I need to know the procedure for calibrating SRV MKV HMI.

 

Hi there,

I am user of a Mark V Speedtronic system and not a user of HMI (maybe in the future). For calibrating the SRV, you first need to change the regulator from type 77 to a 43 in the io configurator (regulator nr 1), download the new iocfg file to the three processors and reset (by switching off/on the power) the three processors one by one.
Now you can manually calibrate the SRV in the user defined display/srv/gcv callibration screen. Note: Don't forget to change the regulator after calibration back to 77 type.

Good luck.

 

There's no difference in the procedure for calibrating the LVDT (Linear Variable Differential Transformer) feedback from the SRV (Stop-Ratio Valve) based on which operator interface is used--<I> or GE Mark V HMI Server.

What Frame size unit do you need to calibrate the SRV on?

What is prompting the need to calibrate the LVDT feedback from the SRV?

Is AutoCalibrate "available" and are you familiar with using it? For AutoCalibrate to work, there must be a file in the unit-specific directory (e.g., F:\UNIT1) called ACALIB.DAT; does this file exist on your GE Mark V HMI Server?

The reason AutoCalibrate is being mentioned is that it temporarily and automatically re-configures the regulator of the SRV from a pressure control loop with position feedback to a position control loop--which makes verification of the LVDT feedback after calibration possible. The SRV job during start-up, acceleration, and FSNL (Full Speed-No Load) operation is to control the pressure (called the "P2 pressure", or "interstage" gas pressure) between the SRV and the GCV (Gas Control Valve)--and it does this as a function of speed.

A P2 pressure reference is calculated during execution of the CSP (Control Sequence Program) and is supplied to the servo-valve output's regulator. The regulator compares the value of the P2 pressure feedback from the P2 pressure transducer(s) (96FG-2 or -2A, and -2B and -2C, if so equipped). If the feedback equals the reference, there is no error and the valve does not need to move. If the feedback does not equal the reference, the error is converted to a position error, and the valve is moved until the pressure feedback equals the reference--and the error is again zero and the valve is held at the position.

If one doesn't somehow re-configure the SRV servo-valve output's regulator, when a reference is given to the valve it looks for pressure feedback. If the gas fuel supply is isolated and bled down (as it should be when performing any checks of the SRV), there will be no pressure between the SRV and GCV when the SRV opens and it will just continue opening until it reaches the open mechanical stop.

There is usually an improperly named User Defined (or Demand) Display called Gas Valve Calibration. If one tries to use that to calibrate or manually position the SRV, it won't work because the regulator isn't reconfigured--the SRV will just wide open as described above.

So, please tell us a little bit about your unit and the requirement to calibrate the SRV LVDT feedback, and let us know if you're familiar with AutoCalibrate and if ACALIB.DAT is present in your unit-specific directory or not, and we'll do our best to produce a procedure for you.

markvguy

 

Frame 5 unit I need to calibrate the SRV.

I need to calibrate the LVDT feedback from the SRV
ACALIB.DAT is present in the unit.

 

the ACALIB.DAT is present in the unit.

 

For calibrate the SRV LVDT feedback, ACALIB.DAT is present in the unit-specific directory.

 

Hi markvguy;
You are advising to use " autocalibrate "
I am strongly advised not to use this.
It seems that there was a TIL ( technical information letter ) from GE about this a few years ago? On our Mark V `s I have remark this in runtime so it isn`t visible any more, so they can not longer use this.

Carlos

 

We also use the Autocal screen at Elwood, but use the "Enable Manual" calibrate to perform the calibration manually to get the LVDT 0 & 100% voltage readings for io configurator entry. Doing this manually allows us to check positons in midrange, although there is no adjustment if the valve / feedback is nonlinear. At this site Autocal can be found in HMI G:, EXEC file, AUTOCAL.EXE. This is true for our Windows NT & 2000 machines.

Please keep in mind the function of the SRV, as it is not to be taken lightly. It is an emergency stop valve and there are no process relief valves to lift if interstage fuel pressure goes to high. Depending on the make of the valve, if the "Close" mechanical stop collar is moved either way SRV may not seal when fuel gas pressure is restored. If this happens the valve will have to be removed from the process, disassembled and the plug ( Fisher V-Ball in our case) will have to centered or it can be shipped back to Moog. Once again this not an impossible task, but decompressing an actuator spring with wire as big as your thumb is potentially very dangerous without proper work practices. To avoid all of the above, if the interstage fuel gas vent line does not leak by when gas is lined up to the unit, the Close stop collar should not be moved.

