GEH-6810

GE Energy GEH-6810

OpFlex* Enhanced Transient Stability (ETS)for GE Gas TurbinesUser Guide

These instructions do not purport to cover all details or variations in equipment, nor to provide for every possible contingencyto be met during installation, operation, and maintenance. The information is supplied for informational purposes only, andGE makes no warranty as to the accuracy of the information included herein. Changes, modifications, and/or improvements toequipment and specifications are made periodically and these changes may or may not be reflected herein. It is understoodthat GE may make changes, modifications, or improvements to the equipment referenced herein or to the document itself atany time. This document is intended for trained personnel familiar with the GE products referenced herein.

GE may have patents or pending patent applications covering subject matter in this document. The furnishing of this documentdoes not provide any license whatsoever to any of these patents.

This document contains proprietary information of General Electric Company, USA and is furnished to its customer solely toassist that customer in the installation, testing, operation, and/or maintenance of the equipment described. This documentshall not be reproduced in whole or in part nor shall its contents be disclosed to any third party without the written approval ofGE Energy.

GE provides the following document and the information included therein as is and without warranty of any kind, expressed orimplied, including but not limited to any implied statutory warranty of merchantability or fitness for particular purpose.

If further assistance or technical information is desired, contact the nearest GE Sales or Service Office, or an authorizedGE Sales Representative.

© 2011 General Electric Company, USA. All rights reserved.Issued: 2011-05-06

* Trademark of General Electric CompanyWindows is a trademark of Microsoft Corporation.

Readers Comments To:GE EnergyDocumentation Design, Rm. 2971501 Roanoke Blvd.Salem, VA 24153–6492 USAFax: 1–540–387–8651(GE Internal DC *278–8651)

We welcome comments and suggestions to make this publication more useful.

Your Name Today’s Date If needed, how can we contactyou?

Your Company’s Name and Address Job Site Fax No ........................................Phone No ...................................

GE Requisition No. E-mail .........................................

Publication No. AddressYour Job Function / How You Use This PublicationPublication Issue/RevisionDate

General Rating

Excellent Good Fair Poor Additional Comments

Contents ...............................................Organization ...............................................Technical Accuracy ...............................................Clarity ...............................................Completeness ...............................................Drawings / Figures ...............................................Tables ...............................................Referencing ...............................................Readability ...............................................

Specific Suggestions (Corrections, information that could be expanded on, and such.)Page No. Comments

............................................................................................. ..................................................................................................

............................................................................................. ..................................................................................................

............................................................................................. ..................................................................................................

............................................................................................. ..................................................................................................

.............................................................................................. ..................................................................................................

.............................................................................................. ..................................................................................................

Other Comments (What you like, what could be added, how to improve, and such) .....................................................................

.....................................................................................................................................................................................................

.....................................................................................................................................................................................................

......................................................................................................................................................................................................

......................................................................................................................................................................................................

......................................................................................................................................................................................................

......................................................................................................................................................................................................

Overall grade (Compared to publications from other manufacturers of similar products, how do you rate this publication?)

Superior Comparable Inferior Do not know Comments .............................................Detach and fax or mail.

....................................................................Fold here and close with staple or tape....................................................................

Placestamphere_______________________

______________________________________________

GE EnergyDocumentation Design, Rm. 2971501 Roanoke Blvd.Salem, VA 24153–6492 USA

........................................................................................Fold here first........................................................................................

Safety Symbol Legend

Warning

Indicates a procedure, condition, or statement that, if not strictlyobserved, could result in personal injury or death.

Caution

Indicates a procedure, condition, or statement that, if not strictlyobserved, could result in damage to or destruction of equipment.

Attention

Indicates a procedure, condition, or statement that should bestrictly followed to improve these applications.

GEH-6810 User Guide 5

Warning

To prevent personal injury or damage to equipment, follow allGE safety procedures, LOTO, and site safety procedures asindicated by EHS.

Warning

This equipment contains a potential hazard of electric shock,burn, or death. Only personnel who are adequately trained andthoroughly familiar with the equipment and the instructionsshould install, operate, or maintain this equipment.

Warning

Isolation of test equipment from the equipment under testpresents potential electrical hazards. If the test equipmentcannot be grounded to the equipment under test, the testequipment’s case must be shielded to prevent contact bypersonnel. To minimize hazard of electrical shock or burn,approved grounding practices and procedures must be strictlyfollowed.

Warning

To prevent personal injury or equipment damage caused byequipment malfunction, only adequately trained personnelshould modify any programmable machine.

Warning

Always ensure that applicable standards and regulations arefollowed and only properly certified equipment is used as acritical component of a safety system. Never assume that theHMI or the operator will close a safety critical control loop.

6 OpFlex Enhanced Transient Stability (ETS) for GE Gas Turbines

ContentsOverview ....................................................................................................................................................... 9

Model-based Control (MBC) – Direct Boundary Control .................................................................................. 9Adaptive Real-time Engine Simulation (ARES) .............................................................................................. 10Control Mode ........................................................................................................................................... 10Parameter Boundaries ................................................................................................................................ 11Enhanced Transient Stability (ETS) .............................................................................................................. 11

Improved Transient (Grid) Response ................................................................................................................... 12

Model-based Coordinated Air-fuel (MBCAF) ................................................................................................. 12Grid Frequency Filter (GFF) ....................................................................................................................... 12Flame Anchoring Stability (Transient Split Bias) ............................................................................................. 14

Input Signal Processing (ISP) ............................................................................................................................. 15

Protective Actions ..................................................................................................................................... 16Sensor Models .......................................................................................................................................... 17

Human-machine Interface (HMI) Screens ............................................................................................................. 18

MBC Sensor Data ..................................................................................................................................... 18MBC Sensor Data Specific Details................................................................................................................ 19MBC Sensor Training ................................................................................................................................ 22MBC Sensor Tuning .................................................................................................................................. 24Combustor Hardware Selection.................................................................................................................... 25

Cycle Reference Parameters............................................................................................................................... 26

Combustion Reference (CRT)...................................................................................................................... 26Turbine Reference (TRT) ............................................................................................................................ 26

Alarms and Unit Response................................................................................................................................. 27

Glossary of Terms ............................................................................................................................................ 36


Notes


OverviewThe 2000s witnessed a boom in combined cycle gas turbine power plants. This trendhas been large enough to significantly impact the generating mix in many countries andfundamentally shift the dynamics of grid operation and dispatch. One outcome of thisis that a combined cycle gas turbine plant is now one of the easiest generation assets tomanipulate. The modern combined cycle power plant is often expected to start and stopmultiple times a week, as well as respond to changing load demands multiple times anhour.

