, the exhaust temperature begins to drop AND that should cause FSRT to NOT be a limiting factor. Also, the axial compressor gets more efficient as the machine speed approaches rated (FSNL) which also helps to drop the exhaust temperature because of the increased air flow through the turbine section and exhaust area so FSRT doesn't limit the fuel flow-rate.

There have been sites which have tweaked the machine acceleration constants to get to FSNL quicker--but that can have serious knock-on effects, some of which include combustion hardware and turbine section mechanical damage, and even HRSG damage (if the machine exhausts into a boiler).

But, we don't know which turbine control system is used on the machine at your site. We also don't know if the machine has conventional or DLN-I combustors, and we don't know if the machine has an air-cooled or hydrogen-cooled generator. AND, we really don't know precisely when the machine "stalls"--meaning at what speed does the machine stop accelerating, and does it's speed decrease or remain steady. I don't recall if L14HS caused the L60BOG sequencing to be disabled or not, but if the machine is accelerating slowly and having trouble reaching 95% speed (taking longer than normal, in some cases, much longer than normal) and L60BOG is disabled when L14HS is reached then that could also contribute to problems. (And when the L60BOG sequencing is active and detects the machine is losing speed at a serious rate then the machine will actually be tripped--during starting).

The Mark* IV and early versions of the Mark* V which didn't have a starting acceleration rate reference and which on rate occasions would not get much past 97- or 98%, or even past 99%. L14HS would pick up at 95% speed but there just wasn't enough "oomph" to get "over the hump" to 100.3% speed (FSNL). AND, several sites I visited had gas fuel supply issues (low supply pressure; inability to maintain flow AND pressure during starting) and a couple had very badly calibrated gas control valve LVDTs. A couple of sites had switched natural gas suppliers and the new gas make-up was not as good as the old gas was (lower BTU/methane content) and the machines just couldn't quite get to 95% speed without adjusting the start-up control constants. (Sometimes this happens even when the natural gas supplier doesn't change; new sources of natural gas sometimes just aren't as good as the original sources so this can happen over time--low-quality natural gas.)

The torque converter is also pretty critical to the acceleration rate of the machine since it is providing a lot of torque during starting up to around 60% speed. Torque converters do require periodic maintenance and refurbishment (contrary to popular belief).

So, there are a lot of factors--including the temperature of the machine during starting. Sometimes, when the machine is cold and the ambient temperature is cold and the density of the air flowing into and through the compressor is higher than when it's warm outside that can also affect how much fuel (and torque converter assistance) is required to accelerate the machine all the way to FSNL (past L14HS). And, sometimes (particularly for machines with DLN-I combustion sytsems) when a machine doesn't make L14HS or FSNL and the machine is stopped and re-started it will "magically" make FSNL as if nothing had happened (this is true of some combustion mode change-overs on DLN-I machines, also). Machine and/or ambient temperature can make a difference, especially if the start-up Control Constants are marginal (and especially if the natural gas quality is/has been decreasing over time).

Finally, axial compressor cleanliness (or "dirtiness") also has an effect on starting reliability. An extremely dirty compressor (including dirty turbine inlet air filters which have a high differential pressure) means not as much air will be flowing through the machine--at any speed--which can impact starting efficiency/reliability, exhaust temperature (causing FSRT to limit fuel during acceleration to keep TTXM in check).

Personally, I've never found any Mark* logic that wasn't working properly when accelerating to FSNL--not to say it can't happen or hasn't happened. But, in my experience there would most likely be other circumstances and alarms which might help to explain the problem with accelerating to FSNL (or past L14HS).

 

@WFT
We used the Mark V HMI control system on site, the gas turbine was GE PG9171E, the machine did not have a DLN burner, and the generator was air-cooled. The unit started up normally until it reached 95% speed, when the speed stopped rising. At this time, the speed could basically remain stable, but it could not reach FSNL. No new alarms were found at the time. Since there had been a similar experience before, we arranged to shut down the unit until the speed was zero, and then restarted the unit, which was able to reach FSNL smoothly after starting.

