IGV Does Not Reach Minimum Operating Angle After Offline Compressor Washing


Thread Starter


We have a Mark VI controlled Hitachi H-25 gas turbine unit (30 MW) at our combined cycle power plant. Today, after compressor offline washing or "cold washing" as it's referred to at the plant, had been completed, the IGV failed (stuck at 40 DGA) to reach its minimum operating angle of 26.9 DGA. The Operator was able to force the IGV to its minimum position with the hydraulic oil pump on manual.

The Operator tells me that the hydraulic oil pump apparently stopped right before the IGV had reached its minimum operating angle with the hydraulic oil pump on manual.

The fact that I was not physically present when this occurred, I could not verify a lot of the things Operations are saying. The turbine is now in operation and the IGV + hydraulic oil pump seem to be working fine. The IGV went from 26.9 DGA to 54 DGA and beyond during start-up and the hydraulic oil pump pressure is also within acceptable limits.

I have looked at the Hydraulic Oil Pump Control logic on Mark VI Control System Toolbox to try and come to a solution.

I looked at L4HQRUNZ (Main Hydraulic Oil Pump Control Enable), which I have identified as the main Hydraulic Oil Pump Control command. This signal needs to be high for the pump to work. According to the logic, the only way the pump can stop prematurely is if the lube oil pressure drops. Could you kindly verify if my assessment so far is valid? Is there possibly something else wrong here?

Thanks for all the help in advance.

I'm not familiar with Hitachi machines, but you seem to be saying that there is no Accessory Gear-driven Main Hydraulic Pump, only an AC motor-driven Hydraulic Pump? Is that correct?

In GE-design non-F-class heavy duty gas turbine Mark VI application code, the AC motor-driven Auxiliary Hydraulic Pump logic signal is usually L4HQZ, which when a logic "0" (zero) will start the Aux Hyd Pump. (Sometimes, L4HQZ is driven by a single NC contact from a rung named L4HQRUN, but that's usually only done on F-class units.) The "Z" at the end of the signal name that begins with "L4" is usually an indication that the signal controls a discrete output AND that when the signal is a logic "0" the device controlled by the output should be running. But, that's GE convention, and Companies that use GE control systems are not bound to use GE convention or control philosophy even though they might use the same or similar signal names.

GE-design heavy duty gas turbines usually have a hydraulic accumulator and because the flow-rates s in a GE-design hydraulic system are usually very low when the pump is shut off it may take as long as a minute for the hydraulic pressure to fall significantly when the unit is at CRANK speed or below.

IGVs generally have no spring to return them to the closed position; the actuators are double-acting pistons. So, you would be correct that in the absence of any hydraulic pressure the IGVs would (at Cranking speed or below) probably stay at the position they were at. (Not knowing how the IGVs are designed, it's hard to say if air flowing into the axial compressor would tend to open or close them in the absence of hydraulic pressure; most GE-design Frame 5 machines the IGVs would tend to open if the air flow across them were high enough and there was no hydraulic pressure).

Many newer GE-design heavy duty gas turbines also have a solenoid, 20TV-1, that when de-energized will cause hydraulic pressure to close the IGVs. Does the H25 have a similar solenoiid that serves the same function?

The reports from Operations seem not to be in line with expectations, but, Operators are not usually very familiar with any other operating conditions other than zero speed and rated speed, and even then they are extremely likely to be heard saying, "It's never done THAT before!" when it's been doing "that" for decades.... It's the rare Operator who knows what's supposed to happen and can say with certainty that it did or didn't happen. Very rare.

Without being able to see the application code running in the Mark VI at your site AND without being able to see the P&IDs for the machine at your site, it's really difficult to say much more. And even much of the above is speculation on my part because of my unfamiliarity with H-25 configuration and operation. Though they are osimilar in many respects to GE-design Frame 5 machines--they are also unique in many ways, so I hope I haven't added to the confusion.

Please write back to let us know what you discover!
Yes, we have two AC motor-drive hydraulic pumps in lead-lag configuration. We also have a hydraulic accumulator in place. I looked at the logic some more and as it turns out, the main hydraulic oil pump will only turn on at 44% TNH which would be 7258*0.44 = 3193.5 rpm in our case. According to Operations, the hydraulic oil pump stopped at about 1000 RPM. If this is indeed the case, it would make sense that the IGV stopped at 40 DGA. This assumes that the hydraulic oil pump works independently and that it only requires the lube oil pressure to be sufficient and the speed to be 44% TNH for it to turn on.