The goal of Enhanced Transient Stability (ETS) is to increase the robustness of the DryLow NOx (DLN)-based gas turbine. GE Energy has re-written the core control softwareof the gas turbine using a Model-based Control (MBC) - Direct Boundary Controlapproach, referred to as MBC technology. This technology improves our control accuracyand capability.

Model-based Control (MBC) – Direct Boundary ControlThe intent of MBC - Direct Boundary Control is to identify operational parameters (suchas exhaust temperature, firing temperature, and emissions) of the physical system andcreate a control loop specific to each parameter to regulate. This ensures that the turbineas a whole, as well as the individual components, is always operating within the intendeddesign space. The Direct Boundary Control concept removes the inherent coupling thatcomes from legacy control methods, such as exhaust temperature control. Instead, gasturbine actuators or effectors such as fuel, air (inlet guide vanes [IGV]), inlet bleed heat(IBH), and fuel splits may be operated independently to provide a more flexible controlsolution with greater ability for optimization.

Effector Coupling

The ARES model is based onthe engineering cycle deck.

In practice, many gas turbine boundaries are often parameters that are not directlymeasured or even measurable (such as firing temperature). To overcome this limitationvarious boundary models are used. The goal of the models is to estimate the behavior ofthe system, based on known physics, to the level of fidelity required for the application.


Adaptive Real-time Engine Simulation (ARES)ARES is a high fidelity model of the gas turbine, continuously tuned in real-time tomatch the performance of the actual gas turbine. This model is derived from the GasTurbine Performance (GTP) Model application and coded to run real-time in the gasturbine controller. In order to make the steady state cycle model function transiently inthe controller, both a heat soak model and filter were added to supplement the basic cyclecalculations. Together they use the existing gas turbine sensors to tune the ARES model tomatch the actual operating conditions of a unit at any given moment.

Refer to GEH-6740,Model-based Control forGE Gas Turbines, for furtherinformation on the AutoTunesystem.

The ARES model is a key enabler in order to execute the Direct Boundary Controlphilosophy. As previously stated, many parameters that make up a component’s designspace are not readily measurable. The ARES model estimates many un-measurable cycleparameters with a high degree of accuracy that can be used directly in control loopsor as inputs to additional downstream sub-system models. The fidelity of the ARESmodel and any additional sub-system models are determined by the precision requiredin order to maintain the component in question within its design space. An example ofsub-system models that are enabled by ARES are the DLN transfer functions used forthe OpFlex*AutoTune* product. These DLN models would not be feasible without firsthaving the ARES model in place.

Control ModeIn the case of a gas turbine, many key parameters are affected by moving a single actuatoror effector. This requires the creation of a priority scheme, or control mode, for eachparameter that an actuator will affect. The typical GE Energy gas turbine continuouslycontrols approximately 20 parameters within the flange-to-flange turbine. The control ofthese parameters must be achieved with only four actuators: total fuel, IGV, IBH, andDLN fuel splits. The way this problem is overcome is by prioritizing certain controlparameters over others. The control mode is a hierarchy of control loops, with increasingpriority to the right.

Note The following figure is for reference only and does not represent an actual design.

Example Control Mode for the IGV Actuator

Each input to the control mode is an independent control loop that is controlling oneparameter. Whichever loop actively makes it though the control mode gate to determinethe command to the actuator is said to be the loop in control (LIC).

In some cases, multiple actuators can control the same parameter. For example, eitherIGVs or total fuel flow could be changed to impact the exhaust temperature. This allowsthe parameter to continue to be controlled even when one or more of the actuators aresaturated (unable to respond further).


Parameter BoundariesEach loop in the control mode must have a boundary to use as the control loop reference.These parameter boundaries can be a constant, such as the rotor torque limit, or complexmulti-variable schedules, such as the compressor operating limit line. Typical gas turbinecycle boundaries include (but are not limited to):

• Hot gas path durability (firing temperature)

• Exhaust frame durability (exhaust temperature)

• Compressor surge

• Compressor icing

• Compressor aero-mechanical limits

• Compressor clearances

• Compressor discharge temperature

• Valve pressure ratio

• DLN boundaries

Enhanced Transient Stability (ETS)With the fundamental philosophy of Direct Boundary Control and the ARES model inplace, the decision was made to structure the software into two separate areas:

• Control of the gas turbine cycle – bulk fuel/air control

• Control of the DLN system – DLN split control

The control structure for the gas turbine cycle is ETS and the control structure for the DLNsystem is AutoTune. This document primarily explains ETS.

The startup control scheme usesthe same logic as the legacypart-speed control logic.

ARES is currently designed for use only when connected to the grid at operating pointsabove full speed no load (FSNL). ETS requires ARES to operate; therefore a separatecontrol scheme, referred to as startup control, is used during turbine startup or shutdown.Startup control consists of all part-speed operation (generator breaker open), and includesall control loops and commands that do not use the ARES model.