I recommend calibrating the SRVs especially if it has not been performed since commisioning. I say this because we found 4 of the 9 7FA GTs on this site stopping at ~80 deg. (on the Open mechanical stop collar) instead of 90 deg. With that condition the valve will not be at its true 90% (watch degrees and % on SRV, the MKV uses 0 - 100% and the valve strokes from 0 - 90 degrees). The MKV issues an automatic shutdown at >90% SRV opening. This happens if there is an upstream gas pressure transient and the SRV goes further open to maintain the interstage fuel pressure setpoint.

Greg Ponto
Lead Combustion Turbine Specialist
Elwood Operations
(815) 423-9883 x2232
Email: [email protected]

 

Carlos,
If you have this infamous TIL (Technical Information Letter)--or if anyone reading this posting has this TIL--can you, or they, please post the number of the TIL on this forum? (TILs are numbered, usually a four-digit number like 1471, sometimes they have a revision number like 1480-R2. The TILs for the two examples cited would be formally identified as TIL-1471 and TIL-1480-R2, respectively.)

There is a widely circulated "wives' tale" (otherwise known as "gossip") that continues to persist in many parts of the world today that AutoCalibrate on the Mark V Speedtronic turbine control system doesn't work. It is patently (unmistakably) false. Like all software tools it has its limitations (which will be described below), but this author has used it literally hundreds of times to quickly and easily calibrate LVDT (Linear Variable Differential Transformer) feedback on all manner of devices on nearly every type of GE-design heavy duty gas turbine manufactured. There has been only one instance when it "failed" to work for this author.

It is this author's belief that the reason that this myth got started was because of a lack of understanding of how AutoCalibrate works--and what is required to make it work properly. For AutoCalibrate to work properly, the information in its "reference" file, ACALIB.DAT, MUST be accurate AND the device must be mechanically "sound" or "firm."

AutoCalibrate is nothing more than a method of automatically driving a servo-valve-operated device equipped with an LVDT or RVDT (Rotary Variable Differential Transformer) to one mechanical stop (usually the closed mechanical stop), waiting for the position feedback voltage to be stable, recording that position feedback voltage, then moving the device to the other mechanical stop (usually the open mechanical stop), waiting for the position feedback voltage to be stable, recording that position feedback voltage, calculating the zero- and 100%-stroke voltages (in other words, the "offset" and "gain", or "zero" and "span" voltages for 0%- and 100% stroke feedback), then returning the device to the first mechanical stop and checking to ensure the voltage feedback at that position is nearly identical to the voltage first measured.

For GE-design heavy-duty gas turbines, most device positions are expressed in percent--not actual physical length of stroke. 0% stroke for most valves (not the LFBV--Liquid Fuel Bypass Valve--since it's a bypass valve) is defined to be the position at which there is zero fuel flow, and 100% stroke is defined to be the physical position at which maximum expected fuel flow will occur (the fuel flow on the coldest possible day expected to be experienced on that site with the fuel characteristics supplied to GE during the requisition phase of purchasing the combustion turbine). In the case of the LFBV, 0% stroke is actually when the LFBV is full open, since this valve bypasses fuel from the discharge of the positive displacement high-pressure liquid fuel pump (back to its suction) and as the valve closes less fuel is recirculated back to the suction of the pump and that which is not is forced to the liquid fuel nozzles.

For many units manufactured in the last decade or so, the positions of the gas valves (SRVs (Stop-Ratio Valves), GCVs (Gas Control Valves), GSVs (Gas Splitter Valves), GTVs (Gas Transfer Valves), Primary Gas Control Valve, Secondary Gas Control Valve, Tertiary Gas Control Valve, and Quatenrary Gas Control Valve) were mechanically set to 0- and 100%. In other words, the strokes were mechanically limited to the required movement (1.250 inches, or 1.500 inches, or 2.250 inches, or 0-90 degrees--whatever length of stroke was specified by GE or provided by the vendor to match GE's flow specifications) such that zero flow occurred at the minimum mechanical stop and maximum expected flow occurred at the maximum mechanical stop.