Another conclusion I can draw from this scenario is that perhaps the hydraulic oil pump is supposed to stop by the time 44% TNH is reached but this assumes that the IGV will have reached its minimum operating angle before that happens. Something I would like to point out is that the hydraulic oil pump is supposed to produce a pressure of 9.5 MPa but it only produces 9.22 MPa. Maybe because of the decreased pressure, it takes longer for the same hydraulic circuit to respond which is only apparent when the IGV has to close as opposed to open. So at 80% TNH, the IGV start to open and the little bit of inertia is compensated by the air flow. I would like to ask if this is valid.

We have a 20TV-1 dump valve but I think it is only de-energized upon tripping.

Another thing that certain senior Operations personnel seem to have is that the hydraulic oil pump operation is somehow timed. I don't think that is true and believe that the logic is condition based as opposed to timing based. The application code/logic seems to agree with me too. Could you kindly verify this as well in light of your experience? Perhaps I am missing something and the logic is indeed timed?

I would REALLY like to see the .m6b file for the Mark VI of the Hitachi H-25 to be able to comment better (and just because I'm curious, as well!).

Starting the lead hydraulic pump at 44% speed seems kind of a little late, because the unit should be firing by then.?.?.? And you would need hydraulic pump pressure to open the SRV/GCV of LFBV when fuel is admitted to the unit.

And, if there's a speed level used for starting the pump, then there is likely a deadband on the drop-out, of say 5%, or 1% or something like that.?.?.?

If there's a hydraulic accumulator, I would think with the low hydraulic flow during CRANKing that if the lead hydraulic pump shut down the accumulator would be capable of supplying the flow required to close the IGVs--but that's just a SWAG (Scientific Wild-Arsed Guess) based on experience with GE-design heavy duty gas turbines. When the pump shuts down on GE machines, it can take as much as a minute for the hydraulic system pressure to decrease from approximately 10 barg to 0 barg.

I think 20TV-1 has to be energized any time the IGVs are to be opened, and is de-energized when the IGVs are supposed to be closed. If the arrangement is similar to a GE-design Frame 5 heavy duty gas turbine, when 20TV-1 is energized it will permit hydraulic oil from the servo-valve to pass to/from the double-acting IGV hydraulic actuator to open and close it as necessary. And when 20TV-1 is de-energized, then hydraulic oil is ported to close the IGVs, regardless of the action of the servo-valve.

I would presume that with two AC motor-driven hydraulic pumps being operated in a lead-lag configuration with no gear-driven hydraulic pump that a low discharge pressure from the lead pump would start the lag pump. But, if the discharge pressure was normal when the lead pump was stopped the lag pump would not be started.

Hope this helps!
I did some more digging around and as it turns out, my entire approach was wrong. I managed to get my hands on the Compressor Cold Washing Procedure that Operations uses and kind of went on from there.

This is what the Cold Washing Sequence looks like:

-> Cold washing happens at 20% TNH.

-> It involves compressor washing for 10 minutes and then drying for 30 minutes, both on cranking mode.

-> Certain conditions have to be met and then the Offline Water Wash Start option will be available to the operator.

-> As soon as the Offline Water Wash Start button is pressed, the IGV reference goes to the max operating angle (KGVMAX) and it opens to about 86 DGA.

-> Water Wash Stop Valve (GWWSTV-1) opens for 10 minutes and the water is injected.

-> GWWSTV-1 is closed on auto and the IGV goes back to following the minimum possible reference angle.

-> Drying proceeds for 30 minutes.

-> The operator presses L1STOP (Master Control Stop) and a fired shutdown happens after drying is completed.

-> Fired shutdown will lead to L94T going High or 1 if : L94XZ = 1 after the fired shutdown timer expires OR no flame is detected.

-> In our case, since no flame was ever produced the turbine should trip right away which is what the event summary is saying.

-> So L94T goes High or 1 and L4 goes 0 and the Main Hydraulic Oil Pump turns off.

Conclusion I have come to based on the above scenario:

-> Because the Main Hydraulic Oil Pump is being operated at lower than normal pressure, the IGV does not close right away after Offline Water Wash Stop is pressed and L83BW_OFF becomes 0.

-> The recommended operating pressure is 9.5 MPa but the pump is being operated at 9.22 MPa.