Startup/Cycle Control Mode Selection


Improved Transient (Grid) ResponseThe ETS product was designed to improve the transient response of GE Energy gasturbines. It accomplishes this objective by using three main methods:

• Maintenance of global fuel/air ratio through coordinated air-fuel control

• Regulation of fuel response (fuel stroke reference [FSR]) by filtering the speed inputto the load governor and controlling fuel response to rapid transients

• Increase of transient lean blowout (LBO) margin, which is accomplished throughtransient DLN fuel split biasing.

Model-based Coordinated Air-fuel (MBCAF)The global fuel-air ratio isthe total fuel entering thecombustor divided by the totalairflow entering the combustor.

The modern DLN combustor only remains operable over a small window of stoichiometricratios. If the ratio is too high, the combustor will experience high combustion dynamicsand NOx emissions. If the ratio is too low, the combustor will flame out or produceexcessive CO. The goal of the Coordinated Air-Fuel (CAF) control is to maintainthe global fuel-air mixture (or stoichiometric ratio) delivered to the combustor in anoperable range. The CAF control typically uses IGVs as its actuator. Therefore, theCAF regulates airflow into the compressor in response to sensed or demanded fuel flowinto the combustor.

The MBCAF control improvesthe transient capability of thegas turbine by adjusting air andfuel flow rates simultaneously.

The basic idea behind the MBCAF control is to create a model of an ideal IGV-to-FSRrelationship (also known as the CAF Map), and to then use that modeled relationship tocontrol IGVs in response to a fast FSR motion instead of the nominal exhaust temperaturefeedback loop. The MBCAF intends to impact IGV control only when FSR is movingfaster than the normal IGV control loop can follow. The target of the MBCAF issignificant grid events, when FSR can load/unload the unit at a rate that can exceed 10 to15 times the nominal loading rate.

Grid Frequency Filter (GFF)Gas turbine robustness to LBO during abrupt frequency disturbances can be a concern,particularly in the emission compliant modes of a DLN combustor. Any change in gridfrequency causes a speed error, and invokes a response in which the speed-based fuelcommand is modified. Rapid changes in commanded fuel flow are not necessarily pairedwith well-coordinated changes in airflow, potentially leading the combustor to a conditionin which it is operating either too rich or too lean. In addition, grid requirements do notcurrently require the kind of rapid fuel flow changes that can occur during grid eventswhen no filtering is applied to sensed speed.

To address this condition, a speed/frequency filter called the Grid Frequency Filter (GFF)is used to shelter the gas turbine from the full effects of extreme frequency disturbances.As grid speed changes dramatically, only a tolerable rate of the change is passed throughto the load governor to set the new fuel command. In effect, this limits the response ofthe engine during grid events to maneuvers which are more aligned with actual machinecapability, as well as only that response required by the relevant grid code(s).


The GFF design is based upon a self-imposed transient power response requirementaligned with the most stringent European grid codes. The assumed transient powerresponse requirement is defined as follows. If measured, the turbine output responseto a 1% (60 Hz) change in grid frequency ramped in over a 10 second period and thensustained for another 20 seconds is such that the power at the end of the 10 seconds haschanged by at least the power response (P) and is sustained for 20 seconds.

Transient Power Response Requirement (as measured)

The magnitude of the power response (P) is expressed as a percent (%) of rated outputand is scheduled as a function of the current gas turbine load (refer to the followingfigure). Holding each gas turbine to such a requirement is a more appropriate balancebetween responsiveness (supporting the grid) and precaution as to not call upon units torespond in a way that is beyond their transient operating capability where they may bemore vulnerable to LBO.

Power Response Requirement as a Function of Gas Turbine Load


The requirement is stated so that the turbine is expected to be most responsive while in theemission compliant mode of operation that is consistent with being at a dispatchable loadlevel. The requirement assumes that a unit operating below the minimum turndown point(outside of emission compliance) has no transient power response expectation tied to it.This is consistent with the fact that these units are most likely loading or unloading to orfrom the emission compliant modes as part of a startup or shutdown, and not being reliedupon to support the grid. If a unit operating just above the turndown point is faced with apositive change in grid frequency, it will be called upon to shed load, but the rate will beless than the maximum and adjusted as load changes as to discourage an actual transferout of the combustion mode. Similarly, a unit operating at base load that is faced with anegative change in grid frequency will not respond as it cannot pick up any more loadfrom the base loaded point without incurring a higher maintenance factor.

Flame Anchoring Stability (Transient Split Bias)The transient DLN split bias function temporarily adjusts pre-determined fuel circuits bypre-determined amounts to ensure sufficient LBO margin during fast transients. Theamount of split bias given to a fuel circuit is calculated differently depending uponwhether the unit is running in AutoTune or not. If the unit is not running in AutoTune, thefuel splits are biased by a constant percentage during every application of split biasing.If the unit is running in AutoTune, the split biases are calculated in real time to ensuresufficient LBO margin while limiting total split levels in an effort to minimize the impacton combustion system dynamics and emissions.


Input Signal Processing (ISP)The accuracy of the ARES model relative to the actual operating turbine is extremelyimportant. If ARES believes a parameter to be one value when in reality it is somethingelse, the control system will have no knowledge that it is in error. The result can be brokenhardware or reduced component life. The accuracy of the ARES model is dependenton the accuracy of the gas turbine input sensors. It is therefore more important that thesensors are kept operational and in good calibration for MBC than for a non-MBC basedcontrol scheme.