Many units manufactured with GE's proprietary combined gas valve assemblies (cast gas valve assemblies containing the SRV and GCV in one assembly) had mechanical stroke lengths which were longer than that at which maximum expected fuel flow (the 100% stroke position) for a specific unit was achieved. In other words, the physical travel of the valve might be something like 1.860 inches, but maximum expected fuel flow, or the 100% stroke position, for that application would be achieved at 1.625 inches.

Also, most--not all--of these combined gas valve assemblies employed what is referred as "closed-end over-travel" meaning the valve stems could continue to move for approximately 0.030-0.050 inches AFTER THE VALVE PLUG WAS FULLY ON THE SEAT, or, fully closed. (This was to prevent the possibility that the actuator might not close fully and prevent the valve from completely closing. Since the LVDTs are attached to the valve stems, it's possible for them to indicate negative position--and this MUST be taken into account when calibrating these types of valves--either when using AutoCalibrate or a manual method.)

For these combined gas valve assemblies, the 100% stroke lengths were listed in the Control Specifications, Control System Settings, document in Section 5, Gas Fuel Control.

The valves whose 0- and 100% strokes are set with mechanical stops are relatively easy to calibrate; those whose 0- and 100% stroke positions do NOT occur at mechanical stops are much more difficult to calibrate.

For AutoCalibrate to work properly, it must be "told" what the physical positions of the devices are at the mechanical stops WITH RESPECT TO THE 0%- AND 100% STROKE POSITIONS SPECIFIED IN THE CONTROL SPECIFICATIONS. It gets this information from entries in ACALIB.DAT--its "reference" file.

There are "default" entries in ACALIB.DAT--usually -0.01 (that's right--a slightly negative number!?!?!?) for the minimum stroke value, and 100.0 for the maximum stroke value. If the mechanical stops are set at 0% and 100% of stroke length, these values work GE ("Good Enough"), or just fine. However, if the valve travels 1.860 inches to the open mechanical stop when 100% stroke length occurs at 1.625 inches, then the maximum stroke value must be changed to 114.462 ((1.860 / 1.6250) * 100 = 114.462) since the physical position of the device when it is at the full open mechanical stop is 114.462% (1.860 inches) of the 100% stroke position (1.625 inches).

If this information isn't determined BEFORE AutoCalibrate is executed and an AutoCalibrate is performed--and someone physically tries to measure 25%, 50%, 75% and 100% stroke positions based on the value in the Control Specification (1.625 inches), then the calibration will be WRONG! Is it AutoCalibrate's fault? NO! It has no idea what the physical stroke of the valve is with respect to the expected maximum fuel flows! It has to be "told" via ACALIB.DAT.

Further complicating the matter is the position of the IGVs (Inlet Guide Vanes) on GE-design heavy duty gas turbines: they don't go from 0 DGA (DeGrees Angle) to 100 DGA. Their mechanical travel is limited to approximately 32 degrees and 86-89 degrees. (In the Mk V, DGA and 'percent of stroke' have the same range: 0-128. The Mk V "knows" when DGA is specified for LVDT feedback to use the same range as devices calibrated in 'percent of stroke'; the specified engineering units (DGA) are used instead of PCT. The 0% stroke voltage for IGV LVDT feedback is therefore, negative--which also causes some people concern.)

The default entries for all eight servo-valve outputs in ACALIB.DAT are - 0.01 for the minimum stroke position and 100.00 for the maximum stroke position. If someone uses AutoCalibrate to calibrate the IGVs WITHOUT FIRST MEASURING THE IGV POSITION AT THE MINIMUM- AND MAXIMUM MECHANICAL STOPS AND INPUTTING THOSE VALUES INTO ACALIB.DAT, the calibration of IGV LVDT feedback will be HORRIBLY wrong! If someone enter 57.0 and 84.0 for IGV for the minimum- and maximum positions, this is ALSO wrong--because there is overtravel on both ends of the normal operating range of the IGVs.

For many machines, the values of mimimum- and maximum IGV mechanical stop positions listed in the Control Specification document are EXPECTED values--not actual values. That's why measurements must be taken to accurately determine exactly what the positions at the mechanical stops are--and then that information must be put into ACALIB.DAT. If one measures 31.75 degrees for the minimum mechanical stop and 86.50 degrees for the maximum mechanical stop, then one enters 31.75 for the minimum mechanical position and 86.50 for the maximum mechanical position of the proper servo-valve output into ACALIB.DAT before performing an AutoCalibration of IGV LVDT feedback.