-> The accumulator circuit should take over during high-flow requirements such as this. The IGV is huge and requires the greatest amount of flow. So I am also assuming there is some sort of passing in the accumulator circuit.

Some confusions:

-> Am I right to assume that the IGV should close right away after water wash is completed or is it going to take some time based on ambient conditions etc.?

You keep referring to the "Main Hydraulic Pump"--is that mechanically driven (from an Accessory Gear) or just the Lead pump of a lead-lag pair of AC motor-driven Aux. Hydr. Pumps? Most of the GE-design Frame 5/6 machines use a similar variable displacement, axial piston pump for both the Main (gear-driven) and Aux. Hydraulic pumps (same pumps; different drivers). And, when the gear-driven Main Hyd. Pump is operating when the unit is cranking (such as during washing), it's output is pretty close to rated. There is usually sufficient pressure/flow during cranking (on a gear-driven hydraulic pump) to power most of the control devices--especially if there's no fuel flowing. The pressure differential you cite (0.28 MPa) isn't enough to cause the problem you were <b><i>told</b></i> occurred.

As for the washing sequence, usually when performing an Off-Line Water Wash (sometimes called a "cold wash"--don't know why, but I have heard them called that), there is a washing "cycle" (where, while cranking and with the IGVs at maximum operating angle (LK83GVMAX) detergent and water are admitted through the wash control valve for usually somewhere around 20-30 minutes, or until the desired amount of detergent has been used. Then, most procedures call for a "soak", and the unit is shut down to ratchet/turning gear to let the detergent do it's thing. During coastdown the IGVs are usually closed.

Then after a 20-30 minute soak (depends on how dirty the compressor was determined to be before the washing was started, and depending on the type of contaminants which may be present), the unit is put back on CRANK and plain water (demin water, usually) is re-admitted for the "rinse cycle." The IGVs are opened back to maximum operating angle, and water is admitted through the wash control valve until the drains are clear of soap and suds (this can take a LONG time depending on how much detergent was used, and IN GENERAL, too much detergent is usually used).

After the rinse is complete (clear water coming from all drains), the wash control valve is usually closed and the unit can be cranked for a few more minutes, or just STOPped. When the valves which were moved for the Off-Line Water Wash are returned back to their normal, running conditions, the unit is then either STARTed in FIRE or AUTO mode, to establish flame and dry out the combustion and turbine sections (especially if the unit isn't going to be put back on line within approximately 24 hours).

From your description, I don't see any indication of wash or soak or rinse cycle. But, whenever water is flowing through the water wash control valve the IGVs are usually opened to maximum operating angle. (Note that opening the IGVs increases the air flow through the axial compressor, which usually causes the shaft speed to decrease. Also, admitting water to the compressor while cranking can also cause the shaft speed to decrease slightly.)

And, while the IGV actuator for GE-design Frame 5/6-size machines is usually larger than the actuator for either the SRV/GCV or LFBV--but not that much larger. (Again, I don't have any personal experience with Hitachi H-25 machines.) You are correct in that the IGV actuator usually requires more flow to move from fully open to fully closed (or vice versa) but I would think a properly functioning hydraulic accumulator of the proper size could easily accommodate that flow even if there was no hydraulic pump pressure/flow, or the pressure/flow from a hydraulic pump was very low (much lower than you have cited).

Most hydraulic accumulators are the bladder type, charged with Nitrogen or some other inert gas, to approximately 50% of working system pressure. MANY hydraulic accumulators are not properly maintained, and not properly valved in, and many are not properly charged, and even if they are charged to the proper pressure unless proper care is used when disconnecting the charging hose the pressure in the bladder can easily be lost. Sometimes, the charging hose isn't even properly fully valved-in during checking/charging. They aren't difficult devices, but they do require some familiarization and training to operate properly.

As for how long it should take for the IGVs to close, a second or two is probably maximum when CRANKing--if Hitachi used typical GE IGV control logic. The reference is switched from minimum to maximum during washing, and when washing is complete it's switched back from maximum to minimum. Sometimes there's a ramp on how fast the output can change, but I've never seen anything that controls how fast the output changes as a function of ambient temperature.

Finally, I'm still not clear on the Main Hydraulic pump comments versus the lead-lag Aux. Hydraulic Pumps.
Thanks for your comments so far. I cannot even begin to describe how helpful they have been especially since this is my first foray into the world of power generation.