Recognizing this potential weakness, a new input signal processing (ISP) functionwas developed for MBC. The ISP function provides fault detection, isolation, andaccommodation (FDIA) for each analog sensor input that is critical to maintaining theaccuracy of ARES across the load envelope. It also initiates appropriate control systemactions based on input sensor status. The sensor measurements monitored by the ISPfunction are those inputs which have the greatest impact on gas turbine operationalparameters across the load and ambient envelope, both those estimated by ARES as wellas standard parameters such as exhaust temperature. A representative list of sensormeasurements in the scope of the ISP function is as follows:

• Ambient pressure

• Inlet dew point temperature

• Inlet bleed heat upstream pressure

• Inlet bleed heat downstream or differential pressure

• Compressor discharge pressure


• Compressor inlet temperature

• Generator power

• Gas fuel pressure

• Gas fuel system differential pressures for PM1, PM2, PM3

• Gas fuel flow

• Gas fuel temperature

• Liquid fuel water injection flow

• 9th stage compressor extraction pressure

• 13th stage compressor extraction pressure


The ISP function uses statistical techniques to provide a complete solution to inputsignal processing diagnostics – out-of-range and in-range fault detection, faulted channelisolation and measured parameter accommodation for single, dual, and triple-redundantsensors. The algorithm is able to distinguish between the following fault types:

• No fault

• Availability fault

• Spike fault

• Shift fault

• Stuck fault (low noise)

• Noise fault (high noise)

• Disagreement fault (able to be isolated to a specific channel)

• Drift fault

• Redundant channel differential (not able to be isolated to a specific channel)

Fault detection is based on specific fault mode confidence calculations. Specificconfidences are combined to determine overall channel confidences and classification offaults, if they exist. The instantaneous channel confidences are combined with recenthistorical health information to derive a final confidence value for each sensor. Lastly, theaccommodation takes into account all system information to decide how to combine eachof the sensor readings to obtain a final output value for the measured parameter, which isused by all downstream control functions.

Protective ActionsThe sensor failures are aggregated from all monitored sensors. Based upon apre-determined protective matrix, the ISP takes the appropriate actions to protect the gasturbine. The following is a representative list of the protective actions that can be taken bythe ISP logic:

• Start inhibit (a start permissive)

• Use a model/surrogate in place of a failed sensor set

• Slew out of ETS (step to spinning reserve)

• Slew out of AutoTune

• Disable liquid fuel water injection

• Fail the inlet bleed heat (IBH) system open

• Disable the IBH DLN turndown schedule (raise the minimum IGV angle)

• Load reject to full speed no load (FSNL)

• Fired shutdown

• Trip

• Fail degraded operation


The application of the ISP strategy brings with it the benefit of more flexibility in theautomated protective actions of the gas turbine when station instrumentation fails. Thismeans that when certain sensors fail, the unit may still operate at a reduced output levelrather than causing the unit to trip. Fail degraded is an operational mode used for situationswhen the impact of sensor failures on key gas turbine operational boundaries has beenquantified, and thus can be conservatively accommodated in the parameter boundaries.The new operating state depends on the specific sensor failure or set of failures. Thefail degraded concept was introduced to maintain power generation, while potentiallyavoiding more severe consequences of failures, such as an automatic shutdown or trip.The magnitude of potential gas turbine derate is indicated by the fail degraded level, on ascale of 1-10, with higher numbers being more severe. The scale is relative and does notindicate a specific impact to the gas turbine, as this can vary with operating condition.

Sensor ModelsAn integral piece of the ISP is a generic tool set of sensor-specific models that may beused to provide additional virtual sensor readings to assist in unit operation and control.As previously stated, sensor models are used to increase the range of protective actionsavailable to ISP, and to assist in fault isolation. The sensor models are all physics-basedmodels, and many are tuned on a machine-to-machine basis, either automatically inreal-time, or at unit commissioning. A representative list of sensor models included withthe ISP function is as follows:

• Ambient pressure

• Inlet dew point temperature

• Inlet bleed heat flow

• Compressor discharge pressure


• Compressor inlet temperature

• Generator power

• Gas fuel flow

• Liquid fuel flow

• Liquid fuel water injection flow

Each sensor model provides an indication of its validity, as well as an alarm for faultedconditions. The validity logical indicates when the model should and should not be used.Some sensor models are expected to not be valid at certain times, for example, there areARES-based models that cannot be valid when the main ARES model is not valid. Inthese cases the alarm is masked, but the model output is not used by the ISP.


Human-machine Interface (HMI) ScreensThree screens have been added to the HMI to facilitate sensor training and tuning aswell as communicate the enhanced level of sensor information provided by this updateto the operator. Details on these screens are provided in the figures in this section (usethe notes for guidance).

Note The screens in this section are illustrations for reference only; actual screensmay vary slightly.

MBC Sensor DataFeedbacks from the LVDTs(gas valves, SRV, IGV, and IBHpositions) are also displayed,but enhanced information is notavailable for these sensors.

This screen displays an overview of the entire gas turbine, including all applicable fuelstreams and inlet, with analog sensor readings displayed in their approximate physicallocation. All of the analog sensor sets with enhanced protection provided by this packagehave a faceplate that turns red if any problems are detected with that sensor set. Forexample, the following figure displays the compressor inlet temperature (CTIM) hasdetected a failure.

MBC Sensor Data HMI Screen


If at any time the user moves the cursor over a faceplate, the application code signal namedisplays on the faceplate as shown in the following figure.

Faceplate Application Code Signal Name Display

If the unit is operating in a fail-degraded mode, the MBC Sensor Data screen also displaysthe fail-degraded level at which the unit is operating. The figure, MBC Sensor Data HMIScreen, displays an example of the unit operating in fail-degraded mode Level 9. Thiselement disappears when the unit is not operating in fail-degraded mode.

MBC Sensor Data Specific DetailsBy clicking on any of the faceplates, the sensor data HMI screen displays sensor specificdetails. There are three possible popup screens that may be displayed depending on theredundancy of the sensor set: simplex (single sensor), duplex (dual-redundant), or triplex(triple-redundant).

Simplex Sensor Faceplate


Duplex Sensor Faceplate

Triplex Sensor Faceplate

Raw Sensor Values – On an individual channel basis, the faceplate displays each sensor’scurrent reading in analog and bar chart form. The bar chart range limits are determinedby each parameter’s engineering range limits, which are set by control constants inapplication code.

Output Selection – For the input parameter being examined (such as CPD and CTD), thedisplayed output value is the result of the input selection processing function. The outputselection is the value of the parameter used by the control system.