Now, for the one time when AutoCalibrate has been known NOT to work. When retrofitting a GE-design Frame 6B unit with a Mk V turbine control system during a DLN (Dry Low NOx) upgrade, it was discovered that one of the pins in the IGV actuator linkage was severely worn, such that the IGVs would not return to the same mechanical position as when the AutoCalibration of IGV LVDT feedback was started. This has happened on other retrofits, as well, when the IGV actuators were found to be severely worn and needed to be replaced.

The same thing also happens occasionally on those gas valves which have closed-end overtravel for the same reason--the device does not return to the same position as when the AutoCalibration was started. In the case of these valves, one must insert something between the actuator rod end and the valve stem to make sure the device returns to the same position after calibration--feeler gauges work just fine for this!

The nice thing about using AutoCalibration for calibrating SRV LVDT feedback is that it AUTOMATICALLY changes the regulator type from a pressure control loop with position feedback to a position control loop during the AutoCalibration procedure! No need to change the value in the I/O Configurator, download, reboot, perform the calibration procedure, then have to change the value in the I/O Configurator back and download and reboot. To this author, it seems MORE dangerous to perform a manual calibration of SRV LVDT feedback for this reason--if the regulator type isn't changed back after the calibration, the SRV may go full open, which could be catastrophic, but usually just results in a turbine trip. This author has been to several sites which forgot to change the SRV regulator type back after performing a manual SRV calibration and the unit just kept tripping during starting.... A TIL should have been issued to warn owners and operators of the possibility of problems if AutoCalibrate is NOT used for calibrating SRV LVDT feedback!

The last thing about AutoCalibrate--which REALLY seems to bother most people--is that after the procedure is finished, one has to manually calculate the AVERAGE value of the 0% and 100% stroke voltages for the two LVDTs (there are usually two LVDTs per device) determined using AutoCalibrate, input those values into the I/O Configurator, and download them to the Mk V. Notice this author didn't say "download and reboot." That's because AutoCalibrate determines the EXACT 0% and 100% stroke voltages FOR EACH CONTROL PROCESSOR IN A TMR (Triple Modular Redundant) CONTROL PANEL--and if one doesn't re-boot the panel after an AutoCalibrate, those values remain in TCQA RAM and the accuracy of LVDT position feedback is the best it can ever be! If one calculates average 0% and 100% stroke voltages and downloads them to each of the three control processors in a TMR control panel and re-boots the processors, the position feedback will not be as accurate as immediately after performing an AutoCalibrate.

So, until and unless someone can produce the number of the TIL which allegedly says that AutoCalibrate isn't safe (which is the main function of TILs--to alert owners and operators to potentially unsafe conditions), it's just a myth.

markvguy

 

By the way, just "commenting out" (remarking) an entry in MENU.DAT does NOT remove it from the <I>--it only means it can't be run from the Main Menu when IDOS is running.

If one exits to a command (DOS) prompt, changes to the unit-specific directory, and types ACALIB and presses ENTER, the AutoCalibrate application will start (if ACALIB.DAT is present in the unit-specific directory, e.g., F:\UNIT1).

markvguy

 

markvguy,

I also believe that autocal is a good thing. But it can also be a nightmare in the hands of the wrong tech/person.

If the stops are not properly set on the valves, null bias settings have not been verified, things can get really nasty.

Try dropping two of the moog servo coil feedbacks by disconnecting them from two of the TMR cores. Autocal will pass the calibration and store the values.

I am just trying to impress that one must ensure that everything is in order before absently just saying it passed Autocal it must be OK.

Improper P2 pressures can cause a "Uncontrolled energy release" if procedures for gas valve calibration are not properly followed.

In a "nutshell". Be careful. I have encountered technicians that BELIEVE that Autocal can cure all ills without performing ANY verification of the valve, servo, or anything.