The Main Hydraulic Pump I have been talking about for the course of our discussion is indeed the lead pump of a lead-lag pair of AC motor driven pumps. We don't have an accessory gear driven main pump. We only have the two pumps I mentioned earlier.

Cold wash happens exactly as you mention. A timer goes on for an hour and the operator is free to go about the sequence as he pleases within that time-frame. He has to fit the washing, the rinsing and the drying etc. in that time-frame. I was just going over the steps in a broad sense. However, the issue occurs towards the very end when the IGV does not fully close after the operator presses L1STOP and the turbine coasts down. It begins closing but it gets stuck at 40 DGA.

The operator does the washing (the soap, the rinse etc.), presses Offline Water Wash Stop, the water wash stop valve closes (GWWSTV-1) and drying proceeds for 30 minutes. After drying is done, the operator presses L1STOP, L4 drops out and the turbine coasts down. The IGV does not fully close as the turbine is coasting down. There is even an IGV Position Servo Trouble Alarm and IGV Control Trouble Trip after the cranking motor stops which leads to L20TV de-energizing but the IGV still does not close fully and the operator has to start the lag pump on manual to make it close all the way.

Possible Cause:
- There is blockage along the return line or drain line of the IGV. Gonna have the team look at the dump valve (see attachment below).
- Low hydraulic system pressure. Gonna have the team look at the accumulator circuit and will have them adjust the pressure of the pumps to about 9.5 MPa just to be on the safe side.

I have two more questions.

According to the application code, the IGV should go back to following the part-speed reference as soon as the operator presses Offline Water Wash Stop on the HMI because LGVMAX becomes 0. This is something I find most confusing because the turbine is on cranking or 20% TNH and the fact that the IGV only starts to modulate open when turbine speed reaches > 80% TNH, shouldn't it begin closing before drying can happen? Wouldn't this affect drying since very little air is flowing in?

- http://pho.to/AEUk0 << the dump valve is circled in green.

Hmmm.... I would really like to see the .m6b file, but if I post my email address here three things happen (none of them good). First, a LOT of people will contact me directly for help with their issues instead of posting to control.com. While I don't mind helping people, I'd rather do it on control.com because MANY people can benefit from helping one person instead of just the person who wrote directly. That's the beauty of control.com--many people read these threads when they're first posted and follow them, and many people use the 'Search' feature to find similar issues or information on a subject they want to know more about. So, instead of just helping one person, we get to help a lot of people. Second, people write asking for a job--and that's not something I can help anyone with over an Internet forum without the benefit of having worked alongside them and being familiar with their experience and drive.

Third, one gets a LOT of spam email from posting their email address directly to forums such as control.com--A LOT!!!

Anyway, since you've found out how to post a graphic image to a web hosting site, perhaps you can post the .m6b file to one and then post a link to it here and I can download it. I know this is not the easiest way to communicate, but it's what needs doing at this point (since we can't post files/graphics to control.com--yet; we keep hoping).

There are two (2) dump valves in the P&ID you posted--and there are two hydraulic accumulators. 20TV-1 drives both dump valves, and when 20TV-1 is in the de-energized position (as shown on the drawing) and the dump valves are in the de-actuated positions as shown on the drawing hydraulic pressure/flow is ported to the IGV actuator to CLOSE it, and the oil on the other side of the IGV actuator's double-acting piston is ported through the other dump valve to drain. There's NO flow through the servo-valve in the positions shown--when 20TV-1 is de-energized, and the dump valves are in the de-actuated position. 20TV-1 has to be energized to allow hydraulic pressure to flow through the servo-valve to the IGV actuator to position the IGVs to the reference.

When 20TV-1 is energized, the "shuttle" in the solenoid valve moves to the left--blocking the flow or oil to drain through 20TV-1. This builds up the pressure on the dump valves causing the shuttles in the dump valves to shift to the left allowing hydraulic pressure/flow through the servo-valve to the IGV actuator.

So, as long as there's sufficient pressure/flow, the IGVs should move--either when the servo-valve is being commanded to go back to the normal IGV reference after washing is completed, or when 20TV-1 is de-energized. As long as there' sufficient hydraulic pressure/flow.

If the Lead hydraulic pump motor is being de-energized when L4 is dropping out, AND the hydraulic accumulators aren't working properly then it's entirely possible there is insufficient pressure/flow to move the IGVs back to the normal IGV control reference.