 

@DGY,

I'm not poo-pooing your experiences. Okay; you have your "tried and true" method for solving the problem at your site. That's good. But everything you've written has been "reported" data, probably from people watching HMI screens (which often don't update faster than once-per-second). We don't know how old the Mark* V is on your machine (does it have DCC/LCC cards, or SLCC/SDCC cards in the control and <C> processors?). We don't know what alarms were present BEFORE and DURING the initial, failed START attempts--and we don't know, with any degree of accuracy because there does not appear to be any recorded data from them, what precisely was happening with speed and speed levels (L14Hx signals) and SRV and GCV positions and P2 pressures, etc. It's all anecdotal information--someone "saw" this or that and attributed the problem to this or that.

I've looked at some Mark* V sequencing (from a 6B and a 9E) and there MIGHT be a possibility that if the machine hit 95% speed and then dropped back to 94.5% speed and then went again almost up to 95% speed that the RAISE/LOWER logic MIGHT get "confused" and not allow a RAISE. The Mark* V wouldn't respond to RAISEs or LOWERs if this happened--manual ones or automatic ones. (By the way, one can't give manual RAISEs or LOWERs until AFTER the machine speed causes L14HS to go to logic "1", or after L14HS goes back to logic "0" during a STOP; the Mark* blocks those to protect the machine from getting stuck at a critical vibration during STARTing or STOPping.) And, giving the machine a STOP would probably un-confuse the sequencing. But it doesn't fully explain why, on a subsequent START the machine hit 95% speed and continued increasing speed without dropping below 94.5% speed. It would be necessary to see actual recorded data of relevant control database signals at a relatively fast rate of speed (faster than once-per-second, such as 8- or 16- or 32 times per second) ALONG with a history of alarms--PROCESS and Diagnostic--to begin to analyze the issue.

As I wrote before, early Mark* Vs DID NOT use an acceleration speed reference to control acceleration during starting. They used a stepped increase in L70R which wasn't always the best way to accelerate the machine, especially when ambient conditions change during the course of a year (especially when winter conditions can be less than 32 deg C (0 deg C) and summer temperatures can be above 90 DEG F (32 D deg C).

There's not a lot of difference between 94.5%- and 95% speed--and if a machine is struggling for some reason to achieve even 95% speed (and is it 95.01% or 95.1%--I've never thought to check that with a frequency generator during commissioning or to ask one of the design engineers at the OEM) it's also possible that the speed may have actually caused L14HS to go to a logic "1" and then if it dipped below 94.5% caused L14HS to go back to logic "0". But anecdotal data isn't good enough to say for sure--only recorded "actionable" data at a high rate is going to tell anyone for sure.

AGAIN, I'm not saying you're wrong. I am saying you are reporting anecdotal data which can't be analyzed. And, your remedy works at your site when the problem occurs.

I'm also saying that even if the logic got "confused" (because the machine went above 95% and then momentarily dipped below 95% speed--which we don't know for certain) there's STILL another reason (likely a mechanical reason--presuming the Mark* V Control Constants and sequencing are standard) why the machine "stalled" at, or around, 95% speed and even if the Mark* was trying to raise speed this other reason was conspiring against an actual fuel flow increase which would have caused machine speed to increase to FSNL. And, I'm also saying that if the Mark* V is an early version of Mark* V and does not have the acceleration reference sequencing/logic that those machines did suffer the occasional odd sequence of events, and very often, tripping or STOPping the machine and simply re-starting it resulted in a successful start and synchronization and loading.

I will add that on at least two sites I visited or supported via phone/email the original conditions of problems of inability to get past 95% speed to FSNL reported were that the machine wouldn't make FSNL and that they'd tried STOPping and STARTing the machine and it still wouldn't make FSNL--implying that NOTHING else was done the Mark* or the machine configuration between the two START attempts. On at least two sites it did finally come out that someone had tried to re-calibrate the SRV and/or the GCV and had sufficiently mucked up the LVDT calibrations that the machine wouldn't make FSNL three or four times, and in fact, one machine flamed out during firing/acceleration more than once. A proper SRV and/or GCV LVDT calibration solved the problems at both sites. AND, one of the sites was attempting to re-start after a maintenance outage when the natural gas piping upstream of the machine had been purged with nitrogen, and then the nitrogen wasn't sufficiently purged with natural gas prior to the first failed START attempt. (The piping run from the isolation valve to the SRV was more than 150 meters, and it was only purged for about 30 seconds, though a 1.5-inch vent line--not nearly enough to remove all the nitrogen and fill it with natural gas.) And, as usual, the blame was placed on the Mark*.... Incorrectly, I will add.