Selection Status – For the input parameter being examined, the method of output selectionis displayed on the faceplate. For example, a triplex sensor with all good input channelscalculates a median output. The selection status options are as follows:

• Median

• Weighted average of A & B

• Weighted average of A & C

• Weighted average of B & C

• Channel A

• Channel B

• Channel C

• Model

• Default Value

Confidence – On an individual channel basis, confidence displays on a scale of 0-1 howconfident the input signal processing is of that sensor’s reading. A confidence of zeroindicates a failure has been detected, while a confidence of one indicates a completelyhealthy sensor. This box turns red if a failure has occurred.

Long-term (LT) Confidence – On an individual channel basis, LT confidence displays ona scale of 0-1 how confident the input signal processing has been in that sensor’s readingover a period of approximately the past 24 hours, with greater emphasis on more recentsensor behavior. This box turns red if long-term confidence is very low.

Refer to the section Input SignalProcessing (ISP) for a list offaults

Fault Status – If a failure has been detected, fault status provides a best guess as to thefailure mode of that sensor. Also identifies when sensors have high spread. This box turnsred if a fault has been detected, and yellow if high spread is detected.

Fail Degraded Box – The fail degraded box displays if the unit is utilizing a sensor modelinput. As displayed in the previous figure, the unit is in Fail Degraded Mode Level 9 dueto a CTIM hardware set failure.


MBC Sensor TrainingAs described previously, one of the fault detection checks looks for abnormally highor low noise relative to a normal baseline. The baseline is established once during thecommissioning of the software, during which adjustments to the default sensor noiselevels are made. The constants are then stored in non-volatile random access memory(NOVRAM) within the gas turbine controller such that they are recalled even if thecontroller is powered down and returned to service. If failed sensors are not replaced inkind, training should be manually initiated to avoid unnecessary protective actions. Thisscreen is used to facilitate this tuning process.

Note Sensor training must be completed before loading the unit beyond spinningreserve for the first time.

Noise Training Initiation Button (On) – This is located in the center of the screen towardthe top in the Sensor Training box. Clicking this button performs training on all sensorsets that have been enabled and meet the necessary permissives. This button changes toblue in color for the duration of the training process.

Individual Sensor Set Training Status Boxes – These comprise the majority of thescreen. The training procedure can be enabled or disabled for an individual sensor set byselecting the Enable or Disable button for that sensor. Successful training automaticallysets the Disable button (displays in blue), but the user can enable noise training at a latertime by manually selecting the Enable button again. All sensor sets are set to enable(Enable button displays in blue) initially for convenience. Upon successful completionof sensor training the sensor displays Trained in green. Otherwise, Not Trained displaysin red.

Example: In the figure MBC Sensor Tuning HMI Screen only the CTIM sensor set hasbeen successfully trained. With the exception of TS2P and WQ, all sensor sets in theright column have met the necessary requirements for training (they display Permittedin green and Enable displays in blue). Clicking the On button would train all of themsimultaneously. In contrast, the sensor sets in the left column have all met the permissive(Permitted displays in green) but are disabled. The sensor sets associated with waterinjection and the Cooling Optimization Package (COP) have not met the permissives (NotPermitted displays in red) because the unit is consuming gas rather than liquid fuel and ata load below where COP may be initiated.


MBC Sensor Training HMI Screen


MBC Sensor TuningIn the event a sensor set measuring a critical input has failed, sensor models are used tosupport the unit in fail degraded operation. For simplicity, sensor models are referredto with an M suffix added to the signal name. As an example, the sensor model forcompressor inlet temperature which has signal name CTIM is CTIMM.

A subset of the sensor sets with model counterparts include automated tuning featuresfor their respective model. This is done to maximize its accuracy. The indications andpushbuttons required to utilize this functionality reside in their respective sensor tuningboxes located in the Sensor Model Tuning field in the upper right hand corner of the screenas displayed in the following figure. In general, the automated tuning process makespermanent adjustments to the sensor model calculations based on the hardware outputsat the time tuning is initiated. Sensor models should be tuned during commissioning aswell as after any hardware changes within the subsystem they reside as they do makeassumptions about subsystem components.

MBC Sensor Tuning HMI Screen

Outputs – The output of the sensor model is displayed in the field with the whitebackground, below its hardware counterpart displayed in the field with the greybackground. If the control logic detects a problem with a given sensor model, the whitefield containing its output displays Invalid. (Refer to the previous figure; inlet bleed heatflow [CQBH] displays Invalid.) Similarly, when a fault is detected within a sensor set, thetext in the grey field containing the hardware output changes from black to white. (Referto compressor discharge pressure [CPD] in the previous figure.) This designates that thehardware is not performing optimally but is still being used by the controller.


Faceplates – If the controller is using the information provided by the hardware, thefaceplate associated with a given signal remains grey. In the event all available hardwareintended to provide that signal has failed, the unit operates in a fail-degraded state usingoutput from the sensor model. This is indicated by the faceplate of the sensor modelchanging to red. (Refer to the compressor inlet temperature (CTIM) in the previousfigure.) The popup screen associated with each faceplate is available on this screen byclicking the faceplate on the Sensor Data screen.

Fail Degraded Box – Similar to the functionality on the Sensor Data screen, the faildegraded box displays if the unit is utilizing a sensor model input. As displayed in theprevious figure, the unit is in Fail Degraded Mode Level 9 due to a CTIM hardwareset failure.

Individual Sensor Set Training Status Boxes – If a sensor model has never been tuned,Not Tuned displays in red. In the previous figure both the water injection flow sensormodel (WQM) and the gas fuel flow sensor model (FQGM) have not been tuned. The firststep to tune is to verify the Permitted indication is green. This means all of the requiredpermissives specific to that particular sensor model have been met. This is the case for theinlet bleed heat flow sensor model (CQBH) in the previous figure. Tuning can then beinitiated by clicking the Enable button. This button remains blue in color for the durationof the tuning process which is different for each sensor model. Once the tuning processhas been completed, Commissioned displays in green on the respective sensor model. Forexample, in the previous figure, the compressor inlet temperature sensor model (CTIMM)and the inlet bleed heat flow sensor model (CQBHM) are commissioned.