 

Hello,

WE have a FRame 6001 B , Control System Speedtronic MKV TMR <I>, Fuel: GAS. (no DLN)

I would like to know the procedure to execute an Autocalibrate on SRV and GCV. To achieve the hydraulic pressure in the circuit we have to run the turbine in crank mode. So apart from that are there signals that I have to Force in the Logic Forcing Display?. if so which are those signals? thanks very much in advance

 

It should go without saying, but be sure the gas fuel is manually isolated upstream of the SRV, and bleed down the gas fuel between the manual isolation valve and the SRV before beginning the procedure.

Once you have established hydraulic pressure you will need to establish Gas Fuel Trip Oil pressure. Since you will be cranking the unit, the 4s will be picked up so all you should need to do is force the Gas Fuel Trip Solenoid logic signal to a logic "1". The name of that signal is usually L20FG1X.

On some units, it's also necessary to energize the IGV Trip Solenoid, 20TV-1. Again, since you will be cranking the unit, I suspect that signal may already true; the signal name is usually L20TV1X.

You really need to review the Piping Schematics (P&IDs) for the Trip Oil system to see exactly what's required.

AutoCalibrate isn't exactly "automatic". All it does is drive the servo in the closed direction until the LVDT feedback isn't changing for some period of time, records that voltage feedback, then drives the servo in the open direction until the LVDT feedback isn't changing for some period of time, records that voltage, then drives the servo back to the closed position until the voltage isn't changing for some period of time (to make sure the device returns to the same zero stroke position), then calculates the zero and 100% stroke voltages to be entered into the I/O Configurator.

AutoCalibrate uses some information in F:\UNIT1\ACALIB.DAT to determine what position the device is in when it's at fully closed and fully open mechanical stops. Because, on some older combined gas valve assemblies, fully open mechanical stop is not always 100% stroke. The 100% stroke setting, which is usually less than fully open, is specified in Sect. 05.02.nn of the Control Specification. You must stroke the device while measuring the stroke with a dial indicator or caliper, and then determine the fully open stroke as a percentage of the specified 100% stroke value, and enter that into the appropriate NEG_SAT_POS field in ACALIB.DAT. When AutoCalibrate runs, it looks at the information in ACALIB.DAT and calculates the proper 0- and 100% stroke voltages.

Once AutoCalibrate has run, for those values to be stored in the EEPROM of the Mark V until the next calibration, you need to enter them (or their average values if the Mark V is TMR) into the I/O Configurator, then download the values to the Mark V <Q> processor(s). You don't need to re-boot at that point, just download them.

You should also verify that when you tell the device to go to 50% stroke, that it goes to 50% of the Control Spec. stroke.

Hope this helps.

 

Hello CSA, sure it helps a lot...great information

I want to perform an Autocalibrate on SRV and GCV just to verify existing calibration and check the mechanical system. So is good to know that I have to Force L20FX1 to 1, because I thought I had only to Force the permissive L3ADJ to 1 according to the permissive field on ACALIB.DAT. The Field on the Autocalibrate dispay for SRV Permissive: SPEED < 28 % (TNH) refers only to assure the Hydraulic Pression and no gas supply?

So as I just want to verify calibration, I don't have to Download any value to the MKV Processors. Is that right? thanks !

 

Wrong. The Mark V operates using values which are downloaded to RAM from EEPROM each time the processor is re-booted. In the case of LVDT feedback values, when an AutoCalibration is performed, the values in RAM are <b>directly</b> changed. <b>But</b>, the values in EEPROM <b>are NOT</b> changed by AutoCalibrate.

If you actually perform an AutoCalibrate procedure, the associated values in processor RAM <b>WILL CHANGE</b>! If you click on the "AutoCalibrate" button, it will run the servo/actuator through the calibration process and will physically affect the values used by the Mark V for that LVDT feedback.

The reason one downloads values to EEPROM is so that the values in EEPROM match the ones in RAM. That way, if the processor is re-booted it will get the same values as were modified in RAM by AutoCalibrate.

Performing an AutoCalibration <b>DOES NOT</b> allow one to <b>verify</b> an existing calibration. You must use the "manual" feature of AutoCalibrate to verify an existing calibration. You would need to be able to accurately measure the stroke of the device whose LVDT feedback you are trying to verify.

 

Could you please post the whole; accurate; end to finish; procedure one should follow for LVDT calibration. Instead of trying to understand just what you mean by all your side-track comments.
That would be much easier for everyone to follow.

Thanks.