During low-speed ("part speed") operation of GE-design Frame 5/6 heavy duty gas turbines the normal IGV reference is 34 DGA (DeGrees Angle). The closed-end mechanical stop is approximately 32 DGA, which is slightly below the minimum operating position (34 DGA). As you've rightly noted, the IGVs don't typically start moving until the unit approaches 80% TNH, then on GE-design Frame 5/6 heavy duty gas turbines they ramp open to 57 DGA (sometimes less if the unit has DLN combustors). But, from START to approximately 80% speed the normal IGV reference is 34 DGA. Drying can happen with the IGVs open or closed--as long as the shaft is spinning and there's air flowing (even if the IGVs are at 34 DGA there is still air flowing). So, part-speed reference is fine. Drying will probably take a little longer with the IGVs ar 34 DGA, but it will still occur.

My money is on something being amiss with the accumulator<b>S</b>--either the bladder(s) is/are blown, or the valves are not in the proper positions (the "block" valve should always be open (NO--normally open), and the "bleed" valve should always be closed (NC-normally closed). Or, the charge has been lost in one or both of the bladders.

I presume the Lead hydraulic pump is shut down when the STOP is initiated, but I would also find it difficult to believe that there is so much flow in the hydraulic system when cranking that the accumulators can't provide sufficient pressure/flow to close the IGVs. Again, there are two accumulators--one seemingly simply for the IGVs, with a 2.4 litre capacity (I believe that's the hydraulic oil volume).

I also didn't ask if this has always been happening, or if it's something that just started happening.?.?.? If it's always been happening, that's one thing. If it just started happening, then something's changed--and the trick with all troubleshooting when something unexpected starts happening like this is to determine what's changed that would cause this to happen.

I want to point out a couple more things about the P&ID and typical plant adjustment of hydraulic pressure. The hydraulic pumps have internal pressure compensators which try to keep the pressure in the system at rated regardless of the speed of the pump (which is fixed because these pumps are driven by AC induction motors). And, because the pumps are positive displacement pumps, there is a relief valve downstream of each pump so that if the internal pressure compensator fails the relief valve will relieve the pressure and protect the hydraulic system components. MOST sites use the relief valve to set/adjust the system pressure--<b><i>WRONG.</i></b> The internal hydraulic pump pressure compensator is to be used to set the system pressure. The relief valve setting is to be HIGHER than the system pressure.

The proper way to set the two pressures (internal pump compensator pressure and relief valve pressure) is to run the one of the pumps, then adjust the internal pump pressure compensator to raise the pump pressure to make the relief valve operate. If the relief valve starts relieving too soon (at too low a pressure), it needs to be adjusted, or if the relief valve relieves too late (at too high a pressure) it needs to be adjusted. Once the relief valve is set, then readjust the internal pump pressure compensator to the rated hydraulic system pressure. Shut down the pump, and do the same with the other pump and it's relief valve.

If the relief valve is relieving too soon (if it's being used to set the system pressure) then there's too much flow through the relief valve (there should be NO flow through the relief valve unless the internal pump pressure compensator has failed, or is misadjusted). Usually, that causes the current drawn by the AC motor-driven pump to be higher than it should be, which can cause the thermal overload of the pump's motor starter to actuate (alarm/trip). And, most sites just re-adjust the thermal overload setting to make the alarm go away....

Again, the internal pump pressure compensator (usually accessible by removing an acorn nut on the outside of the pump housing) is to be used to set/adjust the hydraulic system pressure to the system rated pressure--NOT the relief valve.

You've said the IGV actuator is "huge" so it might need a high volume--but do you think it would take more than 2.4 litres? And, remember, there's another, second, accumulator, with a 5 litre capacity....

Please continue to write back with your findings--it's much appreciated!
firstly pls tell me; you do only one cycle for Offline (cold) water wash?

isn't the full cycle is:
1. 10 min detergent cycle
2. 30 min soaking
3. 4-5 rinsing 10 min cycles until clear water is available in drain
4. drying for 30 mins

if not i think you are risking detergent presence on your compressor blades.

we have H25 (25MW) hitachi machines. we can give water wash start command when m/c is in cranking; at that time IGV gradually opens; then water CV opens.

similarly during stop command CV closes and IGV closes slowly;.then you can give crank stop command at will depending on its soaking or drying.