Note Tuning can be repeated after commissioning if system hardware has beenchanged and/or a sensor model indicates it is invalid, and all of its input parametershave been verified to be in working order.

Combustor Hardware SelectionThe Combustor Hardware Selection screen allows the user to account for differentcombustor hardware configurations. This enables proper GT operation based on actualcombustion hardware installed.


Cycle Reference Parameters

Combustion Reference (CRT)The control variable Combustion Reference (CRT) is used to define combustion modetransfer points (including staying in emissions compliance) and fuel split schedules for theDLN system, replacing the functionality of the traditional TTRF1 signal. The CRT canalso be used as a boundary for control of specific gas turbine cycle effectors as requiredby given engine configuration. Depending on the plant requirements, the signal CRTcan also be used for scheduling the operation of additional plant equipment through thedistributed control system (DCS).

Encoded parameters areproprietary to GE Energy.

Many cycle parameters, including the CRT, are encoded; the value is not given inengineering units but rather in non-dimensional units. The encoded values still allow forfull evaluation and manipulation of gas turbine operation.

Turbine Reference (TRT)The control variable TRT is used to define proper or nominal turbine operation,predominantly at base load. It is an encoded value that allows for verification of predefinedturbine operation by the operators of the power plant to ensure that gas turbine operationis as expected. This encoded value is synonymous with previous usage of TTRF1 by gasturbine operators to verify that the unit is operating correctly on the exhaust temperaturecontrol curves, which are no longer in use with ETS. A correlation between a gas turbineinput parameter such as ambient temperature or compressor inlet temperature and turbinereference is provided to the customer at commissioning of the unit to be used to assessproper gas turbine operation.


Alarms and Unit ResponseAs part of the high-level protection strategy associated with MBC, alarms indicate variousfaults that have an impact on the system. It is important that these guidelines be followedto maintain the integrity and operability of MBC.

Attention

It is imperative that only trained personnel perform any of thefollowing actions, and that all site-wide safety procedures arefollowed.

Alarm List and Gas Turbine Response

Alarm Signal Fault Condition Controller Display / Action Recommended OperatorActions

L83CA_F_A • ARES Model Fails ARES DIAGNOSTIC FAULT -MBC DISABLED

Step the unit to spinning reserve.

Allow at least five minutes forthe alarm to clear. This faultrequires a master reset to clear.If the alarm becomes activeagain (or if the original alarmnever clears), contact the PACcenter for assistance.

L30SUC_LLO • In startup control at toohigh of a load, CRT, or noton minimum IGV angle.

• Unable to enter cyclecontrol.

Start Up Control Load Lock OutAlarm

• ARES model has failed(see L83CA_F_A)

• Ensure compressor bleedvalves are closed.

• Sensor failures havedisabled the ARES model.See LCA_CSENS_Afor details on specificcombinations of sensorfailures).

• All CPD sensorsunavailable AND CPDsensor model not valid

• All CTD sensorsunavailable AND CTDsensor model not valid

• All DWATT sensorsunavailable AND DWATTsensor model not valid

• Refer to the table SensorFault Root Causes andRecommended Actions forrecommended actions tofix sensor failures.



L30TS2PSENS_A • Ejector system notoperational

SENSOR FAULTS - DISABLEEJECTOR SYSTEM

Inspect Ejector sensors. Referto the table Sensor Fault RootCauses and RecommendedActions for troubleshootingtips.

L30TSQPSENS_A • Ejector system notoperational

SENSOR FAULTS - DISABLEEJECTOR ISOLATIONVALVE

Inspect Ejector sensors. Referto the table Sensor Fault RootCauses and RecommendedActions for troubleshootingtips.

LCA_SENSTRN_A • Sensor training has notbeen performed or was notsuccessful

MBC RUNBACK DUE TOINADEQUATE SENSORTRAINING

Perform sensor training asdescribed in GEH-6810.

LTS2P_TRNP_A • Sensor training for TS2Phas not been performed orwas not successful

TS2P SENSOR HAS NOTBEEN TRAINED - PERFORMSENSOR TRAININGPROCEDURE


LTS2QP13_TRNP_A • Sensor training forTS2QP13 has not beenperformed or was notsuccessful

TS2QP13 SENSOR HASNOT BEEN TRAINED- PERFORM SENSORTRAINING PROCEDURE


LTS3QP9_TRNP_A • Sensor training forTS2QP9 has not beenperformed or was notsuccessful

TS3QP9 SENSOR HASNOT BEEN TRAINED- PERFORM SENSORTRAINING PROCEDURE


LWQ_TRNP_A • Sensor training for WQhas not been performed orwas not successful

WQ SENSOR HAS NOTBEEN TRAINED - PERFORMSENSOR TRAININGPROCEDURE


L3SENS_A • One or less CPD sensorsavailable OR

• One or less FPG2 sensorsavailable OR

• One or less CTIM sensorsavailable OR

• Two or more of thefollowing are true:

− One or less AFPAPsensors available

− One or less CPDsensors available

− One or less CTDsensors available

− Zero ITDP sensorsavailable

− Zero CPBH1 sensorsavailable

SENSOR FAULTS - INHIBITSTART

Start Inhibited

Examine sensor faults andsensor model validity changes(with associated alarms) thatcaused protective action. Referto the table Sensor Fault RootCauses and RecommendedActions for recommendedactions to fix sensor failures.




− One or less FPG2sensors availableAND not on totalliquid fuel

− One or less FTGsensors availableAND not on totalliquid fuel

− Zero FQG sensorsavailable AND not ontotal liquid fuel

− Zero FQLM1 sensorsavailable AND ontotal liquid fuel

L30LRSENS_A • Generator breaker closedAND all DWATT sensorsunavailable AND DWATTsensor model not valid

SENSOR FAULTS – LOADREJECT TO FSNL

Load Reject to FSNL


L86SENS_A • All FPG2 sensorsunavailable AND ontotal gas fuel OR

• All CPD sensorsunavailable AND CPDsensor model not validAND at minimumoperating speed ANDgenerator breaker closed

SENSOR FAULTS – TRIPUNIT

Trip


L94SENS_A • All FTG sensorsunavailable AND ontotal gas fuel OR

• Start permissive conditionsnot met AND breaker notclosed AND not trippedAND not already shuttingdown

SENSOR FAULTS –SHUTDOWN UNIT

Fired Shutdown Initiated


L3BHSENS_A • All CTIM sensorsunavailable AND CTIMsensor model not valid OR

• All AFPAP sensorsunavailable

SENSOR FAULTS - FAILBLEED HEAT OPEN

IBH System Failed Open (bysolenoid)




L3BHTSENS_A • CPBH2 sensor notavailable AND CQBHsensor model not valid

SENSOR FAULTS –DISABLEIBH DLN TURNDOWN

IBH DLN Turndown ScheduleDisabled,Minimum IGV AngleIncreased


L3WQSENS_A • WQ sensor failure of anytype detected

SENSOR FAULTS - DISABLEWATER INJECTION

Liquid Fuel Water InjectionSystem Disabled


L30LRSENS_A • Generator breaker closedAND all DWATT sensorsunavailable AND DWATTsensor model not valid

SENSOR FAULTS - LOADREJECT TO FSNL

Load Reject to FSNL


LCA_SENS_FDL30AFPAP_0L30AFPAP_1L30AFPAP_2L30AFPAP_DIFL30CPBH1_0L30CPBH2_0L30CPD_0L30CPD_1L30CPD_2L30CPD_DIFL30CTD_0L30CTD_1L30CTD_2L30CTD_DIFL30CTIM_0L30CTIM_1L30CTIM_2L30CTIM_DIFL30DWATT_0L30DWATT_1L30DWATT_DIFL30FPG2_0L30FPG2_1L30FPG2_2L30FPG2_DIFL30FPGN1_0

• A sensor fault in one ofthe monitored parameters(AFPAP, CPD, CTD,CTIM, DWATT, FPG2,FPGN1, FPGN2, FPGN3,FQG, FQLM1, FTG,ITDP, WQ, TS2P,TS2QP13, or TS3QP9) hasoccurred.

Fail degraded biases applied tomachine boundary targets asappropriate to accommodatethese sensor failures.

Refer to the table SensorFault Root Causes andRecommended Actions forrecommended OperatorActions.



L30FPGN1_1L30FPGN1_2L30FPGN1_DIFL30FPGN2_0L30FPGN2_1L30FPGN2_2L30FPGN2_DIFL30FPGN3_0L30FPGN3_1L30FPGN3_2L30FPGN3_DIFL30FPGN4_0L30FPGN4_1L30FPGN4_2L30FPGN4_DIFL30FQG_0L30FTG_0L30FTG_1L30FTG_2L30FTG_DIFL30ITDP_0L30ITDP_1L30ITDP_2L30ITDP_DIFL30WQ_0L30WQ_1L30WQ_DIFL30TS2P_0L30TS2P_1L30TS2P_2L30TS2P_DIFL30TS2QP13_0L30TS2QP13_1L30TS2QP13_2L30TS2QP13_DIFL30TS3QP9_0L30TS3QP9_1L30TS3QP9_2L30TS3QP9_DIF

LCA_ATSENS_A • All FPGN1 sensorsunavailable OR

• All FPGN2 sensorsunavailable OR

• All FPGN3 sensorsunavailable

SENSOR FAULTS –AUTOTUNE DISABLED

Slew Out of AutotuneMBC,FSR-VPR Loop Disabled




LCA_CSENS_A • All CTD sensorsunavailable AND CTDsensor model not valid OR

• All CTIM sensorsunavailable AND CTIMsensor model not valid OR

• FQLM1 sensor unavailableAND on total liquid fuelOR

• 2 or more of the followingare true:

− One or less AFPAPsensors available

− One or less CPDsensors available

− One or less CTDsensors available

− One or less DWATTsensors available

− Zero ITDP sensorsavailable



− One or less FPG2sensors availableAND not on totalliquid fuel

− One or less FTGsensors availableAND not on totalliquid fuel

− Zero FQG sensorsavailable AND not ontotal liquid fuel

− Zero FQLM1 sensorsavailable AND ontotal liquid fuel

− One or less WQsensors availableAND water injectionis on

SENSOR FAULTS – ARESDISABLED

Slew Out of ALCC (Step toSpinning Reserve)




L30CPDM • CPD sensor model is notvalid

CPD SENSOR MODELINVALID

CPD sensor model output isignored in downstream logic,ex. input signal processing(ISP).

Check health of input sensorsto model first (CPD, DWATT).Verify wiring, calibration,device integrity, and so forth.Replace if necessary.Repeat forall other ARES analog sensorinputs.

L30CTDM • CTD sensor model is notvalid

CTD SENSOR MODELINVALID

CTD sensor model output isignored in downstream logic,ex. input signal processing(ISP).

Check health of input sensorsto model first (CTD, DWATT).Verify wiring, calibration,device integrity, and so forth.Replace if necessary.Repeat forall other ARES analog sensorinputs.

L30DWATTM • DWATT sensor model isnot valid

DWATT SENSOR MODELINVALID

DWATT sensor model output isignored in downstream logic,ex. input signal processing(ISP).

Check health of input sensorsto model first (CPD, CTD).Verify wiring, calibration,device integrity, and so forth.Replace if necessary.Repeat forall other ARES analog sensorinputs.

L30CTIMM • CTIM sensor model is notvalid

CTIM SENSOR MODELINVALID

CTIM sensor model output isignored in downstream logic,ex. input signal processing(ISP).

Check health of input sensorsto model (CTD, CPBH1,CPBH2). Verify wiring,calibration, device integrity,and so forth. Replace ifnecessary.Turn off evaporativecooling.

L30FQGM • FQG sensor model is notvalid

FQG SENSOR MODELINVALID

FQG sensor model output isignored in downstream logic,ex. input signal processing(ISP).

Check health of input sensorsto model (CPD, FPG2, FTG,FPGN1, FPGN2, FPGN3).Verify wiring, calibration,device integrity, and so forth.Replace if necessary.

L30CQBHM • CQBH sensor model is notvalid

CQBH SENSOR MODELINVALID

CQBH sensor model output isignored in downstream logic,ex. input signal processing(ISP).

Check health of input sensors tomodel (CTD, CPBH1). Verifywiring, calibration, deviceintegrity, and so forth. Replaceif necessary.



L30ITDPM • ITDP sensor model is notvalid

ITDP SENSOR MODELINVALID

ITDP sensor model output isignored in downstream logic,ex. input signal processing(ISP).

Check health of input sensorsto model (AFPAP, CTIM).Verify wiring, calibration,device integrity, and so forth.Replace if necessary.

L30WQM • WQ sensor model is notvalid

WQ SENSOR MODELINVALID

WQ sensor model output isignored in downstream logic,ex. input signal processing(ISP).

Check health of input sensorsto model (WQDP). Verifywiring, calibration, deviceintegrity, and so forth. Replaceif necessary.


Sensor Fault Root Causes and Recommended Actions

Possible Causes of Sensor FaultDetection

Recommended Operator Actions

Control hardware failure • Examine I/O board/pack diagnostics log

• Check proper I/O layout/fanning to ensure single panel loss does not resultin sensor signal loss

• Ensure all controllers in controlling state (not inputs enabled, and such)

Dirty pneumatic or sensing lines • Find source of contamination and seal

• Clean or replace sensing lines

External interference • Check if new equipment recently installed near sensing lines/wiring

• Install additional electromagnetic shielding

• Ensure proper signal path separation from power wiring

• Re-route sensing lines/wiring

Faulty or broken transmitter • Physically inspect transmitter for damage, wear, and leakage

• Replace transmitter

• Perform signal injection tests to confirm proper operation

Incorrect I/O settings • Double-check device settings including I/O settings in application software

• Check I/O configuration in application software is consistent with panel layout

Loose, broken, and/or incorrect wiring • Ensure tight terminations in control cabinets and connections at the device

• Perform loop checks

Sensor not properly calibrated • Recalibrate the sensor

Sensors isolated or valved out • Remove isolation block (if present)

• Confirm sensors not in calibration mode

• Disengage isolation valve (if present)

Wrong installation • Thermocouples: check proper well installation and insertion depth

• Differential pressures: ensure lines piped to correct sides of transmitters

• Pressure transmitters: look for leakages

• Check sensor placed in proper physical location/tap


Glossary of TermsAdaptive Real-time Engine Simulation (ARES) is a high-fidelity model of thegas turbine that is continuously tuned in real-time to match the performance of the actualmachine.All Load Cycle Control (ALCC) is a technology that implements MBC directboundary control from breaker closure for the bulk fuel/air boundaries.AutoTune is a software product that adds closed loop DLN split control to ETS,enabling greater allowable MWI variation and elimination of seasonal retunes.Boundary is a limit, such as an operational limit or a design limit. An example wouldbe the typical 9 ppm NOx limitation for a 7FA+e gas turbine.Boundary Models are physics-based models that capture the fundamental behaviorof the operational boundaries.Coordinated Air-Fuel (CAF) is a control strategy used to maintain an operableglobal fuel-air mixture in the combustor during gas turbine transient events.Combustion Reference (CRT) is a control system parameter used to schedulecombustion mode transfer points and split schedules.Effectors are the control elements that alter machine operation; IGV, inlet bleed heat,total fuel flow, fuel temperature, and DLN fuel splits.Enhanced Transient Stability (ETS) is a product that utilizes the technologyplatform of ALCC and provides improved transient response of GE gas turbines usingMBCAF, the GFF, and transient fuel split biasing.Grid Frequency Filter (GFF) is a speed filter specifically designed for the ETSproduct utilized to shelter the gas turbine from the full effects of extreme frequencydisturbances.GE Control System Solutions (CSS toolbox) is a Windows®-based applicationused to configure Mark* VI control hardware and software.Health is a term that defines whether a variable is functioning as expected.Input Signal Processing (ISP) is a signal-processing-based fault detection,isolation, and accommodation strategy applied to all sensor inputs critical to the accurateoperation of ARES.Loop in Control (LIC) is a status indication that displays which control loop isgenerating the output reference for an effector.Model-based Control (MBC) is a control strategy designed to improve theperformance and operational flexibility of a GE gas turbine.Model-based Coordinated Air-Fuel (MBCAF) is a coordinated air-fuel strategyspecific to the ETS product that creates a model of an ideal IGV-to-FSR relationship thenuses that modeled relationship to control IGVs in response to a fast FSR motion.ToolboxST* application is a Windows-based application used to configure Mark Veand Mark VIe control hardware and software.Turbine Reference (TRT) is a control system parameter used to define proper ornominal turbine operation, predominantly at base load.


Notes


GE Energy1501 Roanoke Blvd.Salem, VA 24153–6492 USA

1 540 387 7000www.geenergy.com

http://www.geenergy.com

GEH-6810

Documents

ge products

understoodthat ge

ge requisition

technical information

nearest ge sales

ge energydocumentation

ge internal dc

ge energy geh